BIER

- see draft-ietf-bier-mpls-encapsulation-10
- midpoint, head and tail functions
- supported payload protocols; IPv4 and IPv6 only.

Change-Id: I59d7363bb6fdfdce8e4016a68a9c8f5a5e5791cb
Signed-off-by: Neale Ranns <nranns@cisco.com>

73 files changed, 10133 insertions, 147 deletions

diff --git a/src/vat/api_format.c b/src/vat/api_format.c
index eee5c743959..c75857bc111 100644
--- a/src/vat/api_format.c
+++ b/src/vat/api_format.c
@@ -5180,6 +5180,8 @@ _(ip_mroute_add_del_reply)                              \
 _(mpls_route_add_del_reply)                             \
 _(mpls_table_add_del_reply)                             \
 _(mpls_ip_bind_unbind_reply)                            \
+_(bier_route_add_del_reply)                             \
+_(bier_table_add_del_reply)                             \
 _(proxy_arp_add_del_reply)                              \
 _(proxy_arp_intfc_enable_disable_reply)                 \
 _(sw_interface_set_unnumbered_reply)                    \
@@ -5383,6 +5385,8 @@ _(IP_MROUTE_ADD_DEL_REPLY, ip_mroute_add_del_reply)			\
 _(MPLS_TABLE_ADD_DEL_REPLY, mpls_table_add_del_reply)			\
 _(MPLS_ROUTE_ADD_DEL_REPLY, mpls_route_add_del_reply)			\
 _(MPLS_IP_BIND_UNBIND_REPLY, mpls_ip_bind_unbind_reply)			\
+_(BIER_ROUTE_ADD_DEL_REPLY, bier_route_add_del_reply)			\
+_(BIER_TABLE_ADD_DEL_REPLY, bier_table_add_del_reply)			\
 _(PROXY_ARP_ADD_DEL_REPLY, proxy_arp_add_del_reply)                     \
 _(PROXY_ARP_INTFC_ENABLE_DISABLE_REPLY,                                 \
   proxy_arp_intfc_enable_disable_reply)                                 \
@@ -8548,6 +8552,154 @@ api_mpls_ip_bind_unbind (vat_main_t * vam)
 }
 
 static int
+api_bier_table_add_del (vat_main_t * vam)
+{
+  unformat_input_t *i = vam->input;
+  vl_api_bier_table_add_del_t *mp;
+  u8 is_add = 1;
+  u32 set = 0, sub_domain = 0, hdr_len = 3;
+  mpls_label_t local_label = MPLS_LABEL_INVALID;
+  int ret;
+
+  /* Parse args required to build the message */
+  while (unformat_check_input (i) != UNFORMAT_END_OF_INPUT)
+    {
+      if (unformat (i, "sub-domain %d", &sub_domain))
+	;
+      else if (unformat (i, "set %d", &set))
+	;
+      else if (unformat (i, "label %d", &local_label))
+	;
+      else if (unformat (i, "hdr-len %d", &hdr_len))
+	;
+      else if (unformat (i, "add"))
+	is_add = 1;
+      else if (unformat (i, "del"))
+	is_add = 0;
+      else
+	{
+	  clib_warning ("parse error '%U'", format_unformat_error, i);
+	  return -99;
+	}
+    }
+
+  if (MPLS_LABEL_INVALID == local_label)
+    {
+      errmsg ("missing label\n");
+      return -99;
+    }
+
+  /* Construct the API message */
+  M (BIER_TABLE_ADD_DEL, mp);
+
+  mp->bt_is_add = is_add;
+  mp->bt_label = ntohl (local_label);
+  mp->bt_tbl_id.bt_set = set;
+  mp->bt_tbl_id.bt_sub_domain = sub_domain;
+  mp->bt_tbl_id.bt_hdr_len_id = hdr_len;
+
+  /* send it... */
+  S (mp);
+
+  /* Wait for a reply... */
+  W (ret);
+
+  return (ret);
+}
+
+static int
+api_bier_route_add_del (vat_main_t * vam)
+{
+  unformat_input_t *i = vam->input;
+  vl_api_bier_route_add_del_t *mp;
+  u8 is_add = 1;
+  u32 set = 0, sub_domain = 0, hdr_len = 3, bp = 0;
+  ip4_address_t v4_next_hop_address;
+  ip6_address_t v6_next_hop_address;
+  u8 next_hop_set = 0;
+  u8 next_hop_proto_is_ip4 = 1;
+  mpls_label_t next_hop_out_label = MPLS_LABEL_INVALID;
+  int ret;
+
+  /* Parse args required to build the message */
+  while (unformat_check_input (i) != UNFORMAT_END_OF_INPUT)
+    {
+      if (unformat (i, "%U", unformat_ip4_address, &v4_next_hop_address))
+	{
+	  next_hop_proto_is_ip4 = 1;
+	  next_hop_set = 1;
+	}
+      else if (unformat (i, "%U", unformat_ip6_address, &v6_next_hop_address))
+	{
+	  next_hop_proto_is_ip4 = 0;
+	  next_hop_set = 1;
+	}
+      if (unformat (i, "sub-domain %d", &sub_domain))
+	;
+      else if (unformat (i, "set %d", &set))
+	;
+      else if (unformat (i, "hdr-len %d", &hdr_len))
+	;
+      else if (unformat (i, "bp %d", &bp))
+	;
+      else if (unformat (i, "add"))
+	is_add = 1;
+      else if (unformat (i, "del"))
+	is_add = 0;
+      else if (unformat (i, "out-label %d", &next_hop_out_label))
+	;
+      else
+	{
+	  clib_warning ("parse error '%U'", format_unformat_error, i);
+	  return -99;
+	}
+    }
+
+  if (!next_hop_set || (MPLS_LABEL_INVALID == next_hop_out_label))
+    {
+      errmsg ("next hop / label set\n");
+      return -99;
+    }
+  if (0 == bp)
+    {
+      errmsg ("bit=position not set\n");
+      return -99;
+    }
+
+  /* Construct the API message */
+  M2 (BIER_ROUTE_ADD_DEL, mp, sizeof (vl_api_fib_path3_t));
+
+  mp->br_is_add = is_add;
+  mp->br_tbl_id.bt_set = set;
+  mp->br_tbl_id.bt_sub_domain = sub_domain;
+  mp->br_tbl_id.bt_hdr_len_id = hdr_len;
+  mp->br_bp = ntohs (bp);
+  mp->br_n_paths = 1;
+  mp->br_paths[0].n_labels = 1;
+  mp->br_paths[0].label_stack[0] = ntohl (next_hop_out_label);
+  mp->br_paths[0].afi = (next_hop_proto_is_ip4 ? 0 : 1);
+
+  if (next_hop_proto_is_ip4)
+    {
+      clib_memcpy (mp->br_paths[0].next_hop,
+		   &v4_next_hop_address, sizeof (v4_next_hop_address));
+    }
+  else
+    {
+      clib_memcpy (mp->br_paths[0].next_hop,
+		   &v6_next_hop_address, sizeof (v6_next_hop_address));
+    }
+
+  /* send it... */
+  S (mp);
+
+  /* Wait for a reply... */
+  W (ret);
+
+  return (ret);
+}
+
+static int
 api_proxy_arp_add_del (vat_main_t * vam)
 {
   unformat_input_t *i = vam->input;
@@ -22238,6 +22390,12 @@ _(mpls_ip_bind_unbind,                                                  \
 _(mpls_tunnel_add_del,                                                  \
   " via <addr> [table-id <n>]\n"                                        \
   "sw_if_index <id>] [l2]  [del]")                                      \
+_(bier_table_add_del,                                                   \
+  "<label> <sub-domain> <set> <bsl> [del]")                             \
+_(bier_route_add_del,                                                   \
+  "<bit-position> <sub-domain> <set> <bsl> via <addr> [table-id <n>]\n" \
+  "[<intfc> | sw_if_index <id>]"                                        \
+  "[weight <n>] [del] [multipath]")                                     \
 _(proxy_arp_add_del,                                                    \
   "<lo-ip4-addr> - <hi-ip4-addr> [vrf <n>] [del]")                      \
 _(proxy_arp_intfc_enable_disable,                                       \
diff --git a/src/vnet.am b/src/vnet.am
index f5548307647..3204a5b5598 100644
--- a/src/vnet.am
+++ b/src/vnet.am
@@ -1112,6 +1112,40 @@ libvnet_la_SOURCES +=                    \
   vnet/util/trajectory.c
 
 ########################################
+# BIER
+########################################
+
+libvnet_la_SOURCES +=                           \
+  vnet/bier/bier_bit_string.c                   \
+  vnet/bier/bier_entry.c                        \
+  vnet/bier/bier_fmask.c                        \
+  vnet/bier/bier_fmask_db.c                     \
+  vnet/bier/bier_input.c                   	\
+  vnet/bier/bier_lookup.c                  	\
+  vnet/bier/bier_output.c                  	\
+  vnet/bier/bier_table.c                        \
+  vnet/bier/bier_types.c                        \
+  vnet/bier/bier_test.c                         \
+  vnet/bier/bier_api.c                          \
+  vnet/bier/bier_drop.c                         \
+  vnet/bier/bier_update.c			\
+  vnet/bier/bier_imp_node.c			\
+  vnet/bier/bier_imp.c				\
+  vnet/bier/bier_disp_entry.c			\
+  vnet/bier/bier_disp_lookup_node.c		\
+  vnet/bier/bier_disp_dispatch_node.c		\
+  vnet/bier/bier_disp_table.c
+
+nobase_include_HEADERS +=			\
+  vnet/bier/bier_types.h                        \
+  vnet/bier/bier_entry.h                        \
+  vnet/bier/bier_update.h                       \
+  vnet/bier/bier.api.h                          \
+  vnet/bier/bier_table.h
+
+API_FILES += vnet/bier/bier.api
+
+########################################
 # Plugin client library
 ########################################
 
diff --git a/src/vnet/bier/bier.api b/src/vnet/bier/bier.api
new file mode 100644
index 00000000000..466524cc6a3
--- /dev/null
+++ b/src/vnet/bier/bier.api
@@ -0,0 +1,268 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/** \file
+
+    This file defines vpp BIER control-plane API messages which are generally
+    called through a shared memory interface.
+*/
+vl_api_version 1.0.0
+
+/** \brief BIER Table Indentifier
+    @param bt_set
+    @param bt_sub_domain
+    @param bt_bit_header_length
+*/
+typeonly define bier_table_id
+{
+  u8 bt_set;
+  u8 bt_sub_domain;
+  u8 bt_hdr_len_id;
+};
+
+/** \brief BIER Table Add / del route
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param bt_tbl_id - The BIER table-id the route is added in
+    @param bt_mpls_label - The MPLS label for the table
+    @param bt_is_add - Is this a route add or delete
+*/
+autoreply define bier_table_add_del
+{
+  u32 client_index;
+  u32 context;
+  vl_api_bier_table_id_t bt_tbl_id;
+  u32 bt_label;
+  u8 bt_is_add;
+};
+
+define bier_table_dump
+{
+  u32 client_index;
+  u32 context;
+};
+
+define bier_table_details
+{
+  u32 context;
+  u32 bt_label;
+  vl_api_bier_table_id_t bt_tbl_id;
+};
+
+/** \brief FIB path
+    @param sw_if_index - index of the interface
+    @param weight - The weight, for UCMP
+    @param preference - The preference of the path. lowest preference is prefered
+    @param is_local - local if non-zero, else remote
+    @param is_drop - Drop the packet
+    @param is_unreach - Drop the packet and rate limit send ICMP unreachable
+    @param is_prohibit - Drop the packet and rate limit send ICMP prohibited
+    @param afi - the afi of the next hop, IP46_TYPE_IP4=1, IP46_TYPE_IP6=2
+    @param next_hop[16] - the next hop address
+
+    WARNING: this type is replicated, pending cleanup completion
+*/
+typeonly define fib_path3
+{
+  u32 sw_if_index;
+  u32 table_id;
+  u8 weight;
+  u8 preference;
+  u8 is_local;
+  u8 is_drop;
+  u8 is_unreach;
+  u8 is_prohibit;
+  u8 afi;
+  u8 next_hop[16];
+  u32 rpf_id;
+  u8 n_labels;
+  u32 label_stack[16];
+};
+
+/** \brief BIER Route Add / del route
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param br_bp - The Bit-position value
+    @param br_tbl_id - The BIER table-id the route is added in
+    @param br_is_add - Is this a route add or delete
+    @param br_is_replace - Are the paths specfied replacing those already
+                           present or are they to be combined.
+    @param br_n_paths - The number of paths
+    @param br_paths - The array of paths
+*/
+autoreply define bier_route_add_del
+{
+  u32 client_index;
+  u32 context;
+  u16 br_bp;
+  u8 br_is_add;
+  u8 br_is_replace;
+  vl_api_bier_table_id_t br_tbl_id;
+  u8 br_n_paths;
+  vl_api_fib_path3_t br_paths[br_n_paths];
+};
+
+define bier_route_dump
+{
+  u32 client_index;
+  u32 context;
+  vl_api_bier_table_id_t br_tbl_id;
+};
+
+define bier_route_details
+{
+  u32 client_index;
+  u32 context;
+  u16 br_bp;
+  vl_api_bier_table_id_t br_tbl_id;
+  u32 br_n_paths;
+  vl_api_fib_path3_t br_paths[br_n_paths];
+};
+
+/** \brief BIER Imposition Add
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param bi_tbl_id - The BIER table-id used to forward post encap
+    @param bi_src - The source Bit-position in the encap.
+    @param bi_is_add - Is this a route add or delete
+    @param bi_n_bytes - The number of bytes in the following bit-string
+    @param bi_bytes - The bit-string represented as a byte array
+*/
+define bier_imp_add
+{
+  u32 client_index;
+  u32 context;
+  vl_api_bier_table_id_t bi_tbl_id;
+  u16 bi_src;
+  u8 bi_is_add;
+  u8 bi_n_bytes;
+  u8 bi_bytes[bi_n_bytes];
+};
+
+/** \brief Reply for BIER route add / del request
+    @param context - returned sender context, to match reply w/ request
+    @param retval - return code
+    @param bi_index - The index of the created imposition object.
+*/
+define bier_imp_add_reply
+{
+  u32 context;
+  i32 retval;
+  u32 bi_index;
+};
+
+/** \brief BIER Imposition Del
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param bi_index - The index of the imposition object (as returned
+		      from the ADD)
+*/
+autoreply define bier_imp_del
+{
+  u32 client_index;
+  u32 context;
+  u32 bi_index;
+};
+
+define bier_imp_dump
+{
+  u32 client_index;
+  u32 context;
+};
+
+define bier_imp_details
+{
+  u32 client_index;
+  u32 context;
+  vl_api_bier_table_id_t bi_tbl_id;
+  u16 bi_src;
+  u8 bi_n_bytes;
+  u8 bi_bytes[bi_n_bytes];
+};
+
+/** \brief BIER Disposition Table Add / del route
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param bt_tbl_id - The BIER Disposition table-id.
+*/
+autoreply define bier_disp_table_add_del
+{
+  u32 client_index;
+  u32 context;
+  u32 bdt_tbl_id;
+  u8 bdt_is_add;
+};
+
+define bier_disp_table_dump
+{
+  u32 client_index;
+  u32 context;
+};
+
+define bier_disp_table_details
+{
+  u32 context;
+  u32 bdt_tbl_id;
+};
+
+/** \brief BIER Disposition Entry Add / del
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param bde_bp - The Bit-position value for the entry
+    @param bde_tbl_id - The BIER dispositiontable-id the route is added in
+    @param bde_next_hop_sw_if_index - the nextop interface
+    @param bde_is_add - Is this a route add or delete
+    @param bde_payload_proto - The payload protocol for which the next-hop
+			       is added
+    @param bde_next_hop_table_id - The table ID for the next-hop
+    @param bde_next_hop_proto_is_ip4 - The next-hop is IPV4
+    @param bde_next_hop[16] - the nextop address.
+                              Set this to all 0s for dispostion.
+*/
+autoreply define bier_disp_entry_add_del
+{
+  u32 client_index;
+  u32 context;
+  u16 bde_bp;
+  u32 bde_tbl_id;
+  u8 bde_is_add;
+  u8 bde_payload_proto;
+  u8 bde_n_paths;
+  vl_api_fib_path3_t bde_paths[bde_n_paths];
+};
+
+define bier_disp_entry_dump
+{
+  u32 client_index;
+  u32 context;
+  u32 bde_tbl_id;
+};
+
+define bier_disp_entry_details
+{
+  u32 context;
+  u16 bde_bp;
+  u32 bde_tbl_id;
+  u8 bde_is_add;
+  u8 bde_payload_proto;
+  u8 bde_n_paths;
+  vl_api_fib_path3_t bde_paths[bde_n_paths];
+};
+
+/*
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */
diff --git a/src/vnet/bier/bier_api.c b/src/vnet/bier/bier_api.c
new file mode 100644
index 00000000000..aacee0910bf
--- /dev/null
+++ b/src/vnet/bier/bier_api.c
@@ -0,0 +1,690 @@
+/*
+ *------------------------------------------------------------------
+ * bier_api.c - vnet BIER api
+ *
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *------------------------------------------------------------------
+ */
+
+#include <vnet/vnet.h>
+#include <vlibmemory/api.h>
+
+#include <vnet/api_errno.h>
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_imp.h>
+#include <vnet/bier/bier_disp_table.h>
+#include <vnet/bier/bier_disp_entry.h>
+#include <vnet/bier/bier_bit_string.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+#include <vnet/fib/fib_path_list.h>
+#include <vnet/fib/fib_api.h>
+#include <vnet/fib/fib_table.h>
+#include <vnet/mfib/mfib_table.h>
+
+#include <vnet/vnet_msg_enum.h>
+
+#define vl_typedefs		/* define message structures */
+#include <vnet/vnet_all_api_h.h>
+#undef vl_typedefs
+
+#define vl_endianfun		/* define message structures */
+#include <vnet/vnet_all_api_h.h>
+#undef vl_endianfun
+
+/* instantiate all the print functions we know about */
+#define vl_print(handle, ...) vlib_cli_output (handle, __VA_ARGS__)
+#define vl_printfun
+#include <vnet/vnet_all_api_h.h>
+#undef vl_printfun
+
+#include <vlibapi/api_helper_macros.h>
+
+#define foreach_bier_api_msg                            \
+    _(BIER_TABLE_ADD_DEL, bier_table_add_del)           \
+    _(BIER_TABLE_DUMP, bier_table_dump)                 \
+    _(BIER_ROUTE_ADD_DEL, bier_route_add_del)           \
+    _(BIER_ROUTE_DUMP, bier_route_dump)                 \
+    _(BIER_IMP_ADD, bier_imp_add)                       \
+    _(BIER_IMP_DEL, bier_imp_del)                       \
+    _(BIER_IMP_DUMP, bier_imp_dump)                     \
+    _(BIER_DISP_TABLE_ADD_DEL, bier_disp_table_add_del) \
+    _(BIER_DISP_TABLE_DUMP, bier_disp_table_dump)       \
+    _(BIER_DISP_ENTRY_ADD_DEL, bier_disp_entry_add_del) \
+    _(BIER_DISP_ENTRY_DUMP, bier_disp_entry_dump)
+
+static void
+vl_api_bier_table_add_del_t_handler (vl_api_bier_table_add_del_t * mp)
+{
+    vl_api_bier_table_add_del_reply_t *rmp;
+    vnet_main_t *vnm;
+    int rv;
+
+    vnm = vnet_get_main ();
+    vnm->api_errno = 0;
+
+    bier_table_id_t bti = {
+        .bti_set = mp->bt_tbl_id.bt_set,
+        .bti_sub_domain = mp->bt_tbl_id.bt_sub_domain,
+        .bti_hdr_len = mp->bt_tbl_id.bt_hdr_len_id,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+
+    if (mp->bt_is_add)
+    {
+        bier_table_add_or_lock(&bti, ntohl(mp->bt_label));
+    }
+    else
+    {
+        bier_table_unlock(&bti);
+    }
+
+    rv = vnm->api_errno;
+
+    REPLY_MACRO (VL_API_BIER_TABLE_ADD_DEL_REPLY);
+}
+
+static void
+send_bier_table_details (unix_shared_memory_queue_t * q,
+                         u32 context,
+                         const bier_table_t *bt)
+{
+    vl_api_bier_table_details_t *mp;
+
+    mp = vl_msg_api_alloc(sizeof(*mp));
+    if (!mp)
+        return;
+    memset(mp, 0, sizeof(*mp));
+    mp->_vl_msg_id = ntohs(VL_API_BIER_TABLE_DETAILS);
+    mp->context = context;
+
+    mp->bt_label = bt->bt_ll;
+    mp->bt_tbl_id.bt_set = bt->bt_id.bti_set;
+    mp->bt_tbl_id.bt_sub_domain = bt->bt_id.bti_sub_domain;
+    mp->bt_tbl_id.bt_hdr_len_id = bt->bt_id.bti_hdr_len;
+
+    vl_msg_api_send_shmem (q, (u8 *) & mp);
+}
+
+static void
+vl_api_bier_table_dump_t_handler (vl_api_bier_table_dump_t * mp)
+{
+    unix_shared_memory_queue_t *q;
+    bier_table_t *bt;
+
+    q = vl_api_client_index_to_input_queue (mp->client_index);
+    if (q == 0)
+        return;
+
+    pool_foreach(bt, bier_table_pool,
+    ({
+        /*
+         * skip the ecmp tables.
+         */
+        if (bier_table_is_main(bt))
+        {
+            send_bier_table_details(q, mp->context, bt);
+        }
+    }));
+}
+
+static void
+vl_api_bier_route_add_del_t_handler (vl_api_bier_route_add_del_t * mp)
+{
+    vl_api_bier_route_add_del_reply_t *rmp;
+    fib_route_path_t *brpaths, *brpath;
+    vnet_main_t *vnm;
+    bier_bp_t bp;
+    int rv = 0;
+    u8 ii, jj;
+
+    vnm = vnet_get_main ();
+    vnm->api_errno = 0;
+
+    bp = ntohs(mp->br_bp);
+    brpaths = NULL;
+
+    if (0 == bp || bp > 0xffff)
+    {
+        rv = -1;
+        goto done;
+    }
+
+    bier_table_id_t bti = {
+        .bti_set = mp->br_tbl_id.bt_set,
+        .bti_sub_domain = mp->br_tbl_id.bt_sub_domain,
+        .bti_hdr_len = mp->br_tbl_id.bt_hdr_len_id,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+
+    vec_validate(brpaths, mp->br_n_paths - 1);
+
+    vec_foreach_index(ii, brpaths)
+    {
+        brpath = &brpaths[ii];
+        memset(brpath, 0, sizeof(*brpath));
+        brpath->frp_flags = FIB_ROUTE_PATH_BIER_FMASK;
+
+        vec_validate(brpath->frp_label_stack,
+                     mp->br_paths[ii].n_labels);
+        for (jj = 0; jj < mp->br_paths[ii].n_labels; jj++)
+        {
+            brpath->frp_label_stack[jj] =
+                ntohl(mp->br_paths[ii].label_stack[jj]);
+        }
+
+        if (0 == mp->br_paths[ii].afi)
+        {
+            clib_memcpy (&brpath->frp_addr.ip4,
+                         mp->br_paths[ii].next_hop,
+                         sizeof (brpath->frp_addr.ip4));
+        }
+        else
+        {
+            clib_memcpy (&brpath->frp_addr.ip6,
+                         mp->br_paths[ii].next_hop,
+                         sizeof (brpath->frp_addr.ip6));
+        }
+        if (ip46_address_is_zero(&brpath->frp_addr))
+        {
+            index_t bdti;
+
+            bdti = bier_disp_table_find(ntohl(mp->br_paths[ii].table_id));
+
+            if (INDEX_INVALID != bdti)
+                brpath->frp_fib_index = bdti;
+            else
+            {
+                rv = VNET_API_ERROR_NO_SUCH_FIB;
+                goto done;
+            }
+        }
+    }
+
+    if (mp->br_is_add)
+    {
+        bier_table_route_add(&bti, ntohs(mp->br_bp), brpaths);
+    }
+    else
+    {
+        bier_table_route_remove(&bti, ntohs(mp->br_bp), brpaths);
+    }
+
+done:
+    vec_free(brpaths);
+    rv = (rv == 0) ? vnm->api_errno : rv;
+
+    REPLY_MACRO (VL_API_BIER_ROUTE_ADD_DEL_REPLY);
+}
+
+typedef struct bier_route_details_walk_t_
+{
+    unix_shared_memory_queue_t * q;
+    u32 context;
+} bier_route_details_walk_t;
+
+static void
+send_bier_route_details (const bier_table_t *bt,
+                         const bier_entry_t *be,
+                         void *args)
+{
+    fib_route_path_encode_t *api_rpaths = NULL, *api_rpath;
+    bier_route_details_walk_t *ctx = args;
+    vl_api_bier_route_details_t *mp;
+    vl_api_fib_path3_t *fp;
+    u32 n_paths, m_size;
+
+    n_paths = fib_path_list_get_n_paths(be->be_path_list);
+    m_size = sizeof(*mp) + (n_paths * sizeof(vl_api_fib_path3_t));
+    mp = vl_msg_api_alloc(m_size);
+    if (!mp)
+        return;
+
+    memset(mp, 0, m_size);
+    mp->_vl_msg_id = ntohs(VL_API_BIER_ROUTE_DETAILS);
+    mp->context = ctx->context;
+
+    mp->br_tbl_id.bt_set = bt->bt_id.bti_set;
+    mp->br_tbl_id.bt_sub_domain = bt->bt_id.bti_sub_domain;
+    mp->br_tbl_id.bt_hdr_len_id = bt->bt_id.bti_hdr_len;
+    mp->br_bp = htons(be->be_bp);
+    mp->br_n_paths = htonl(n_paths);
+
+    fib_path_list_walk(be->be_path_list, fib_path_encode, &api_rpaths);
+
+    fp = mp->br_paths;
+    vec_foreach (api_rpath, api_rpaths)
+    {
+        fp->weight = api_rpath->rpath.frp_weight;
+        fp->preference = api_rpath->rpath.frp_preference;
+        fp->sw_if_index = htonl (api_rpath->rpath.frp_sw_if_index);
+        fp->n_labels = 0;
+        copy_fib_next_hop (api_rpath, fp);
+        fp++;
+    }
+
+    vl_msg_api_send_shmem (ctx->q, (u8 *) & mp);
+}
+
+static void
+vl_api_bier_route_dump_t_handler (vl_api_bier_route_dump_t * mp)
+{
+    unix_shared_memory_queue_t *q;
+
+    q = vl_api_client_index_to_input_queue (mp->client_index);
+    if (q == 0)
+        return;
+
+    bier_table_id_t bti = {
+        .bti_set = mp->br_tbl_id.bt_set,
+        .bti_sub_domain = mp->br_tbl_id.bt_sub_domain,
+        .bti_hdr_len = mp->br_tbl_id.bt_hdr_len_id,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+    bier_route_details_walk_t ctx = {
+        .q = q,
+        .context = mp->context,
+    };
+    bier_table_walk(&bti, send_bier_route_details, &ctx);
+}
+
+static void
+vl_api_bier_imp_add_t_handler (vl_api_bier_imp_add_t * mp)
+{
+    vl_api_bier_imp_add_reply_t *rmp;
+    vnet_main_t *vnm;
+    index_t bii;
+    int rv = 0;
+
+    vnm = vnet_get_main ();
+    vnm->api_errno = 0;
+
+    bier_table_id_t bti = {
+        .bti_set = mp->bi_tbl_id.bt_set,
+        .bti_sub_domain = mp->bi_tbl_id.bt_sub_domain,
+        .bti_hdr_len = mp->bi_tbl_id.bt_hdr_len_id,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+    bier_bit_string_t bs = {
+        .bbs_len = mp->bi_n_bytes,
+        .bbs_buckets = mp->bi_bytes,
+    };
+
+    bii = bier_imp_add_or_lock(&bti, ntohs(mp->bi_src), &bs);
+
+    /* *INDENT-OFF* */
+    REPLY_MACRO2 (VL_API_BIER_IMP_ADD_REPLY,
+    ({
+        rmp->bi_index = bii;
+    }));
+    /* *INDENT-OM* */
+}
+
+static void
+vl_api_bier_imp_del_t_handler (vl_api_bier_imp_del_t * mp)
+{
+    vl_api_bier_imp_del_reply_t *rmp;
+    vnet_main_t *vnm;
+    int rv = 0;
+
+    vnm = vnet_get_main ();
+    vnm->api_errno = 0;
+
+    bier_imp_unlock(ntohl(mp->bi_index));
+
+    REPLY_MACRO(VL_API_BIER_IMP_DEL_REPLY);
+}
+
+static void
+send_bier_imp_details (unix_shared_memory_queue_t * q,
+                       u32 context,
+                       const bier_imp_t *bi)
+{
+    vl_api_bier_imp_details_t *mp;
+    bier_hdr_t copy;
+    u8 n_bytes;
+
+    copy = bi->bi_hdr;
+    bier_hdr_ntoh(&copy);
+
+    n_bytes = bier_hdr_len_id_to_num_bytes(
+                  bier_hdr_get_len_id(&copy));
+    mp = vl_msg_api_alloc(sizeof(*mp) + n_bytes);
+    if (!mp)
+        return;
+    memset(mp, 0, sizeof(*mp)+n_bytes);
+    mp->_vl_msg_id = ntohs(VL_API_BIER_IMP_DETAILS);
+    mp->context = context;
+
+    mp->bi_tbl_id.bt_set = bi->bi_tbl.bti_set;
+    mp->bi_tbl_id.bt_sub_domain = bi->bi_tbl.bti_sub_domain;
+    mp->bi_tbl_id.bt_hdr_len_id = bi->bi_tbl.bti_hdr_len;
+
+
+    mp->bi_src = htons(bier_hdr_get_src_id(&copy));
+    mp->bi_n_bytes = n_bytes;
+    memcpy(mp->bi_bytes, bi->bi_bits.bits, n_bytes);
+
+    vl_msg_api_send_shmem (q, (u8 *) & mp);
+}
+
+static void
+vl_api_bier_imp_dump_t_handler (vl_api_bier_imp_dump_t * mp)
+{
+    unix_shared_memory_queue_t *q;
+    bier_imp_t *bi;
+
+    q = vl_api_client_index_to_input_queue (mp->client_index);
+    if (q == 0)
+        return;
+
+    pool_foreach(bi, bier_imp_pool,
+    ({
+        send_bier_imp_details(q, mp->context, bi);
+    }));
+}
+
+static void
+vl_api_bier_disp_table_add_del_t_handler (vl_api_bier_disp_table_add_del_t * mp)
+{
+    vl_api_bier_disp_table_add_del_reply_t *rmp;
+    vnet_main_t *vnm;
+    u32 table_id;
+    int rv;
+
+    vnm = vnet_get_main ();
+    vnm->api_errno = 0;
+    table_id = ntohl(mp->bdt_tbl_id);
+
+    if (mp->bdt_is_add)
+    {
+        bier_disp_table_add_or_lock(table_id);
+    }
+    else
+    {
+        bier_disp_table_unlock_w_table_id(table_id);
+    }
+
+    rv = vnm->api_errno;
+
+    REPLY_MACRO (VL_API_BIER_DISP_TABLE_ADD_DEL_REPLY);
+}
+
+static void
+send_bier_disp_table_details (unix_shared_memory_queue_t * q,
+                              u32 context,
+                              const bier_disp_table_t *bdt)
+{
+    vl_api_bier_disp_table_details_t *mp;
+
+    mp = vl_msg_api_alloc(sizeof(*mp));
+    if (!mp)
+        return;
+    memset(mp, 0, sizeof(*mp));
+    mp->_vl_msg_id = ntohs(VL_API_BIER_DISP_TABLE_DETAILS);
+    mp->context = context;
+
+    mp->bdt_tbl_id = htonl(bdt->bdt_table_id);
+
+    vl_msg_api_send_shmem (q, (u8 *) & mp);
+}
+
+static void
+vl_api_bier_disp_table_dump_t_handler (vl_api_bier_disp_table_dump_t * mp)
+{
+    unix_shared_memory_queue_t *q;
+    bier_disp_table_t *bdt;
+
+    q = vl_api_client_index_to_input_queue (mp->client_index);
+    if (q == 0)
+        return;
+
+    pool_foreach(bdt, bier_disp_table_pool,
+    ({
+        send_bier_disp_table_details(q, mp->context, bdt);
+    }));
+}
+
+static void
+vl_api_bier_disp_entry_add_del_t_handler (vl_api_bier_disp_entry_add_del_t * mp)
+{
+    vl_api_bier_disp_entry_add_del_reply_t *rmp;
+    fib_route_path_t *brps = NULL, *brp;
+    vnet_main_t *vnm;
+    bier_bp_t bp;
+    u32 table_id;
+    int rv = 0;
+    u32 ii;
+
+    vnm = vnet_get_main ();
+    vnm->api_errno = 0;
+    table_id = ntohl(mp->bde_tbl_id);
+    bp = ntohs(mp->bde_bp);
+
+    if (0 == bp || bp > 0xffff)
+    {
+        rv = -1;
+        goto done;
+    }
+
+    vec_validate(brps, mp->bde_n_paths - 1);
+    vec_foreach_index(ii, brps)
+    {
+        brp = &brps[ii];
+        brp->frp_fib_index = ntohl(mp->bde_paths[ii].table_id);
+        brp->frp_sw_if_index = ntohl(mp->bde_paths[ii].sw_if_index);
+
+        if (~0 != ntohl(mp->bde_paths[ii].rpf_id))
+        {
+            brp->frp_flags = FIB_ROUTE_PATH_RPF_ID;
+            brp->frp_rpf_id = ntohl(mp->bde_paths[ii].rpf_id);
+        }
+
+        if (0 == mp->bde_paths[ii].afi)
+        {
+            clib_memcpy (&brp->frp_addr.ip4,
+                         mp->bde_paths[ii].next_hop,
+                         sizeof (brp->frp_addr.ip4));
+        }
+        else
+        {
+            clib_memcpy (&brp->frp_addr.ip6,
+                         mp->bde_paths[ii].next_hop,
+                         sizeof (brp->frp_addr.ip6));
+        }
+        if (ip46_address_is_zero(&brp->frp_addr))
+        {
+            index_t fti;
+
+            switch (mp->bde_payload_proto)
+            {
+            case BIER_HDR_PROTO_INVALID:
+            case BIER_HDR_PROTO_MPLS_DOWN_STREAM:
+            case BIER_HDR_PROTO_MPLS_UP_STREAM:
+            case BIER_HDR_PROTO_ETHERNET:
+            case BIER_HDR_PROTO_VXLAN:
+            case BIER_HDR_PROTO_CTRL:
+            case BIER_HDR_PROTO_OAM:
+                rv = VNET_API_ERROR_UNSUPPORTED;
+                goto done;
+                break;
+            case BIER_HDR_PROTO_IPV4:
+            case BIER_HDR_PROTO_IPV6:
+            {
+                fib_protocol_t fproto;
+
+                fproto = (mp->bde_payload_proto == BIER_HDR_PROTO_IPV4 ?
+                          FIB_PROTOCOL_IP4 :
+                          FIB_PROTOCOL_IP6);
+
+                if (brp->frp_flags & FIB_ROUTE_PATH_RPF_ID)
+                {
+                    fti = mfib_table_find (fproto,
+                                           ntohl (mp->bde_paths[ii].table_id));
+                }
+                else
+                {
+                    fti = fib_table_find (fproto,
+                                          ntohl (mp->bde_paths[ii].table_id));
+                }
+
+                if (INDEX_INVALID != fti)
+                {
+                    brp->frp_fib_index = fti;
+                }
+                else
+                {
+                    rv = VNET_API_ERROR_NO_SUCH_FIB;
+                    goto done;
+                }
+                break;
+            }
+            }
+        }
+    }
+
+    if (mp->bde_is_add)
+    {
+        bier_disp_table_entry_path_add(table_id, bp,
+                                       mp->bde_payload_proto,
+                                       brps);
+    }
+    else
+    {
+        bier_disp_table_entry_path_remove(table_id, bp,
+                                          mp->bde_payload_proto,
+                                          brps);
+    }
+
+done:
+    vec_free(brps);
+    rv = (rv == 0) ? vnm->api_errno : rv;
+
+    REPLY_MACRO (VL_API_BIER_DISP_ENTRY_ADD_DEL_REPLY);
+}
+
+typedef struct bier_disp_entry_details_walk_t_
+{
+    unix_shared_memory_queue_t * q;
+    u32 context;
+} bier_disp_entry_details_walk_t;
+
+static void
+send_bier_disp_entry_details (const bier_disp_table_t *bdt,
+                              const bier_disp_entry_t *bde,
+                              u16 bp,
+                              void *args)
+{
+    fib_route_path_encode_t *api_rpaths = NULL, *api_rpath;
+    bier_disp_entry_details_walk_t *ctx = args;
+    vl_api_bier_disp_entry_details_t *mp;
+    bier_hdr_proto_id_t pproto;
+    vl_api_fib_path3_t *fp;
+    u32 n_paths, m_size;
+
+    FOR_EACH_BIER_HDR_PROTO(pproto)
+    {
+        fib_node_index_t pl = bde->bde_pl[pproto];
+        if (INDEX_INVALID != pl)
+        {
+            n_paths = fib_path_list_get_n_paths(pl);
+            m_size = sizeof(*mp) + (n_paths * sizeof(vl_api_fib_path3_t));
+            mp = vl_msg_api_alloc(m_size);
+            if (!mp)
+                return;
+
+            memset(mp, 0, m_size);
+            mp->_vl_msg_id = ntohs(VL_API_BIER_DISP_ENTRY_DETAILS);
+            mp->context = ctx->context;
+
+            mp->bde_tbl_id = htonl(bdt->bdt_table_id);
+            mp->bde_n_paths = htonl(n_paths);
+            mp->bde_payload_proto = pproto;
+            mp->bde_bp = htons(bp);
+
+            fib_path_list_walk(pl, fib_path_encode, &api_rpaths);
+
+            fp = mp->bde_paths;
+            vec_foreach (api_rpath, api_rpaths)
+            {
+                fp->weight = api_rpath->rpath.frp_weight;
+                fp->preference = api_rpath->rpath.frp_preference;
+                fp->sw_if_index = htonl (api_rpath->rpath.frp_sw_if_index);
+                fp->n_labels = 0;
+                copy_fib_next_hop (api_rpath, fp);
+                fp++;
+            }
+
+            vl_msg_api_send_shmem (ctx->q, (u8 *) & mp);
+        }
+    }
+}
+
+static void
+vl_api_bier_disp_entry_dump_t_handler (vl_api_bier_disp_entry_dump_t * mp)
+{
+    unix_shared_memory_queue_t *q;
+
+    q = vl_api_client_index_to_input_queue (mp->client_index);
+    if (q == 0)
+        return;
+
+    bier_disp_entry_details_walk_t ctx = {
+        .q = q,
+        .context = mp->context,
+    };
+    bier_disp_table_walk(ntohl(mp->bde_tbl_id),
+                         send_bier_disp_entry_details,
+                         &ctx);
+}
+
+#define vl_msg_name_crc_list
+#include <vnet/bier/bier.api.h>
+#undef vl_msg_name_crc_list
+
+static void
+setup_message_id_table (api_main_t * am)
+{
+#define _(id,n,crc) vl_msg_api_add_msg_name_crc (am, #n "_" #crc, id);
+    foreach_vl_msg_name_crc_bier;
+#undef _
+}
+
+static clib_error_t *
+bier_api_hookup (vlib_main_t * vm)
+{
+    api_main_t *am = &api_main;
+
+#define _(N,n)                                          \
+    vl_msg_api_set_handlers(VL_API_##N, #n,             \
+                            vl_api_##n##_t_handler,     \
+                            vl_noop_handler,            \
+                            vl_api_##n##_t_endian,      \
+                            vl_api_##n##_t_print,       \
+                            sizeof(vl_api_##n##_t), 1);
+    foreach_bier_api_msg;
+#undef _
+
+    /*
+     * Set up the (msg_name, crc, message-id) table
+     */
+    setup_message_id_table (am);
+
+    return 0;
+}
+
+VLIB_API_INIT_FUNCTION (bier_api_hookup);
diff --git a/src/vnet/bier/bier_bit_string.c b/src/vnet/bier/bier_bit_string.c
new file mode 100644
index 00000000000..aa42e083be8
--- /dev/null
+++ b/src/vnet/bier/bier_bit_string.c
@@ -0,0 +1,109 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/vnet.h>
+
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_bit_string.h>
+
+/*
+ * the first bit in the first byte is bit position 1.
+ * bit position 0 is not valid
+ */
+#define BIER_GET_STRING_POS(_bp, _byte, _bit, _str)             \
+{                                                               \
+    _bp--;                                                      \
+    _byte = ((BIER_BBS_LEN_TO_BUCKETS((_str)->bbs_len) - 1 ) -  \
+             (_bp / BIER_BIT_MASK_BITS_PER_BUCKET));            \
+    _bit = _bp % BIER_BIT_MASK_BITS_PER_BUCKET;                 \
+}
+
+static inline int
+bier_bit_pos_is_valid (bier_bp_t bp,  const bier_bit_string_t *bbs)
+{
+    if (!((bp <= BIER_BBS_LEN_TO_BITS((bbs)->bbs_len)) &&
+          (bp >= 1))) {
+        return (0);
+    }
+    return (1);
+}
+
+/*
+ * Validate a bit poistion
+ */
+#define BIER_BIT_POS_IS_VALID(_bp, _str)                                \
+{                                                                       \
+    if (!bier_bit_pos_is_valid(_bp, _str)) return;                      \
+}
+
+void
+bier_bit_string_set_bit (bier_bit_string_t *bit_string,
+                         bier_bp_t bp)
+{
+    bier_bit_mask_bucket_t bmask;
+    uint16_t byte_pos, bit_pos;
+
+    BIER_BIT_POS_IS_VALID(bp, bit_string);
+    BIER_GET_STRING_POS(bp, byte_pos, bit_pos, bit_string);
+
+    bmask = ((bier_bit_mask_bucket_t)1 << bit_pos);
+    bit_string->bbs_buckets[byte_pos] |= bmask;
+}
+
+void
+bier_bit_string_clear_bit (bier_bit_string_t *bit_string,
+                           bier_bp_t bp)
+{
+    uint16_t byte_pos, bit_pos;
+
+    BIER_BIT_POS_IS_VALID(bp, bit_string);
+    BIER_GET_STRING_POS(bp, byte_pos, bit_pos, bit_string);
+
+    bit_string->bbs_buckets[byte_pos] &= ~(1 << bit_pos);
+}
+
+u8 *
+format_bier_bit_string (u8 * string,
+                        va_list * args)
+{
+    bier_bit_string_t *bs = va_arg(*args, bier_bit_string_t *);
+    int leading_marker = 0;
+    int suppress_zero = 0;
+    u16 index;
+    u32 *ptr;
+
+    ptr = (u32 *)bs->bbs_buckets;
+
+    string = format(string, "%d#", (8 * bs->bbs_len));
+
+    for (index = 0; index < (bs->bbs_len/4); index++) {
+        if (!ptr[index]) {
+            if (!leading_marker) {
+                leading_marker = 1;
+                suppress_zero = 1;
+                string = format(string, ":");
+                continue;
+            }
+            if (suppress_zero) continue;
+        } else {
+            suppress_zero = 0;
+        }
+
+        string = format(string, "%s%X", index ? ":" : "",
+                        clib_net_to_host_u32(ptr[index]));
+    }
+
+    return (string);
+}
diff --git a/src/vnet/bier/bier_bit_string.h b/src/vnet/bier/bier_bit_string.h
new file mode 100644
index 00000000000..9baebfa6e69
--- /dev/null
+++ b/src/vnet/bier/bier_bit_string.h
@@ -0,0 +1,107 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef __BIER_BIT_STRING_H__
+#define __BIER_BIT_STRING_H__
+
+#include <vppinfra/byte_order.h>
+#include <vppinfra/format.h>
+
+#include <vnet/bier/bier_types.h>
+
+#define BIER_BBS_LEN_TO_BUCKETS(_len) (_len)
+#define BIER_BBS_LEN_TO_BITS(_len) (_len * 8)
+#define BIER_BBS_LEN_TO_INTS(_len) ((_len) / sizeof(int))
+#define BIER_BIT_MASK_BITS_PER_INT (sizeof(int) * 8)
+
+/*
+ * bier_find_first_bit_set
+ *
+ * find the position of the first bit set in a long
+ */
+static inline int
+bier_find_first_bit_string_set (int mask)
+{
+    return (__builtin_ffs(clib_net_to_host_u32(mask)));
+}
+
+extern void bier_bit_string_set_bit(bier_bit_string_t *mask,
+                                    bier_bp_t bp);
+
+
+extern void bier_bit_string_clear_bit(bier_bit_string_t *mask,
+                                      bier_bp_t bp);
+
+
+extern u8 *format_bier_bit_string(u8 * s, va_list * args);
+
+#define BIER_BBS_NUM_INT_BUKCETS(_bbs) \
+    (BIER_BBS_LEN_TO_BUCKETS(_bbs->bbs_len) / sizeof(int))
+
+always_inline int
+bier_bit_string_is_zero (const bier_bit_string_t *src)
+{
+    uint16_t index;
+
+    for (index = 0;
+         index < BIER_BBS_NUM_INT_BUKCETS(src);
+         index++) {
+        if (((int*)src->bbs_buckets)[index] != 0) {
+            return (0);
+        }
+    }
+    return (1);
+}
+
+always_inline void
+bier_bit_string_clear_string (const bier_bit_string_t *src,
+                              bier_bit_string_t *dest)
+{
+    uint16_t index;
+
+    ASSERT(src->bbs_len == dest->bbs_len);
+
+    for (index = 0;
+         index < BIER_BBS_NUM_INT_BUKCETS(src);
+         index++) {
+        ((int*)dest->bbs_buckets)[index] &= ~(((int*)src->bbs_buckets)[index]);
+    }
+}
+
+always_inline void
+bier_bit_string_logical_and_string (const bier_bit_string_t *src,
+                                    bier_bit_string_t *dest)
+{
+    uint16_t index;
+
+    ASSERT(src->bbs_len == dest->bbs_len);
+
+    for (index = 0;
+         index < BIER_BBS_NUM_INT_BUKCETS(src);
+         index++) {
+        ((int*)dest->bbs_buckets)[index] &= ((int*)src->bbs_buckets)[index];
+    }
+}
+
+always_inline void
+bier_bit_string_init (bier_bit_string_t *bbs,
+                      bier_hdr_len_id_t len,
+                      bier_bit_mask_bucket_t *buckets)
+{
+    bbs->bbs_len = bier_hdr_len_id_to_num_bytes(len);
+    bbs->bbs_buckets = buckets;
+}
+
+#endif
diff --git a/src/vnet/bier/bier_disp_dispatch_node.c b/src/vnet/bier/bier_disp_dispatch_node.c
new file mode 100644
index 00000000000..a00c2eea5f9
--- /dev/null
+++ b/src/vnet/bier/bier_disp_dispatch_node.c
@@ -0,0 +1,150 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/bier/bier_disp_entry.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+
+/**
+ * @brief A struct to hold tracing information for the MPLS label imposition
+ * node.
+ */
+typedef struct bier_disp_dispatch_trace_t_
+{
+    /**
+     * BIER payload protocol used to dispatch
+     */
+    bier_hdr_proto_id_t pproto;
+
+    /**
+     * RPF-ID packet is tagged with
+     */
+    u32 rpf_id;
+} bier_disp_dispatch_trace_t;
+
+always_inline uword
+bier_disp_dispatch_inline (vlib_main_t * vm,
+                         vlib_node_runtime_t * node,
+                         vlib_frame_t * from_frame)
+{
+    u32 n_left_from, next_index, * from, * to_next;
+
+    from = vlib_frame_vector_args (from_frame);
+    n_left_from = from_frame->n_vectors;
+
+    next_index = node->cached_next_index;
+
+    while (n_left_from > 0)
+    {
+        u32 n_left_to_next;
+
+        vlib_get_next_frame(vm, node, next_index, to_next, n_left_to_next);
+
+        while (n_left_from > 0 && n_left_to_next > 0)
+        {
+            bier_hdr_proto_id_t pproto0;
+            bier_disp_entry_t *bde0;
+            u32 next0, bi0, bdei0;
+            const dpo_id_t *dpo0;
+            vlib_buffer_t * b0;
+            bier_hdr_t *hdr0;
+            u32 entropy0;
+
+            bi0 = from[0];
+            to_next[0] = bi0;
+            from += 1;
+            to_next += 1;
+            n_left_from -= 1;
+            n_left_to_next -= 1;
+
+            b0 = vlib_get_buffer (vm, bi0);
+            bdei0 = vnet_buffer(b0)->ip.adj_index[VLIB_TX];
+            hdr0 = vlib_buffer_get_current(b0);
+            bde0 = bier_disp_entry_get(bdei0);
+
+            /*
+             * header is in network order - flip it, we are about to
+             * consume it anyway
+             */
+            bier_hdr_ntoh(hdr0);
+            pproto0 = bier_hdr_get_proto_id(hdr0);
+            entropy0 = bier_hdr_get_entropy(hdr0);
+
+            /*
+             * strip the header and copy the entropy value into
+             * the packets flow-hash field
+             * DSCP mumble mumble...
+             */
+            vlib_buffer_advance(b0, (vnet_buffer(b0)->bier.n_bytes +
+                                     sizeof(*hdr0)));
+            vnet_buffer(b0)->ip.flow_hash = entropy0;
+
+            /*
+             * use the payload proto to dispatch to the
+             * correct stacked DPO.
+             */
+            dpo0 = &bde0->bde_fwd[pproto0].bde_dpo;
+            next0 = dpo0->dpoi_next_node;
+            vnet_buffer(b0)->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
+            vnet_buffer(b0)->ip.rpf_id = bde0->bde_fwd[pproto0].bde_rpf_id;
+
+            if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+            {
+                bier_disp_dispatch_trace_t *tr =
+                    vlib_add_trace (vm, node, b0, sizeof (*tr));
+                tr->pproto = pproto0;
+                tr->rpf_id = vnet_buffer(b0)->ip.rpf_id;
+            }
+
+            vlib_validate_buffer_enqueue_x1(vm, node, next_index, to_next,
+                                            n_left_to_next, bi0, next0);
+        }
+        vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+    return from_frame->n_vectors;
+}
+
+static u8 *
+format_bier_disp_dispatch_trace (u8 * s, va_list * args)
+{
+    CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+    CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+    bier_disp_dispatch_trace_t * t;
+
+    t = va_arg (*args, bier_disp_dispatch_trace_t *);
+    s = format (s, "%U", format_bier_hdr_proto, t->pproto);
+
+    return (s);
+}
+
+static uword
+bier_disp_dispatch (vlib_main_t * vm,
+                  vlib_node_runtime_t * node,
+                  vlib_frame_t * frame)
+{
+    return (bier_disp_dispatch_inline(vm, node, frame));
+}
+
+VLIB_REGISTER_NODE (bier_disp_dispatch_node) = {
+    .function = bier_disp_dispatch,
+    .name = "bier-disp-dispatch",
+    .vector_size = sizeof (u32),
+
+    .format_trace = format_bier_disp_dispatch_trace,
+    .n_next_nodes = 1,
+    .next_nodes = {
+        [0] = "bier-drop",
+    }
+};
+VLIB_NODE_FUNCTION_MULTIARCH (bier_disp_dispatch_node, bier_disp_dispatch)
diff --git a/src/vnet/bier/bier_disp_entry.c b/src/vnet/bier/bier_disp_entry.c
new file mode 100644
index 00000000000..3326aba2f86
--- /dev/null
+++ b/src/vnet/bier/bier_disp_entry.c
@@ -0,0 +1,385 @@
+/*
+ * Copyright (c) 2017 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * bier_dispositon : The BIER dispositon object
+ *
+ * A BIER dispositon object is present in the IP mcast output list
+ * and represents the dispositon of a BIER bitmask. After BIER header
+ * dispositon the packet is forward within the appropriate/specifid
+ * BIER table
+ */
+
+#include <vnet/bier/bier_disp_entry.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+#include <vnet/fib/fib_path_list.h>
+#include <vnet/dpo/drop_dpo.h>
+
+/**
+ * The memory pool of all imp objects
+ */
+bier_disp_entry_t *bier_disp_entry_pool;
+
+/**
+ * When constructing the BIER imp ID from an index and BSL, shift
+ * the BSL this far
+ */
+#define BIER_DISP_ENTRY_ID_HLEN_SHIFT 24
+
+static void
+bier_disp_entry_lock_i (bier_disp_entry_t *bde)
+{
+    bde->bde_locks++;
+}
+
+void
+bier_disp_entry_lock (index_t bdei)
+{
+    bier_disp_entry_lock_i(bier_disp_entry_get(bdei));
+}
+
+static index_t
+bier_disp_entry_get_index(bier_disp_entry_t *bde)
+{
+    return (bde - bier_disp_entry_pool);
+}
+
+index_t
+bier_disp_entry_add_or_lock (void)
+{
+    dpo_id_t invalid = DPO_INVALID;
+    bier_hdr_proto_id_t pproto;
+    bier_disp_entry_t *bde;
+
+    pool_get_aligned(bier_disp_entry_pool, bde, CLIB_CACHE_LINE_BYTES);
+
+    bde->bde_locks = 0;
+
+    FOR_EACH_BIER_HDR_PROTO(pproto)
+    {
+        bde->bde_fwd[pproto].bde_dpo = invalid;
+        bde->bde_fwd[pproto].bde_rpf_id = ~0;
+        bde->bde_pl[pproto] = FIB_NODE_INDEX_INVALID;
+    }
+
+    bier_disp_entry_lock_i(bde);
+    return (bier_disp_entry_get_index(bde));
+}
+
+void
+bier_disp_entry_unlock (index_t bdei)
+{
+    bier_disp_entry_t *bde;
+
+    if (INDEX_INVALID == bdei)
+    {
+        return;
+    }
+
+    bde = bier_disp_entry_get(bdei);
+
+    bde->bde_locks--;
+
+    if (0 == bde->bde_locks)
+    {
+        bier_hdr_proto_id_t pproto;
+
+        FOR_EACH_BIER_HDR_PROTO(pproto)
+        {
+            dpo_unlock(&bde->bde_fwd[pproto].bde_dpo);
+            bde->bde_fwd[pproto].bde_rpf_id = ~0;
+            fib_path_list_unlock(bde->bde_pl[pproto]);
+        }
+        pool_put(bier_disp_entry_pool, bde);
+    }
+}
+
+typedef struct bier_disp_entry_path_list_walk_ctx_t_
+{
+    u32 bdew_rpf_id;
+} bier_disp_entry_path_list_walk_ctx_t;
+
+static fib_path_list_walk_rc_t
+bier_disp_entry_path_list_walk (fib_node_index_t pl_index,
+                                fib_node_index_t path_index,
+                                void *arg)
+{
+    bier_disp_entry_path_list_walk_ctx_t *ctx = arg;
+
+    ctx->bdew_rpf_id = fib_path_get_rpf_id(path_index);
+
+    if (~0 != ctx->bdew_rpf_id)
+    {
+        return (FIB_PATH_LIST_WALK_STOP);
+    }
+    return (FIB_PATH_LIST_WALK_CONTINUE);
+}
+
+static void
+bier_disp_entry_restack (bier_disp_entry_t *bde,
+                         bier_hdr_proto_id_t pproto)
+{
+    dpo_id_t via_dpo = DPO_INVALID;
+    fib_node_index_t pli;
+
+    pli = bde->bde_pl[pproto];
+
+    if (FIB_NODE_INDEX_INVALID == pli)
+    {
+        dpo_copy(&via_dpo,
+                 drop_dpo_get(bier_hdr_proto_to_dpo(pproto)));
+    }
+    else
+    {
+        fib_path_list_contribute_forwarding(pli,
+                                            fib_forw_chain_type_from_dpo_proto(
+                                                bier_hdr_proto_to_dpo(pproto)),
+                                            &via_dpo);
+
+        bier_disp_entry_path_list_walk_ctx_t ctx = {
+            .bdew_rpf_id = ~0,
+        };
+
+        fib_path_list_walk(pli, bier_disp_entry_path_list_walk, &ctx);
+        bde->bde_fwd[pproto].bde_rpf_id = ctx.bdew_rpf_id;
+    }
+
+    dpo_stack(DPO_BIER_DISP_ENTRY,
+              DPO_PROTO_BIER,
+              &bde->bde_fwd[pproto].bde_dpo,
+              &via_dpo);
+}
+
+void
+bier_disp_entry_path_add (index_t bdei,
+                          bier_hdr_proto_id_t pproto,
+                          const fib_route_path_t *rpaths)
+{
+    fib_node_index_t *pli, old_pli;
+    bier_disp_entry_t *bde;
+
+    bde = bier_disp_entry_get(bdei);
+    pli = &bde->bde_pl[pproto];
+    old_pli = *pli;
+
+    /*
+     * create a new or update the exisitng path-list for this
+     * payload protocol
+     */
+    if (FIB_NODE_INDEX_INVALID == *pli)
+    {
+        *pli = fib_path_list_create((FIB_PATH_LIST_FLAG_SHARED |
+                                     FIB_PATH_LIST_FLAG_NO_URPF),
+                                    rpaths);
+    }
+    else
+    {
+        *pli = fib_path_list_copy_and_path_add(old_pli,
+                                               (FIB_PATH_LIST_FLAG_SHARED |
+                                                FIB_PATH_LIST_FLAG_NO_URPF),
+                                               rpaths);
+    }
+
+    fib_path_list_lock(*pli);
+    fib_path_list_unlock(old_pli);
+
+    bier_disp_entry_restack(bde, pproto);
+}
+
+int
+bier_disp_entry_path_remove (index_t bdei,
+                             bier_hdr_proto_id_t pproto,
+                             const fib_route_path_t *rpaths)
+{
+    fib_node_index_t *pli, old_pli;
+    bier_disp_entry_t *bde;
+
+    bde = bier_disp_entry_get(bdei);
+    pli = &bde->bde_pl[pproto];
+    old_pli = *pli;
+
+    /*
+     * update the exisitng path-list for this payload protocol
+     */
+    if (FIB_NODE_INDEX_INVALID != *pli)
+    {
+        *pli = fib_path_list_copy_and_path_remove(old_pli,
+                                                  (FIB_PATH_LIST_FLAG_SHARED |
+                                                   FIB_PATH_LIST_FLAG_NO_URPF),
+                                                  rpaths);
+
+        fib_path_list_lock(*pli);
+        fib_path_list_unlock(old_pli);
+
+        bier_disp_entry_restack(bde, pproto);
+    }
+
+    /*
+     * if there are no path-list defined for any payload protocol
+     * then this entry is OK for removal
+     */
+    int remove = 1;
+
+    FOR_EACH_BIER_HDR_PROTO(pproto)
+    {
+        if (FIB_NODE_INDEX_INVALID != bde->bde_pl[pproto])
+        {
+            remove = 0;
+            break;
+        }
+    }
+
+    return (remove);
+}
+
+u8*
+format_bier_disp_entry (u8* s, va_list *args)
+{
+    index_t bdei = va_arg (*args, index_t);
+    u32 indent = va_arg(*args, u32);
+    bier_show_flags_t flags = va_arg(*args, bier_show_flags_t);
+    bier_hdr_proto_id_t pproto;
+    bier_disp_entry_t *bde;
+
+    bde = bier_disp_entry_get(bdei);
+
+    s = format(s, "bier-disp:[%d]", bdei);
+
+    FOR_EACH_BIER_HDR_PROTO(pproto)
+    {
+        if (INDEX_INVALID != bde->bde_pl[pproto])
+        {
+            s = format(s, "\n");
+            s = fib_path_list_format(bde->bde_pl[pproto], s);
+
+            if (flags & BIER_SHOW_DETAIL)
+            {
+                s = format(s, "\n%UForwarding:",
+                           format_white_space, indent);
+                s = format(s, "\n%Urpf-id:%d",
+                           format_white_space, indent+1,
+                           bde->bde_fwd[pproto].bde_rpf_id);
+                s = format(s, "\n%U%U",
+                           format_white_space, indent+1,
+                           format_dpo_id, &bde->bde_fwd[pproto].bde_dpo, indent+2);
+            }
+        }
+    }
+    return (s);
+}
+
+void
+bier_disp_entry_contribute_forwarding (index_t bdei,
+                                       dpo_id_t *dpo)
+{
+    dpo_set(dpo, DPO_BIER_DISP_ENTRY, DPO_PROTO_BIER, bdei);
+}
+
+const static char* const bier_disp_entry_bier_nodes[] =
+{
+    "bier-disp-dispatch",
+    NULL,
+};
+
+const static char* const * const bier_disp_entry_nodes[DPO_PROTO_NUM] =
+{
+    [DPO_PROTO_BIER]  = bier_disp_entry_bier_nodes,
+};
+
+static void
+bier_disp_entry_dpo_lock (dpo_id_t *dpo)
+{
+    bier_disp_entry_lock(dpo->dpoi_index);
+}
+
+static void
+bier_disp_entry_dpo_unlock (dpo_id_t *dpo)
+{
+    bier_disp_entry_unlock(dpo->dpoi_index);
+}
+
+static void
+bier_disp_entry_dpo_mem_show (void)
+{
+    fib_show_memory_usage("BIER dispositon",
+                          pool_elts(bier_disp_entry_pool),
+                          pool_len(bier_disp_entry_pool),
+                          sizeof(bier_disp_entry_t));
+}
+
+static u8*
+format_bier_disp_entry_dpo (u8* s, va_list *ap)
+{
+    index_t index = va_arg(*ap, index_t);
+    u32 indent = va_arg(*ap, u32);
+
+    s = format(s, "%U", format_bier_disp_entry, index, indent, BIER_SHOW_DETAIL);
+
+    return (s);
+}
+
+const static dpo_vft_t bier_disp_entry_vft = {
+    .dv_lock = bier_disp_entry_dpo_lock,
+    .dv_unlock = bier_disp_entry_dpo_unlock,
+    .dv_format = format_bier_disp_entry_dpo,
+    .dv_mem_show = bier_disp_entry_dpo_mem_show,
+};
+
+clib_error_t *
+bier_disp_entry_db_module_init (vlib_main_t *vm)
+{
+    dpo_register(DPO_BIER_DISP_ENTRY,
+                 &bier_disp_entry_vft,
+                 bier_disp_entry_nodes);
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_disp_entry_db_module_init);
+
+static clib_error_t *
+show_bier_disp_entry (vlib_main_t * vm,
+               unformat_input_t * input,
+               vlib_cli_command_t * cmd)
+{
+    index_t bdei;
+
+    bdei = INDEX_INVALID;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "%d", &bdei))
+            ;
+        else
+        {
+            break;
+        }
+    }
+
+    if (INDEX_INVALID == bdei)
+    {
+        return (NULL);
+    }
+    else
+    {
+        vlib_cli_output(vm, "%U", format_bier_disp_entry, bdei, 1,
+                        BIER_SHOW_DETAIL);
+    }
+    return (NULL);
+}
+
+VLIB_CLI_COMMAND (show_bier_disp_entry_node, static) = {
+    .path = "show bier disp entry",
+    .short_help = "show bier disp entry index",
+    .function = show_bier_disp_entry,
+};
diff --git a/src/vnet/bier/bier_disp_entry.h b/src/vnet/bier/bier_disp_entry.h
new file mode 100644
index 00000000000..34ca5d4a09a
--- /dev/null
+++ b/src/vnet/bier/bier_disp_entry.h
@@ -0,0 +1,81 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * bier_dispositon : The BIER dispositon object
+ *
+ * A BIER dispositon object is used to pop the BIER header for for-us
+ * packets and steer the packet down the payload protocol specific graph
+ */
+
+#ifndef __BIER_DISP_ENTRY_H__
+#define __BIER_DISP_ENTRY_H__
+
+#include <vnet/bier/bier_types.h>
+#include <vnet/fib/fib_types.h>
+#include <vnet/dpo/dpo.h>
+
+/**
+ * The BIER dispositon object
+ */
+typedef struct bier_disp_entry_t_ {
+    /**
+     * The DPO contirubted from the per-payload protocol parents
+     * on cachline 1.
+     */
+    struct
+    {
+        dpo_id_t bde_dpo;
+        u32 bde_rpf_id;
+    } bde_fwd[BIER_HDR_N_PROTO];
+
+    /**
+     * number of locks
+     */
+    u32 bde_locks;
+
+    /**
+     * The path-lists used by per-payload protocol parents.
+     * We don't add the disp entry to the graph as a sibling
+     * since there is nothing we can do with the updates to
+     * forwarding.
+     */
+    fib_node_index_t bde_pl[BIER_HDR_N_PROTO];
+} bier_disp_entry_t;
+
+extern index_t bier_disp_entry_add_or_lock(void);
+extern void bier_disp_entry_path_add(index_t bdei,
+                                     bier_hdr_proto_id_t pproto,
+                                     const fib_route_path_t *rpaths);
+extern int bier_disp_entry_path_remove(index_t bdei,
+                                       bier_hdr_proto_id_t pproto,
+                                       const fib_route_path_t *rpaths);
+
+extern void bier_disp_entry_unlock(index_t bdi);
+extern void bier_disp_entry_lock(index_t bdi);
+
+extern u8* format_bier_disp_entry(u8* s, va_list *ap);
+
+extern void bier_disp_entry_contribute_forwarding(index_t bdi,
+                                                  dpo_id_t *dpo);
+
+extern bier_disp_entry_t *bier_disp_entry_pool;
+
+always_inline bier_disp_entry_t*
+bier_disp_entry_get (index_t bdi)
+{
+    return (pool_elt_at_index(bier_disp_entry_pool, bdi));
+}
+
+#endif
diff --git a/src/vnet/bier/bier_disp_lookup_node.c b/src/vnet/bier/bier_disp_lookup_node.c
new file mode 100644
index 00000000000..15515f467ca
--- /dev/null
+++ b/src/vnet/bier/bier_disp_lookup_node.c
@@ -0,0 +1,147 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/bier/bier_disp_table.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+
+/**
+ * @brief A struct to hold tracing information for the MPLS label imposition
+ * node.
+ */
+typedef struct bier_disp_lookup_trace_t_
+{
+    /**
+     * BIER source BP used in the lookup - host order
+     */
+    bier_bp_t bp;
+    /**
+     * BIER disp table
+     */
+    index_t bdti;
+} bier_disp_lookup_trace_t;
+
+/**
+ * Next nodes from BIER disposition lookup
+ */
+typedef enum bier_disp_lookup_next_t_
+{
+    BIER_DISP_LOOKUP_NEXT_DROP,
+    BIER_DISP_LOOKUP_NEXT_DISPATCH,
+} bier_disp_lookup_next_t;
+#define BIER_DISP_LOOKUP_N_NEXT (BIER_DISP_LOOKUP_NEXT_DISPATCH+1)
+
+always_inline uword
+bier_disp_lookup_inline (vlib_main_t * vm,
+                         vlib_node_runtime_t * node,
+                         vlib_frame_t * from_frame)
+{
+    u32 n_left_from, next_index, * from, * to_next;
+
+    from = vlib_frame_vector_args (from_frame);
+    n_left_from = from_frame->n_vectors;
+
+    next_index = node->cached_next_index;
+
+    while (n_left_from > 0)
+    {
+        u32 n_left_to_next;
+
+        vlib_get_next_frame(vm, node, next_index, to_next, n_left_to_next);
+
+        while (n_left_from > 0 && n_left_to_next > 0)
+        {
+            const bier_hdr_t *hdr0;
+            vlib_buffer_t * b0;
+            u32 bdei0, bdti0;
+            u32 next0, bi0;
+
+            bi0 = from[0];
+            to_next[0] = bi0;
+            from += 1;
+            to_next += 1;
+            n_left_from -= 1;
+            n_left_to_next -= 1;
+
+            b0 = vlib_get_buffer (vm, bi0);
+
+            bdti0 = vnet_buffer(b0)->ip.adj_index[VLIB_TX];
+            hdr0 = vlib_buffer_get_current(b0);
+
+            /*
+             * lookup - source is in network order.
+             */
+            bdei0 = bier_disp_table_lookup(bdti0, bier_hdr_get_src_id(hdr0));
+
+            if (PREDICT_FALSE(INDEX_INVALID == bdei0))
+            {
+                next0 = BIER_DISP_LOOKUP_NEXT_DROP;
+            }
+            else
+            {
+                next0 = BIER_DISP_LOOKUP_NEXT_DISPATCH;
+            }
+
+            vnet_buffer(b0)->ip.adj_index[VLIB_TX] = bdei0;
+
+            if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+            {
+                bier_disp_lookup_trace_t *tr =
+                    vlib_add_trace (vm, node, b0, sizeof (*tr));
+                tr->bp = clib_net_to_host_u16(bier_hdr_get_src_id(hdr0));
+                tr->bdti = bdti0;
+            }
+
+            vlib_validate_buffer_enqueue_x1(vm, node, next_index, to_next,
+                                            n_left_to_next, bi0, next0);
+        }
+        vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+    return from_frame->n_vectors;
+}
+
+static u8 *
+format_bier_disp_lookup_trace (u8 * s, va_list * args)
+{
+    CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+    CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+    bier_disp_lookup_trace_t * t;
+
+    t = va_arg (*args, bier_disp_lookup_trace_t *);
+    s = format (s, "tbl:%d src:%d", t->bdti, t->bp);
+
+    return (s);
+}
+
+static uword
+bier_disp_lookup (vlib_main_t * vm,
+                  vlib_node_runtime_t * node,
+                  vlib_frame_t * frame)
+{
+    return (bier_disp_lookup_inline(vm, node, frame));
+}
+
+VLIB_REGISTER_NODE (bier_disp_lookup_node) = {
+    .function = bier_disp_lookup,
+    .name = "bier-disp-lookup",
+    .vector_size = sizeof (u32),
+
+    .format_trace = format_bier_disp_lookup_trace,
+    .n_next_nodes = BIER_DISP_LOOKUP_N_NEXT,
+    .next_nodes = {
+        [BIER_DISP_LOOKUP_NEXT_DROP] = "bier-drop",
+        [BIER_DISP_LOOKUP_NEXT_DISPATCH] = "bier-disp-dispatch",
+    }
+};
+VLIB_NODE_FUNCTION_MULTIARCH (bier_disp_lookup_node, bier_disp_lookup)
diff --git a/src/vnet/bier/bier_disp_table.c b/src/vnet/bier/bier_disp_table.c
new file mode 100644
index 00000000000..5caf86d3705
--- /dev/null
+++ b/src/vnet/bier/bier_disp_table.c
@@ -0,0 +1,396 @@
+/*
+ * Copyright (c) 2017 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/bier/bier_disp_table.h>
+#include <vnet/bier/bier_disp_entry.h>
+
+/**
+ * memory pool for disposition tables
+ */
+bier_disp_table_t *bier_disp_table_pool;
+
+/**
+ * Hash table to map client table IDs to VPP index
+ */
+static uword *bier_disp_table_id_to_index;
+
+static index_t
+bier_disp_table_get_index (const bier_disp_table_t *bdt)
+{
+    return (bdt - bier_disp_table_pool);
+}
+
+static void
+bier_disp_table_lock_i (bier_disp_table_t *bdt)
+{
+    bdt->bdt_locks++;
+}
+
+index_t
+bier_disp_table_find(u32 table_id)
+{
+    uword *p;
+
+    p = hash_get(bier_disp_table_id_to_index, table_id);
+
+    if (NULL != p)
+    {
+        return (p[0]);
+    }
+
+    return (INDEX_INVALID);
+}
+
+index_t
+bier_disp_table_add_or_lock (u32 table_id)
+{
+    bier_disp_table_t *bdt;
+    index_t bdti;
+
+    bdti = bier_disp_table_find(table_id);
+
+    if (INDEX_INVALID == bdti)
+    {
+        pool_get_aligned(bier_disp_table_pool, bdt,
+                         CLIB_CACHE_LINE_BYTES);
+
+        bdt->bdt_table_id = table_id;
+        bdt->bdt_locks = 0;
+
+        hash_set(bier_disp_table_id_to_index, table_id,
+                 bier_disp_table_get_index(bdt));
+
+        /**
+         * Set the result for each entry in the DB to be invalid
+         */
+        memset(bdt->bdt_db, 0xff, sizeof(bdt->bdt_db));
+    }
+    else
+    {
+        bdt = pool_elt_at_index(bier_disp_table_pool, bdti);
+    }
+
+    bier_disp_table_lock_i(bdt);
+
+    return (bier_disp_table_get_index(bdt));
+}
+
+void
+bier_disp_table_unlock_w_table_id (u32 table_id)
+{
+    index_t bdti;
+
+    bdti = bier_disp_table_find(table_id);
+
+    if (INDEX_INVALID != bdti)
+    {
+        bier_disp_table_unlock(bdti);
+    }
+}
+
+void
+bier_disp_table_unlock (index_t bdti)
+{
+    bier_disp_table_t *bdt;
+
+    bdt = bier_disp_table_get(bdti);
+
+    bdt->bdt_locks--;
+
+    if (0 == bdt->bdt_locks)
+    {
+        u32 ii;
+
+        for (ii = 0; ii < BIER_BP_MAX; ii++)
+        {
+            bier_disp_entry_unlock(bdt->bdt_db[ii]);
+        }
+        hash_unset(bier_disp_table_id_to_index, bdt->bdt_table_id);
+        pool_put(bier_disp_table_pool, bdt);
+    }
+}
+
+void
+bier_disp_table_lock (index_t bdti)
+{
+    bier_disp_table_lock_i(bier_disp_table_get(bdti));
+}
+
+void
+bier_disp_table_contribute_forwarding (index_t bdti,
+                                       dpo_id_t *dpo)
+{
+    dpo_set(dpo,
+            DPO_BIER_DISP_TABLE,
+            DPO_PROTO_BIER,
+            bdti);
+}
+
+
+u8*
+format_bier_disp_table (u8* s, va_list *ap)
+{
+    index_t bdti = va_arg(*ap, index_t);
+    u32 indent = va_arg(*ap, u32);
+    bier_show_flags_t flags = va_arg(*ap, bier_show_flags_t);
+    bier_disp_table_t *bdt;
+
+    bdt = bier_disp_table_get(bdti);
+
+    s = format(s, "bier-disp-table:[%d]; table-id:%d locks:%d",
+               bdti, bdt->bdt_table_id, bdt->bdt_locks);
+
+    if (flags & BIER_SHOW_DETAIL)
+    {
+        u32 ii;
+
+        for (ii = 0; ii < BIER_BP_MAX; ii++)
+        {
+            if (INDEX_INVALID != bdt->bdt_db[ii])
+            {
+                u16 src = ii;
+                s = format(s, "\n%Usrc:%d", format_white_space, indent,
+                           clib_host_to_net_u16(src));
+                s = format(s, "\n%U%U", format_white_space, indent+2,
+                           format_bier_disp_entry, bdt->bdt_db[ii],
+                           indent+4, BIER_SHOW_BRIEF);
+            }
+        }
+    }
+    return (s);
+}
+
+static u8*
+format_bier_disp_table_dpo (u8* s, va_list *ap)
+{
+    index_t bdti = va_arg(*ap, index_t);
+    u32 indent = va_arg(*ap, u32);
+
+    return (format(s, "%U",
+                   format_bier_disp_table, bdti, indent,
+                   BIER_SHOW_BRIEF));
+}
+
+static void
+bier_disp_table_entry_insert (index_t bdti,
+                              bier_bp_t src,
+                              index_t bdei)
+{
+    bier_disp_table_t *bdt;
+
+    bdt = bier_disp_table_get(bdti);
+    bdt->bdt_db[clib_host_to_net_u16(src)] = bdei;
+}
+
+static void
+bier_disp_table_entry_remove (index_t bdti,
+                              bier_bp_t src)
+{
+    bier_disp_table_t *bdt;
+
+    bdt = bier_disp_table_get(bdti);
+    bdt->bdt_db[clib_host_to_net_u16(src)] = INDEX_INVALID;
+}
+
+static index_t
+bier_disp_table_lookup_hton(index_t bdti,
+                            bier_bp_t src)
+{
+    return (bier_disp_table_lookup(bdti, clib_host_to_net_u16(src)));
+}
+
+void
+bier_disp_table_entry_path_add (u32 table_id,
+                                bier_bp_t src,
+                                bier_hdr_proto_id_t payload_proto,
+                                const fib_route_path_t *rpaths)
+{
+    index_t bdti, bdei;
+
+    bdti = bier_disp_table_find(table_id);
+
+    if (INDEX_INVALID == bdti)
+    {
+        return;
+    }
+
+    bdei = bier_disp_table_lookup_hton(bdti, src);
+
+    if (INDEX_INVALID == bdei)
+    {
+        bdei = bier_disp_entry_add_or_lock();
+        bier_disp_table_entry_insert(bdti, src, bdei);
+    }
+
+    bier_disp_entry_path_add(bdei, payload_proto, rpaths);
+}
+
+void
+bier_disp_table_entry_path_remove (u32 table_id,
+                                   bier_bp_t src,
+                                   bier_hdr_proto_id_t payload_proto,
+                                   const fib_route_path_t *rpath)
+{
+    index_t bdti, bdei;
+
+    bdti = bier_disp_table_find(table_id);
+
+    if (INDEX_INVALID == bdti)
+    {
+        return;
+    }
+
+    bdei = bier_disp_table_lookup_hton(bdti, src);
+
+    if (INDEX_INVALID != bdei)
+    {
+        int remove;
+
+        remove = bier_disp_entry_path_remove(bdei, payload_proto, rpath);
+
+        if (remove)
+        {
+            bier_disp_table_entry_remove(bdti, src);
+            bier_disp_entry_unlock(bdei);
+        }
+    }
+}
+
+void
+bier_disp_table_walk (u32 table_id,
+                      bier_disp_table_walk_fn_t fn,
+                      void *ctx)
+{
+    const bier_disp_table_t *bdt;
+    const bier_disp_entry_t *bde;
+    index_t bdti;
+    u32 ii;
+
+    bdti = bier_disp_table_find(table_id);
+
+    if (INDEX_INVALID != bdti)
+    {
+        bdt = bier_disp_table_get(bdti);
+
+        for (ii = 0; ii < BIER_BP_MAX; ii++)
+        {
+            if (INDEX_INVALID != bdt->bdt_db[ii])
+            {
+                u16 src = ii;
+
+                bde = bier_disp_entry_get(bdt->bdt_db[ii]);
+
+                fn(bdt, bde, clib_host_to_net_u16(src), ctx);
+            }
+        }
+    }
+}
+
+static void
+bier_disp_table_dpo_lock (dpo_id_t *dpo)
+{
+    bier_disp_table_lock(dpo->dpoi_index);
+}
+
+static void
+bier_disp_table_dpo_unlock (dpo_id_t *dpo)
+{
+    bier_disp_table_unlock(dpo->dpoi_index);
+}
+
+static void
+bier_disp_table_dpo_mem_show (void)
+{
+    fib_show_memory_usage("BIER disposition table",
+                          pool_elts(bier_disp_table_pool),
+                          pool_len(bier_disp_table_pool),
+                          sizeof(bier_disp_table_t));
+}
+
+const static dpo_vft_t bier_disp_table_dpo_vft = {
+    .dv_lock = bier_disp_table_dpo_lock,
+    .dv_unlock = bier_disp_table_dpo_unlock,
+    .dv_mem_show = bier_disp_table_dpo_mem_show,
+    .dv_format = format_bier_disp_table_dpo,
+};
+
+const static char *const bier_disp_table_bier_nodes[] =
+{
+    "bier-disp-lookup"
+};
+
+const static char * const * const bier_disp_table_nodes[DPO_PROTO_NUM] =
+{
+    [DPO_PROTO_BIER] = bier_disp_table_bier_nodes,
+};
+
+clib_error_t *
+bier_disp_table_module_init (vlib_main_t *vm)
+{
+    dpo_register(DPO_BIER_DISP_TABLE,
+                 &bier_disp_table_dpo_vft,
+                 bier_disp_table_nodes);
+
+    bier_disp_table_id_to_index = hash_create(0, sizeof(index_t));
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_disp_table_module_init);
+
+static clib_error_t *
+show_bier_disp_table (vlib_main_t * vm,
+                      unformat_input_t * input,
+                      vlib_cli_command_t * cmd)
+{
+    bier_disp_table_t *bdt;
+    index_t bdti;
+
+    bdti = INDEX_INVALID;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "%d", &bdti))
+            ;
+        else if (unformat (input, "%d", &bdti))
+            ;
+        else
+        {
+            break;
+        }
+    }
+
+    if (INDEX_INVALID == bdti)
+    {
+        pool_foreach(bdt, bier_disp_table_pool,
+        ({
+            vlib_cli_output(vm, "%U", format_bier_disp_table,
+                            bier_disp_table_get_index(bdt),
+                            1,
+                            BIER_SHOW_BRIEF);
+        }));
+    }
+    else
+    {
+        vlib_cli_output(vm, "%U", format_bier_disp_table, bdti, 1,
+                        BIER_SHOW_DETAIL);
+    }
+    return (NULL);
+}
+
+VLIB_CLI_COMMAND (show_bier_disp_table_node, static) = {
+    .path = "show bier disp table",
+    .short_help = "show bier disp table [index]",
+    .function = show_bier_disp_table,
+};
diff --git a/src/vnet/bier/bier_disp_table.h b/src/vnet/bier/bier_disp_table.h
new file mode 100644
index 00000000000..18726c8ee6a
--- /dev/null
+++ b/src/vnet/bier/bier_disp_table.h
@@ -0,0 +1,110 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef __BIER_DISP_TABLE_H__
+#define __BIER_DISP_TABLE_H__
+
+#include <vnet/ip/ip.h>
+#include <vnet/adj/adj.h>
+#include <vnet/dpo/replicate_dpo.h>
+
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_disp_entry.h>
+
+/**
+ * @brief
+ *   A protocol Independent IP multicast FIB table
+ */
+typedef struct bier_disp_table_t_
+{
+    /**
+     * number of locks on the table
+     */
+    u16 bdt_locks;
+
+    /**
+     * Table ID (hash key) for this FIB.
+     */
+    u32 bdt_table_id;
+
+    /**
+     * The lookup DB based on sender BP. Value is the index of the
+     * BIER disp object.
+     */
+    index_t bdt_db[BIER_BP_MAX];
+} bier_disp_table_t;
+
+/**
+ * @brief
+ *  Format the description/name of the table
+ */
+extern u8* format_bier_disp_table(u8* s, va_list *ap);
+
+extern void bier_disp_table_entry_path_add(u32 table_id,
+                                           bier_bp_t src,
+                                           bier_hdr_proto_id_t payload_proto,
+                                           const fib_route_path_t *rpath);
+
+extern void bier_disp_table_entry_path_remove(u32 table_id,
+                                              bier_bp_t src,
+                                              bier_hdr_proto_id_t payload_proto,
+                                              const fib_route_path_t *paths);
+
+extern index_t bier_disp_table_find(u32 table_id);
+
+
+extern index_t bier_disp_table_add_or_lock(u32 table_id);
+extern void bier_disp_table_unlock_w_table_id(u32 table_id);
+
+extern void bier_disp_table_unlock(index_t bdti);
+extern void bier_disp_table_lock(index_t bdti);
+extern void bier_disp_table_contribute_forwarding(index_t bdti,
+                                                  dpo_id_t *dpo);
+
+/**
+ * Types and functions to walk all the entries in one BIER Table
+ */
+typedef void (*bier_disp_table_walk_fn_t)(const bier_disp_table_t *bdt,
+                                          const bier_disp_entry_t *bde,
+                                          u16 bp,
+                                          void *ctx);
+extern void bier_disp_table_walk(u32 table_id,
+                                 bier_disp_table_walk_fn_t fn,
+                                 void *ctx);
+
+/**
+ * @brief
+ * Get a pointer to a FIB table
+ */
+extern bier_disp_table_t *bier_disp_table_pool;
+
+static inline bier_disp_table_t *
+bier_disp_table_get (index_t bdti)
+{
+    return (pool_elt_at_index(bier_disp_table_pool, bdti));
+}
+
+static inline index_t
+bier_disp_table_lookup (index_t bdti,
+                        bier_bp_t src)
+{
+    bier_disp_table_t *bdt;
+
+    bdt = bier_disp_table_get(bdti);
+
+    return (bdt->bdt_db[src]);
+}
+
+#endif
diff --git a/src/vnet/bier/bier_drop.c b/src/vnet/bier/bier_drop.c
new file mode 100644
index 00000000000..eb873acf7ba
--- /dev/null
+++ b/src/vnet/bier/bier_drop.c
@@ -0,0 +1,101 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/buffer.h>
+#include <vlib/vlib.h>
+#include <vnet/dpo/dpo.h>
+
+#include <vnet/bier/bier_hdr_inlines.h>
+
+typedef struct bier_drop_trace_t_
+{
+    index_t dpi;
+} bier_drop_trace_t;
+
+static void
+bier_drop_trace (vlib_main_t * vm,
+                 vlib_node_runtime_t * node,
+                 vlib_frame_t * frame)
+{
+  u32 *from, n_left;
+
+  n_left = frame->n_vectors;
+  from = vlib_frame_vector_args (frame);
+  while (n_left >= 1)
+    {
+      u32 bi0;
+      vlib_buffer_t *b0;
+      bier_drop_trace_t *t0;
+
+      bi0 = from[0];
+
+      b0 = vlib_get_buffer (vm, bi0);
+
+      if (b0->flags & VLIB_BUFFER_IS_TRACED)
+      {
+          t0 = vlib_add_trace (vm, node, b0, sizeof(*t0));
+
+          t0->dpi = vnet_buffer (b0)->ip.adj_index[VLIB_TX];
+      }
+      from += 1;
+      n_left -= 1;
+    }
+}
+
+static uword
+bier_drop (vlib_main_t * vm,
+           vlib_node_runtime_t * node,
+           vlib_frame_t * frame)
+{
+    u32 *buffers = vlib_frame_vector_args (frame);
+    uword n_packets = frame->n_vectors;
+
+    vlib_error_drop_buffers (vm, node, buffers,
+                             /* stride */ 1,
+                             n_packets,
+                             /* next */ 0,
+                             0, // bier_input_node.index,
+                             0);
+
+    if (node->flags & VLIB_NODE_FLAG_TRACE)
+        bier_drop_trace (vm, node, frame);
+
+    return n_packets;
+}
+
+static u8 *
+format_bier_drop_trace (u8 * s, va_list * args)
+{
+  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+  bier_drop_trace_t *t = va_arg (*args, bier_drop_trace_t *);
+
+  s = format (s, "dpo-idx %d", t->dpi);
+
+  return s;
+}
+
+/* *INDENT-OFF* */
+VLIB_REGISTER_NODE (bier_drop_node, static) =
+{
+    .function = bier_drop,.
+    name = "bier-drop",
+    .vector_size = sizeof (u32),
+    .format_trace = format_bier_drop_trace,
+    .n_next_nodes = 1,
+    .next_nodes = {
+        [0] = "error-drop",
+    },
+};
diff --git a/src/vnet/bier/bier_entry.c b/src/vnet/bier/bier_entry.c
new file mode 100644
index 00000000000..88be8125f51
--- /dev/null
+++ b/src/vnet/bier/bier_entry.c
@@ -0,0 +1,388 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/bier/bier_entry.h>
+#include <vnet/bier/bier_update.h>
+
+#include <vnet/fib/fib_path_list.h>
+
+#include <vnet/bier/bier_fmask_db.h>
+#include <vnet/bier/bier_fmask.h>
+#include <vnet/bier/bier_table.h>
+
+bier_entry_t *bier_entry_pool;
+
+static index_t
+bier_entry_get_index (const bier_entry_t *be)
+{
+    return (be - bier_entry_pool);
+}
+
+static fib_path_list_walk_rc_t
+bier_entry_link_walk (fib_node_index_t pl_index,
+                      fib_node_index_t path_index,
+                      void *arg)
+{
+    bier_entry_t *be = arg;
+    index_t bfmi;
+
+    bfmi = fib_path_get_resolving_index(path_index);
+    bier_fmask_link(bfmi, be->be_bp);
+
+    return (FIB_PATH_LIST_WALK_CONTINUE);
+}
+
+static fib_path_list_walk_rc_t
+bier_entry_unlink_walk (fib_node_index_t pl_index,
+                        fib_node_index_t path_index,
+                        void *arg)
+{
+    bier_entry_t *be = arg;
+    index_t bfmi;
+
+    bfmi = fib_path_get_resolving_index(path_index);
+    bier_fmask_unlink(bfmi, be->be_bp);
+
+    return (FIB_PATH_LIST_WALK_CONTINUE);
+}
+
+index_t
+bier_entry_create (index_t bti,
+                   bier_bp_t bp)
+{
+    bier_entry_t *be;
+
+    pool_get(bier_entry_pool, be);
+
+    be->be_bp = bp;
+    be->be_bti = bti;
+    be->be_path_list = FIB_NODE_INDEX_INVALID;
+
+    return (bier_entry_get_index(be));
+}
+
+void
+bier_entry_delete (index_t bei)
+{
+    bier_entry_t *be;
+
+    be = bier_entry_get(bei);
+
+    /*
+     * if we still ahve a path-list, unlink from it
+     */
+    if (FIB_NODE_INDEX_INVALID != be->be_path_list)
+    {
+        fib_path_list_walk(be->be_path_list,
+                           bier_entry_unlink_walk,
+                           be);
+        fib_path_list_child_remove(be->be_path_list,
+                                   be->be_sibling_index);
+    }
+
+    pool_put(bier_entry_pool, be);
+}
+
+static void
+bier_entry_table_ecmp_walk_add_fmask (index_t btei,
+                                      void *arg)
+{
+    bier_entry_t *be = arg;;
+
+    /*
+     * choose a fmask from the entry's resolved set to add
+     * to ECMP table's lookup table
+     */
+    if (FIB_NODE_INDEX_INVALID != be->be_path_list)
+    {
+        const bier_table_id_t *btid;
+        dpo_id_t dpo = DPO_INVALID;
+        const dpo_id_t *choice;
+        load_balance_t *lb;
+
+        btid = bier_table_get_id(btei);
+
+        fib_path_list_contribute_forwarding(be->be_path_list,
+                                            FIB_FORW_CHAIN_TYPE_BIER,
+                                            &dpo);
+
+        /*
+         * select the appropriate bucket from the LB
+         */
+        ASSERT(dpo.dpoi_type == DPO_LOAD_BALANCE);
+
+        lb = load_balance_get(dpo.dpoi_index);
+
+        choice = load_balance_get_bucket_i(lb,
+                                           btid->bti_ecmp &
+                                           (lb->lb_n_buckets_minus_1));
+
+        if (choice->dpoi_type == DPO_BIER_FMASK)
+        {
+            bier_table_ecmp_set_fmask(btei, be->be_bp,
+                                      choice->dpoi_index);
+        }
+        else
+        {
+            /*
+             * any other type results in a drop, which we represent
+             * with an empty bucket
+             */
+            bier_table_ecmp_set_fmask(btei, be->be_bp,
+                                      INDEX_INVALID);
+        }
+
+        dpo_reset(&dpo);
+    }
+    else
+    {
+        /*
+         * no fmasks left. insert a drop
+         */
+        bier_table_ecmp_set_fmask(btei, be->be_bp, INDEX_INVALID);
+    }
+}
+
+void
+bier_entry_path_add (index_t bei,
+                     const fib_route_path_t *rpaths)
+{
+    fib_node_index_t old_pl_index;
+    bier_entry_t *be;
+
+    be = bier_entry_get(bei);
+    old_pl_index = be->be_path_list;
+
+    /*
+     * lock the path-list so it does not go away before we unlink
+     * from its resolved fmasks
+     */
+    fib_path_list_lock(old_pl_index);
+
+    if (FIB_NODE_INDEX_INVALID == be->be_path_list)
+    {
+        old_pl_index = FIB_NODE_INDEX_INVALID;
+        be->be_path_list = fib_path_list_create((FIB_PATH_LIST_FLAG_SHARED |
+                                                 FIB_PATH_LIST_FLAG_NO_URPF),
+                                                rpaths);
+        be->be_sibling_index = fib_path_list_child_add(be->be_path_list,
+                                                       FIB_NODE_TYPE_BIER_ENTRY,
+                                                       bier_entry_get_index(be));
+    }
+    else
+    {
+
+        old_pl_index = be->be_path_list;
+
+        be->be_path_list =
+            fib_path_list_copy_and_path_add(old_pl_index,
+                                            (FIB_PATH_LIST_FLAG_SHARED |
+                                             FIB_PATH_LIST_FLAG_NO_URPF),
+                                            rpaths);
+
+        fib_path_list_child_remove(old_pl_index,
+                                   be->be_sibling_index);
+        be->be_sibling_index = fib_path_list_child_add(be->be_path_list,
+                                                       FIB_NODE_TYPE_BIER_ENTRY,
+                                                       bier_entry_get_index(be));
+    }
+    /*
+     * link the entry's bit-position to each fmask in the new path-list
+     * then unlink from the old.
+     */
+    fib_path_list_walk(be->be_path_list,
+                       bier_entry_link_walk,
+                       be);
+    if (FIB_NODE_INDEX_INVALID != old_pl_index)
+    {
+        fib_path_list_walk(old_pl_index,
+                           bier_entry_unlink_walk,
+                           be);
+    }
+
+    /*
+     * update the ECNP tables with the new choice
+     */
+    bier_table_ecmp_walk(be->be_bti,
+                         bier_entry_table_ecmp_walk_add_fmask,
+                         be);
+
+    /*
+     * symmetric unlock. The old path-list may not exist hereinafter
+     */
+    fib_path_list_unlock(old_pl_index);
+}
+
+int
+bier_entry_path_remove (index_t bei,
+                        const fib_route_path_t *rpaths)
+{
+    fib_node_index_t old_pl_index;
+    bier_entry_t *be;
+
+    be = bier_entry_get(bei);
+    old_pl_index = be->be_path_list;
+
+    fib_path_list_lock(old_pl_index);
+
+    ASSERT (FIB_NODE_INDEX_INVALID != be->be_path_list);
+
+    be->be_path_list =
+        fib_path_list_copy_and_path_remove(old_pl_index,
+                                           (FIB_PATH_LIST_FLAG_SHARED |
+                                            FIB_PATH_LIST_FLAG_NO_URPF),
+                                           rpaths);
+
+    if (be->be_path_list != old_pl_index)
+    {
+        /*
+         * a path was removed
+         */
+        fib_path_list_child_remove(old_pl_index,
+                                   be->be_sibling_index);
+
+        if (FIB_NODE_INDEX_INVALID != be->be_path_list)
+        {
+            /*
+             * link the entry's bit-position to each fmask in the new path-list
+             * then unlink from the old.
+             */
+            fib_path_list_walk(be->be_path_list,
+                               bier_entry_link_walk,
+                               be);
+            be->be_sibling_index =
+                fib_path_list_child_add(be->be_path_list,
+                                        FIB_NODE_TYPE_BIER_ENTRY,
+                                        bier_entry_get_index(be));
+        }
+
+        fib_path_list_walk(old_pl_index,
+                           bier_entry_unlink_walk,
+                           be);
+    }
+    fib_path_list_unlock(old_pl_index);
+
+
+    /*
+     * update the ECNP tables with the new choice
+     */
+    bier_table_ecmp_walk(be->be_bti,
+                         bier_entry_table_ecmp_walk_add_fmask,
+                         be);
+
+    return (fib_path_list_get_n_paths(be->be_path_list));
+}
+
+void
+bier_entry_contribute_forwarding(index_t bei,
+                                 dpo_id_t *dpo)
+{
+    bier_entry_t *be = bier_entry_get(bei);
+
+    fib_path_list_contribute_forwarding(be->be_path_list,
+                                        FIB_FORW_CHAIN_TYPE_BIER,
+                                        dpo);
+}
+
+u8*
+format_bier_entry (u8* s, va_list *ap)
+{
+    index_t bei = va_arg(*ap, index_t);
+    bier_show_flags_t flags = va_arg(*ap, bier_show_flags_t);
+
+    bier_entry_t *be = bier_entry_get(bei);
+
+    s = format(s, " bp:%d\n", be->be_bp);
+    s = fib_path_list_format(be->be_path_list, s);
+
+    if (flags & BIER_SHOW_DETAIL)
+    {
+        dpo_id_t dpo = DPO_INVALID;
+
+        bier_entry_contribute_forwarding(bei, &dpo);
+
+        s = format(s, " forwarding:\n");
+        s = format(s, "  %U",
+                   format_dpo_id, &dpo, 2);
+        s = format(s, "\n");
+    }
+
+    return (s);
+}
+
+static fib_node_t *
+bier_entry_get_node (fib_node_index_t index)
+{
+    bier_entry_t *be = bier_entry_get(index);
+    return (&(be->be_node));
+}
+
+static bier_entry_t*
+bier_entry_get_from_node (fib_node_t *node)
+{
+    return ((bier_entry_t*)(((char*)node) -
+                            STRUCT_OFFSET_OF(bier_entry_t,
+                                             be_node)));
+}
+
+static void
+bier_entry_last_lock_gone (fib_node_t *node)
+{
+    /*
+     * the lifetime of the entry is managed by the table.
+     */
+    ASSERT(0);
+}
+
+/*
+ * A back walk has reached this BIER entry
+ */
+static fib_node_back_walk_rc_t
+bier_entry_back_walk_notify (fib_node_t *node,
+                             fib_node_back_walk_ctx_t *ctx)
+{
+    /*
+     * re-populate the ECMP tables with new choices
+     */
+    bier_entry_t *be = bier_entry_get_from_node(node);
+
+    bier_table_ecmp_walk(be->be_bti,
+                         bier_entry_table_ecmp_walk_add_fmask,
+                         be);
+
+    /*
+     * no need to propagate further up the graph.
+     */
+    return (FIB_NODE_BACK_WALK_CONTINUE);
+}
+
+/*
+ * The BIER fmask's graph node virtual function table
+ */
+static const fib_node_vft_t bier_entry_vft = {
+    .fnv_get = bier_entry_get_node,
+    .fnv_last_lock = bier_entry_last_lock_gone,
+    .fnv_back_walk = bier_entry_back_walk_notify,
+};
+
+clib_error_t *
+bier_entry_module_init (vlib_main_t * vm)
+{
+    fib_node_register_type (FIB_NODE_TYPE_BIER_ENTRY, &bier_entry_vft);
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_entry_module_init);
diff --git a/src/vnet/bier/bier_entry.h b/src/vnet/bier/bier_entry.h
new file mode 100644
index 00000000000..e514c64bb24
--- /dev/null
+++ b/src/vnet/bier/bier_entry.h
@@ -0,0 +1,94 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * bier_entry : The BIER entry
+ *
+ * The interface to the BIER entry is through a bier_entry_t* rather
+ * than an index. This is becuase the BIER table allocates the entries
+ * in a contiguous array once and only once when the table is created.
+ * this is done for forwarding performance. The entry is thus not subject
+ * to realloc, and does not need to be malloc'd when a route to that
+ * bit-position is first learned.
+ *
+ */
+
+#ifndef __BIER_ENTRY_H__
+#define __BIER_ENTRY_H__
+
+#include <vlib/vlib.h>
+#include <vnet/fib/fib_node.h>
+#include <vnet/bier/bier_types.h>
+
+/**
+ * Forward declarations
+ */
+struct bier_route_update_t_;
+struct bier_fmask_db_t_;
+
+/**
+ * The BIER entry
+ *
+ * the BIER entry is the representation of a BIER forwarding egress router (BFER)
+ * (or the egress PE) that is assigned a bit position.
+ */
+typedef struct bier_entry_t_ {
+    /**
+     * linkage into the FIB graph
+     */
+    fib_node_t be_node;
+
+    /**
+     * The index of the BIER table in which this entry resides
+     */
+    index_t be_bti;
+
+    /**
+     * the bit position this entry represents.
+     *  this is the key table insertion
+     */
+    bier_bp_t be_bp;
+
+    /**
+     * the FIB path-list this entry resolves through.
+     * the path-list is itself resoved on the entry's fmasks
+     */
+    fib_node_index_t be_path_list;
+    /**
+     * sibling index on the path list
+     */
+    fib_node_index_t be_sibling_index;
+} bier_entry_t;
+
+extern index_t bier_entry_create(index_t bti,
+                                 bier_bp_t bp);
+extern void bier_entry_delete(index_t bei);
+
+extern void bier_entry_path_add(index_t bei,
+                                const fib_route_path_t *brp);
+
+extern int bier_entry_path_remove(index_t bei,
+                                  const fib_route_path_t *brp);
+
+extern u8* format_bier_entry(u8* s, va_list *ap);
+
+extern void bier_entry_contribute_forwarding(index_t bei,
+                                             dpo_id_t *dpo);
+
+extern bier_entry_t *bier_entry_pool;
+always_inline bier_entry_t* bier_entry_get(index_t bei)
+{
+    return (&bier_entry_pool[bei]);
+}
+#endif
diff --git a/src/vnet/bier/bier_fmask.c b/src/vnet/bier/bier_fmask.c
new file mode 100644
index 00000000000..e30425cda4e
--- /dev/null
+++ b/src/vnet/bier/bier_fmask.c
@@ -0,0 +1,527 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/fib/fib_entry.h>
+#include <vnet/fib/fib_table.h>
+#include <vnet/fib/fib_walk.h>
+
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_fmask.h>
+#include <vnet/bier/bier_bit_string.h>
+#include <vnet/bier/bier_disp_table.h>
+
+#include <vnet/mpls/mpls.h>
+#include <vnet/dpo/drop_dpo.h>
+#include <vnet/dpo/load_balance.h>
+
+/*
+ * attributes names for formatting
+ */
+static const char *const bier_fmask_attr_names[] = BIER_FMASK_ATTR_NAMES;
+
+/*
+ * pool of BIER fmask objects
+ */
+bier_fmask_t *bier_fmask_pool;
+
+static inline index_t
+bier_fmask_get_index (const bier_fmask_t *bfm)
+{
+    return (bfm - bier_fmask_pool);
+}
+
+static void
+bier_fmask_bits_init (bier_fmask_bits_t *bits,
+                      bier_hdr_len_id_t hlid)
+{
+    bits->bfmb_refs = clib_mem_alloc(sizeof(bits->bfmb_refs[0]) *
+                                     bier_hdr_len_id_to_num_bits(hlid));
+    memset(bits->bfmb_refs,
+           0,
+           (sizeof(bits->bfmb_refs[0]) *
+            bier_hdr_len_id_to_num_bits(hlid)));
+
+    bits->bfmb_input_reset_string.bbs_len =
+        bier_hdr_len_id_to_num_buckets(hlid);
+
+    /*
+     * The buckets are accessed in the switch path
+     */
+    bits->bfmb_input_reset_string.bbs_buckets =
+        clib_mem_alloc_aligned(
+            sizeof(bits->bfmb_input_reset_string.bbs_buckets[0]) *
+            bier_hdr_len_id_to_num_buckets(hlid),
+            CLIB_CACHE_LINE_BYTES);
+    memset(bits->bfmb_input_reset_string.bbs_buckets,
+           0,
+           sizeof(bits->bfmb_input_reset_string.bbs_buckets[0]) *
+           bier_hdr_len_id_to_num_buckets(hlid));
+}
+
+static void
+bier_fmask_stack (bier_fmask_t *bfm)
+{
+    dpo_id_t via_dpo = DPO_INVALID;
+
+    if (bfm->bfm_flags & BIER_FMASK_FLAG_DISP)
+    {
+        bier_disp_table_contribute_forwarding(bfm->bfm_disp,
+                                              &via_dpo);
+    }
+    else
+    {
+        fib_entry_contribute_forwarding(bfm->bfm_fei,
+                                        FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS,
+                                        &via_dpo);
+    }
+
+    /*
+     * If the via fib entry provides no forwarding (i.e. a drop)
+     * then niether does this fmask. That way children consider this fmask
+     * unresolved and other ECMP options are used instead.
+     */
+    if (dpo_is_drop(&via_dpo) ||
+        load_balance_is_drop(&via_dpo))
+    {
+        bfm->bfm_flags &= ~BIER_FMASK_FLAG_FORWARDING;
+    }
+    else
+    {
+        bfm->bfm_flags |= BIER_FMASK_FLAG_FORWARDING;
+    }
+
+    dpo_stack(DPO_BIER_FMASK,
+              DPO_PROTO_BIER,
+              &bfm->bfm_dpo,
+              &via_dpo);
+    dpo_reset(&via_dpo);
+}
+
+void
+bier_fmask_contribute_forwarding (index_t bfmi,
+                                  dpo_id_t *dpo)
+{
+    bier_fmask_t *bfm;
+
+    bfm = bier_fmask_get(bfmi);
+
+    if (bfm->bfm_flags & BIER_FMASK_FLAG_FORWARDING)
+    {
+        dpo_set(dpo,
+                DPO_BIER_FMASK,
+                DPO_PROTO_BIER,
+                bfmi);
+    }
+    else
+    {
+        dpo_copy(dpo, drop_dpo_get(DPO_PROTO_BIER));
+    }
+}
+
+static void
+bier_fmask_resolve (bier_fmask_t *bfm)
+{
+    if (bfm->bfm_flags & BIER_FMASK_FLAG_DISP)
+    {
+        bier_disp_table_lock(bfm->bfm_disp);
+    }
+    else
+    {
+        /*
+         * source a recursive route through which we resolve.
+         */
+        fib_prefix_t pfx = {
+            .fp_addr = bfm->bfm_id.bfmi_nh,
+            .fp_proto = (ip46_address_is_ip4(&(bfm->bfm_id.bfmi_nh)) ?
+                         FIB_PROTOCOL_IP4 :
+                         FIB_PROTOCOL_IP6),
+            .fp_len = (ip46_address_is_ip4(&(bfm->bfm_id.bfmi_nh)) ? 32 : 128),
+        };
+
+        bfm->bfm_fei = fib_table_entry_special_add(0, // default table
+                                                   &pfx,
+                                                   FIB_SOURCE_RR,
+                                                   FIB_ENTRY_FLAG_NONE);
+
+        bfm->bfm_sibling = fib_entry_child_add(bfm->bfm_fei,
+                                               FIB_NODE_TYPE_BIER_FMASK,
+                                               bier_fmask_get_index(bfm));
+    }
+
+    bier_fmask_stack(bfm);
+}
+
+static void
+bier_fmask_unresolve (bier_fmask_t *bfm)
+{
+    if (bfm->bfm_flags & BIER_FMASK_FLAG_DISP)
+    {
+        bier_disp_table_unlock(bfm->bfm_disp);
+    }
+    else
+    {
+        /*
+         * un-source the recursive route through which we resolve.
+         */
+        fib_prefix_t pfx = {
+            .fp_addr = bfm->bfm_id.bfmi_nh,
+            .fp_proto = (ip46_address_is_ip4(&(bfm->bfm_id.bfmi_nh)) ?
+                         FIB_PROTOCOL_IP4 :
+                         FIB_PROTOCOL_IP6),
+            .fp_len = (ip46_address_is_ip4(&(bfm->bfm_id.bfmi_nh)) ? 32 : 128),
+        };
+
+        fib_entry_child_remove(bfm->bfm_fei, bfm->bfm_sibling);
+        fib_table_entry_special_remove(0, &pfx, FIB_SOURCE_RR);
+    }
+    dpo_reset(&bfm->bfm_dpo);
+}
+
+u32
+bier_fmask_child_add (fib_node_index_t bfmi,
+                     fib_node_type_t child_type,
+                     fib_node_index_t child_index)
+{
+    return (fib_node_child_add(FIB_NODE_TYPE_BIER_FMASK,
+                               bfmi,
+                               child_type,
+                               child_index));
+};
+
+void
+bier_fmask_child_remove (fib_node_index_t bfmi,
+                         u32 sibling_index)
+{
+    fib_node_child_remove(FIB_NODE_TYPE_BIER_FMASK,
+                          bfmi,
+                          sibling_index);
+}
+
+static void
+bier_fmask_init (bier_fmask_t *bfm,
+                 const bier_fmask_id_t *fmid,
+                 index_t bti,
+                 const fib_route_path_t *rpath)
+{
+    const bier_table_id_t *btid;
+    mpls_label_t olabel;
+
+    bfm->bfm_id = *fmid;
+    bfm->bfm_fib_index = bti;
+    dpo_reset(&bfm->bfm_dpo);
+
+    if (ip46_address_is_zero(&(bfm->bfm_id.bfmi_nh)))
+    {
+        bfm->bfm_flags |= BIER_FMASK_FLAG_DISP;
+    }
+
+    if (!(bfm->bfm_flags & BIER_FMASK_FLAG_DISP))
+    {
+        olabel = rpath->frp_label_stack[0];
+        vnet_mpls_uc_set_label(&bfm->bfm_label, olabel);
+        vnet_mpls_uc_set_exp(&bfm->bfm_label, 0);
+        vnet_mpls_uc_set_s(&bfm->bfm_label, 1);
+        vnet_mpls_uc_set_ttl(&bfm->bfm_label, 0xff);
+        bfm->bfm_label = clib_host_to_net_u32(bfm->bfm_label);
+    }
+    else
+    {
+        bfm->bfm_disp = rpath->frp_bier_fib_index;
+    }
+
+    btid = bier_table_get_id(bfm->bfm_fib_index);
+    bier_fmask_bits_init(&bfm->bfm_bits, btid->bti_hdr_len);
+    bier_fmask_resolve(bfm);
+}
+
+static void
+bier_fmask_destroy (bier_fmask_t *bfm)
+{
+    clib_mem_free(bfm->bfm_bits.bfmb_refs);
+    clib_mem_free(bfm->bfm_bits.bfmb_input_reset_string.bbs_buckets);
+
+    bier_fmask_db_remove(bfm->bfm_fib_index, &(bfm->bfm_id));
+    bier_fmask_unresolve(bfm);
+    pool_put(bier_fmask_pool, bfm);
+}
+
+void
+bier_fmask_unlock (index_t bfmi)
+{
+    bier_fmask_t *bfm;
+
+    if (INDEX_INVALID == bfmi)
+    {
+        return;
+    }
+
+    bfm = bier_fmask_get(bfmi);
+
+    fib_node_unlock(&bfm->bfm_node);
+}
+
+void
+bier_fmask_lock (index_t bfmi)
+{
+    bier_fmask_t *bfm;
+
+    if (INDEX_INVALID == bfmi)
+    {
+        return;
+    }
+
+    bfm = bier_fmask_get(bfmi);
+
+    fib_node_lock(&bfm->bfm_node);
+}
+
+index_t
+bier_fmask_create_and_lock (const bier_fmask_id_t *fmid,
+                            index_t bti,
+                            const fib_route_path_t *rpath)
+{
+    bier_fmask_t *bfm;
+
+    pool_get_aligned(bier_fmask_pool, bfm, CLIB_CACHE_LINE_BYTES);
+
+    memset(bfm, 0, sizeof(*bfm));
+
+    fib_node_init(&bfm->bfm_node, FIB_NODE_TYPE_BIER_FMASK);
+    bier_fmask_init(bfm, fmid, bti, rpath);
+
+    bier_fmask_lock(bier_fmask_get_index(bfm));
+
+    return (bier_fmask_get_index(bfm));
+}
+
+void
+bier_fmask_link (index_t bfmi,
+                 bier_bp_t bp)
+{
+    bier_fmask_t *bfm;
+
+    bfm = bier_fmask_get(bfmi);
+
+    if (0 == bfm->bfm_bits.bfmb_refs[BIER_BP_TO_INDEX(bp)])
+    {
+        /*
+         * 0 -> 1 transistion - set the bit in the string
+         */
+        bier_bit_string_set_bit(&bfm->bfm_bits.bfmb_input_reset_string, bp);
+    }
+
+    ++bfm->bfm_bits.bfmb_refs[BIER_BP_TO_INDEX(bp)];
+    ++bfm->bfm_bits.bfmb_count;
+}
+
+void
+bier_fmask_unlink (index_t bfmi,
+                   bier_bp_t bp)
+{
+    bier_fmask_t *bfm;
+
+    bfm = bier_fmask_get(bfmi);
+
+    --bfm->bfm_bits.bfmb_refs[BIER_BP_TO_INDEX(bp)];
+    --bfm->bfm_bits.bfmb_count;
+
+    if (0 == bfm->bfm_bits.bfmb_refs[BIER_BP_TO_INDEX(bp)])
+    {
+        /*
+         * 1 -> 0 transistion - clear the bit in the string
+         */
+        bier_bit_string_clear_bit(&bfm->bfm_bits.bfmb_input_reset_string, bp);
+    }
+}
+
+u8*
+format_bier_fmask (u8 *s, va_list *ap)
+{
+    index_t bfmi = va_arg(*ap, index_t);
+    u32 indent = va_arg(*ap, u32);
+    bier_fmask_attributes_t attr;
+    bier_fmask_t *bfm;
+
+    if (pool_is_free_index(bier_fmask_pool, bfmi))
+    {
+        return (format(s, "No BIER f-mask %d", bfmi));
+    }
+
+    bfm = bier_fmask_get(bfmi);
+
+    s = format(s, "fmask: nh:%U bs:%U locks:%d ",
+               format_ip46_address, &bfm->bfm_id.bfmi_nh, IP46_TYPE_ANY,
+               format_bier_bit_string, &bfm->bfm_bits.bfmb_input_reset_string,
+               bfm->bfm_node.fn_locks);
+    s = format(s, "flags:");
+    FOR_EACH_BIER_FMASK_ATTR(attr) {
+        if ((1<<attr) & bfm->bfm_flags) {
+            s = format (s, "%s,", bier_fmask_attr_names[attr]);
+        }
+    }
+    s = format(s, "\n%U%U",
+               format_white_space, indent,
+               format_dpo_id, &bfm->bfm_dpo, indent+2);
+
+    return (s);
+}
+
+
+static fib_node_t *
+bier_fmask_get_node (fib_node_index_t index)
+{
+    bier_fmask_t *bfm = bier_fmask_get(index);
+    return (&(bfm->bfm_node));
+}
+
+static bier_fmask_t*
+bier_fmask_get_from_node (fib_node_t *node)
+{
+    return ((bier_fmask_t*)(((char*)node) -
+                            STRUCT_OFFSET_OF(bier_fmask_t,
+                                             bfm_node)));
+}
+
+/*
+ * bier_fmask_last_lock_gone
+ */
+static void
+bier_fmask_last_lock_gone (fib_node_t *node)
+{
+    bier_fmask_destroy(bier_fmask_get_from_node(node));
+}
+
+/*
+ * bier_fmask_back_walk_notify
+ *
+ * A back walk has reached this BIER fmask
+ */
+static fib_node_back_walk_rc_t
+bier_fmask_back_walk_notify (fib_node_t *node,
+                             fib_node_back_walk_ctx_t *ctx)
+{
+    /*
+     * re-stack the fmask on the n-eos of the via
+     */
+    bier_fmask_t *bfm = bier_fmask_get_from_node(node);
+
+    bier_fmask_stack(bfm);
+
+    /*
+     * propagate further up the graph.
+     * we can do this synchronously since the fan out is small.
+     */
+    fib_walk_sync(FIB_NODE_TYPE_BIER_FMASK, bier_fmask_get_index(bfm), ctx);
+
+    return (FIB_NODE_BACK_WALK_CONTINUE);
+}
+
+/*
+ * The BIER fmask's graph node virtual function table
+ */
+static const fib_node_vft_t bier_fmask_vft = {
+    .fnv_get = bier_fmask_get_node,
+    .fnv_last_lock = bier_fmask_last_lock_gone,
+    .fnv_back_walk = bier_fmask_back_walk_notify,
+};
+
+static void
+bier_fmask_dpo_lock (dpo_id_t *dpo)
+{
+}
+
+static void
+bier_fmask_dpo_unlock (dpo_id_t *dpo)
+{
+}
+
+static void
+bier_fmask_dpo_mem_show (void)
+{
+    fib_show_memory_usage("BIER-fmask",
+                          pool_elts(bier_fmask_pool),
+                          pool_len(bier_fmask_pool),
+                          sizeof(bier_fmask_t));
+}
+
+const static dpo_vft_t bier_fmask_dpo_vft = {
+    .dv_lock = bier_fmask_dpo_lock,
+    .dv_unlock = bier_fmask_dpo_unlock,
+    .dv_mem_show = bier_fmask_dpo_mem_show,
+    .dv_format = format_bier_fmask,
+};
+
+const static char *const bier_fmask_mpls_nodes[] =
+{
+    "bier-output"
+};
+const static char * const * const bier_fmask_nodes[DPO_PROTO_NUM] =
+{
+    [DPO_PROTO_BIER] = bier_fmask_mpls_nodes,
+    [DPO_PROTO_MPLS] = bier_fmask_mpls_nodes,
+};
+
+clib_error_t *
+bier_fmask_module_init (vlib_main_t * vm)
+{
+    fib_node_register_type (FIB_NODE_TYPE_BIER_FMASK, &bier_fmask_vft);
+    dpo_register(DPO_BIER_FMASK, &bier_fmask_dpo_vft, bier_fmask_nodes);
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_fmask_module_init);
+
+static clib_error_t *
+bier_fmask_show (vlib_main_t * vm,
+                 unformat_input_t * input,
+                 vlib_cli_command_t * cmd)
+{
+    bier_fmask_t *bfm;
+    index_t bfmi;
+
+    bfmi = INDEX_INVALID;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "%d", &bfmi))
+        {
+            ;
+        } else
+        {
+            break;
+        }
+    }
+
+    if (INDEX_INVALID == bfmi)
+    {
+        pool_foreach(bfm, bier_fmask_pool,
+        ({
+            vlib_cli_output (vm, "%U",
+                             format_bier_fmask, bier_fmask_get_index(bfm), 0);
+        }));
+    }
+    else
+    {
+        vlib_cli_output (vm, "%U", format_bier_fmask, bfmi, 0);
+    }
+
+    return (NULL);
+}
+
+VLIB_CLI_COMMAND (show_bier_fmask, static) = {
+    .path = "show bier fmask",
+    .short_help = "show bier fmask",
+    .function = bier_fmask_show,
+};
diff --git a/src/vnet/bier/bier_fmask.h b/src/vnet/bier/bier_fmask.h
new file mode 100644
index 00000000000..d5ff4590ffe
--- /dev/null
+++ b/src/vnet/bier/bier_fmask.h
@@ -0,0 +1,198 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * @brief bier_fmask : The BIER fmask
+ *
+ * The BIER fmask contains the bitString that is applied to packets that
+ * egress towards the next-hop. As such the fmask is part of the rewrite
+ * (adj) for that next-hop. It it thus an extension of the next-hop and in
+ * no way associated with the bit-position(s) that are reachable through it.
+ * Fmasks are thus shared by bit-positions that egress throught the same
+ * nh (BFR-NBR).
+ * Deag fmasks are also shread in the event that a router has local
+ * bit-positions. This is necessary to prevent the router recieving two copies
+ * of each packet. Consequently it also means that they share the same
+ * disposition data for the global data.
+ */
+
+#ifndef __BIER_FMASK_H__
+#define __BIER_FMASK_H__
+
+#include <vlib/vlib.h>
+
+#include <vnet/fib/fib_node.h>
+#include <vnet/mpls/packet.h>
+#include <vnet/dpo/dpo.h>
+
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_fmask_db.h>
+
+/**
+ * A struct that represents the reference counting of the bits
+ */
+typedef struct bier_fmask_bits_t_ {
+    /**
+     * each bit in the mask needs to be reference counted
+     * and set/cleared on the 0->1 and 1->0 transitions.
+     */
+    bier_bit_string_t bfmb_input_reset_string;
+    u32 *bfmb_refs;
+
+    /**
+     * The total number of references to bits set on this mask
+     * in effect a count of the number of children.
+     */
+    uint32_t bfmb_count;
+} bier_fmask_bits_t;
+
+/**
+ * Flags on fmask
+ */
+typedef enum bier_fmask_attributes_t_
+{
+    BIER_FMASK_ATTR_FIRST,
+    BIER_FMASK_ATTR_FORWARDING = BIER_FMASK_ATTR_FIRST,
+    BIER_FMASK_ATTR_DISP,
+    BIER_FMASK_ATTR_LAST = BIER_FMASK_ATTR_DISP,
+} bier_fmask_attributes_t;
+
+#define BIER_FMASK_ATTR_NAMES {                         \
+     [BIER_FMASK_ATTR_FORWARDING] = "forwarding",       \
+     [BIER_FMASK_ATTR_DISP] = "disposition",            \
+}
+
+#define FOR_EACH_BIER_FMASK_ATTR(_item)          \
+    for (_item =  BIER_FMASK_ATTR_FIRST;         \
+         _item <= BIER_FMASK_ATTR_LAST;          \
+         _item++)
+
+typedef enum bier_fmask_flags_t_
+{
+    BIER_FMASK_FLAG_FORWARDING = (1 << BIER_FMASK_ATTR_FORWARDING),
+    BIER_FMASK_FLAG_DISP = (1 << BIER_FMASK_ATTR_DISP),
+} bier_fmask_flags_t;
+
+/**
+ * An outgoing BIER mask. aka forwarding bit mask (in the RFCs)
+ *
+ * This mask's function is two-fold
+ *  1 - it is logical-AND with the input packet header to produce the
+ *      output packet header
+ *  2 - it is logical NAND with the input packet header to modify the bit-mask
+ *      for the next lookup
+ */
+typedef struct bier_fmask_t_ {
+    /**
+     * The BIER fmask is a child of a FIB entry in the FIB graph.
+     */
+    fib_node_t bfm_node;
+
+    /**
+     * operational/state flags on the fmask
+     */
+    bier_fmask_flags_t bfm_flags;
+
+    /**
+     * The bits, and their ref counts, that are set on this mask
+     * This mask changes as BIER entries link to and from this fmask
+     */
+    bier_fmask_bits_t bfm_bits;
+
+    struct
+    {
+        /**
+         * The key to this fmask - used for store/lookup in the DB
+         */
+        bier_fmask_id_t bfm_id;
+
+        /**
+         * The BIER Table this Fmask is used in
+         */
+        index_t bfm_fib_index;
+    };
+
+    union
+    {
+        /**
+         * For forwarding via a next-hop
+         */
+        struct
+        {
+            /**
+             * The parent fib entry
+             */
+            fib_node_index_t bfm_fei;
+            /**
+             * The MPLS label to paint on the header during forwarding
+             */
+            mpls_label_t bfm_label;
+        };
+
+        /**
+         * For disposition
+         */
+        struct
+        {
+            /**
+             * The parent disposition table object
+             */
+            index_t bfm_disp;
+        };
+    };
+
+    /**
+     * the index of this fmask in the parent's child list.
+     */
+    u32 bfm_sibling;
+
+    /**
+     * The index into the adj table for the adj that
+     * this fmask resolves via
+     */
+    dpo_id_t bfm_dpo;
+} bier_fmask_t;
+
+extern void bier_fmask_link(index_t bfmi, bier_bp_t bp);
+extern void bier_fmask_unlink(index_t bfmi, bier_bp_t bp);
+extern void bier_fmask_unlock(index_t bfmi);
+extern void bier_fmask_lock(index_t bfmi);
+
+extern index_t bier_fmask_create_and_lock(const bier_fmask_id_t *fmid,
+                                          index_t bti,
+                                          const fib_route_path_t *rpath);
+
+extern u8* format_bier_fmask(u8 *s, va_list *ap);
+
+extern void bier_fmask_contribute_forwarding(index_t bfmi,
+                                             dpo_id_t *dpo);
+
+extern u32 bier_fmask_child_add (fib_node_index_t fib_entry_index,
+                                 fib_node_type_t child_type,
+                                 fib_node_index_t child_index);
+extern void bier_fmask_child_remove (fib_node_index_t fib_entry_index,
+                                     u32 sibling_index);
+
+/*
+ * provided for fast data-path access
+ */
+bier_fmask_t *bier_fmask_pool;
+
+static inline bier_fmask_t *
+bier_fmask_get (u32 index)
+{
+    return (pool_elt_at_index(bier_fmask_pool, index));
+}
+
+#endif
diff --git a/src/vnet/bier/bier_fmask_db.c b/src/vnet/bier/bier_fmask_db.c
new file mode 100644
index 00000000000..37cbb365897
--- /dev/null
+++ b/src/vnet/bier/bier_fmask_db.c
@@ -0,0 +1,161 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/bier/bier_fmask_db.h>
+#include <vnet/bier/bier_fmask.h>
+
+/**
+ * Global Table of fmask objects
+ * The key into this table includes the table's key and the fmask's key,
+ * so there could be a DB per-table. But it is more efficient
+ * at forwarding time to extract the fmask from a single global table
+ * which is hot in dcache.
+ *
+ * The table's key is part of this DB key, since the fmasks therein build up
+ * their forwarding mask based on the routes that resolve through
+ * it, so cross polination would be bad.
+ */
+typedef struct bier_fmask_db_t_ {
+    /**
+     * hash table for underlying storage
+     */
+    mhash_t bfdb_hash;
+
+    /**
+     * Pool for memory
+     */
+    struct bier_fmask_t_ *bfdb_pool;
+} bier_fmask_db_t;
+
+/**
+ * The key used in the fmask DB to compare fmask objects.
+ * There is one global DB, so we need to use the table's ID and the fmasks ID
+ */
+typedef struct bier_fmask_db_key_t_ {
+    bier_fmask_id_t bfmdbk_fm_id;
+    index_t bfmdbk_tbl_id;
+} bier_fmask_db_key_t;
+// TODO packed?
+
+/**
+ * Single fmask DB
+ */
+static bier_fmask_db_t bier_fmask_db;
+
+
+u32
+bier_fmask_get_index (const bier_fmask_t *bfm)
+{
+    return (bfm - bier_fmask_db.bfdb_pool);
+}
+
+u32
+bier_fmask_db_find_or_create_and_lock (index_t bti,
+                                       const bier_fmask_id_t *fmid,
+                                       const fib_route_path_t *rpath)
+{
+    bier_fmask_db_key_t key;
+    u32 index;
+    uword *p;
+
+    /*
+     * there be padding in that thar key, and it's
+     * used as a memcmp in the mhash.
+     */
+    memset(&key, 0, sizeof(key));
+    key.bfmdbk_tbl_id = bti;
+    key.bfmdbk_fm_id = *fmid;
+
+    index = INDEX_INVALID;
+    p = mhash_get (&bier_fmask_db.bfdb_hash, &key);
+
+    if (NULL == p)
+    {
+        /*
+         * adding a new fmask object
+         */
+        index = bier_fmask_create_and_lock(fmid, bti, rpath);
+
+        mhash_set (&bier_fmask_db.bfdb_hash, &key, index, 0 /*old_value*/);
+    }
+    else
+    {
+        index = p[0];
+        bier_fmask_lock(index);
+    }
+
+    return (index);
+}
+
+u32
+bier_fmask_db_find (index_t bti,
+                    const bier_fmask_id_t *fmid)
+{
+    bier_fmask_db_key_t key;
+    u32 index;
+    uword *p;
+
+    /*
+     * there be padding in that thar key, and it's
+     * used as a memcmp in the mhash.
+     */
+    memset(&key, 0, sizeof(key));
+    key.bfmdbk_tbl_id = bti;
+    key.bfmdbk_fm_id = *fmid;
+
+    index = INDEX_INVALID;
+    p = mhash_get(&bier_fmask_db.bfdb_hash, &key);
+
+    if (NULL != p)
+    {
+        index = p[0];
+    }
+
+    return (index);
+}
+
+void
+bier_fmask_db_remove (index_t bti,
+                      const bier_fmask_id_t *fmid)
+{
+    bier_fmask_db_key_t key = {
+        .bfmdbk_tbl_id = bti,
+        .bfmdbk_fm_id = *fmid,
+    };
+    uword *p;
+
+    p = mhash_get (&bier_fmask_db.bfdb_hash, &key);
+
+    if (NULL == p) {
+        /*
+         * remove a non-exitant entry - oops
+         */
+        ASSERT (!"remove non-existant fmask");
+    } else {
+        mhash_unset (&(bier_fmask_db.bfdb_hash), &key, 0);
+    }
+}
+
+clib_error_t *
+bier_fmask_db_module_init (vlib_main_t *vm)
+{
+    mhash_init (&bier_fmask_db.bfdb_hash,
+                sizeof(uword),
+                sizeof(bier_fmask_db_key_t));
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_fmask_db_module_init);
diff --git a/src/vnet/bier/bier_fmask_db.h b/src/vnet/bier/bier_fmask_db.h
new file mode 100644
index 00000000000..6ba40f3a839
--- /dev/null
+++ b/src/vnet/bier/bier_fmask_db.h
@@ -0,0 +1,61 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * @brief bier_fmask_db : The BIER fmask Database
+ */
+
+#ifndef __BIER_FMASK_DB_H__
+#define __BIER_FMASK_DB_H__
+
+#include <vnet/ip/ip.h>
+
+#include <vnet/bier/bier_types.h>
+
+/**
+ * Foward declarations
+ */
+struct bier_fmask_t_;
+
+typedef enum bier_hdr_type_t_ {
+    BIER_HDR_IN_IP6,
+    BIER_HDR_O_MPLS,
+} bier_hdr_type_t;
+
+typedef struct bier_fmask_id_t_ {
+    /**
+     * Type of BIER header this fmask supports
+     */
+    bier_hdr_type_t bfmi_hdr_type;
+
+    /**
+     * next-hop of the peer
+     */
+    ip46_address_t bfmi_nh;
+} bier_fmask_id_t;
+
+extern u32
+bier_fmask_db_find_or_create_and_lock(index_t bti,
+                                      const bier_fmask_id_t *fmid,
+                                      const fib_route_path_t *rpath);
+
+extern u32
+bier_fmask_db_find(index_t bti,
+                   const bier_fmask_id_t *fmid);
+
+extern void
+bier_fmask_db_remove(index_t bti,
+                     const bier_fmask_id_t *fmid);
+
+#endif
diff --git a/src/vnet/bier/bier_hdr_inlines.h b/src/vnet/bier/bier_hdr_inlines.h
new file mode 100644
index 00000000000..4437dc82e24
--- /dev/null
+++ b/src/vnet/bier/bier_hdr_inlines.h
@@ -0,0 +1,217 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * @brief The BIER inline functions acting on the bier header
+ */
+
+#ifndef __BIER_HDR_INLINES_H__
+#define __BIER_HDR_INLINES_H__
+
+#include <vppinfra/byte_order.h>
+#include <vppinfra/string.h>
+
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_bit_string.h>
+#include <vnet/ip/ip6_packet.h>
+
+/**
+ * Mask and shift values for the fields incorporated
+ * into the header's first word
+ */
+#define BIER_HDR_1ST_NIBBLE_MASK    0xf0000000
+#define BIER_HDR_VERSION_FIELD_MASK 0x0f000000
+#define BIER_HDR_LEN_FIELD_MASK     0x00f00000
+#define BIER_HDR_ENTROPY_FIELD_MASK 0x000fffff
+
+#define BIER_HDR_1ST_NIBBLE_SHIFT    28
+#define BIER_HDR_VERSION_FIELD_SHIFT 24
+#define BIER_HDR_LEN_FIELD_SHIFT     20
+#define BIER_HDR_ENTROPY_FIELD_SHIFT  0
+
+#define BIER_HDR_1ST_NIBBLE_VALUE 0x5
+
+/**
+ * Mask and shift values for fields in the headers trainling word
+ */
+#define BIER_HDR_PROTO_FIELD_MASK   0x003f
+#define BIER_HDR_OAM_FIELD_MASK     0xc000
+#define BIER_HDR_DSCP_FIELD_MASK    0x0fc0
+#define BIER_HDR_DSCP_FIELD_SHIFT   6
+#define BIER_HDR_PROTO_FIELD_SHIFT  0
+#define BIER_HDR_OAM_FIELD_SHIFT    14
+
+static inline bier_hdr_version_t
+bier_hdr_get_version (const bier_hdr_t *bier_hdr)
+{
+    return ((bier_hdr->bh_first_word &
+             BIER_HDR_VERSION_FIELD_MASK) >>
+            BIER_HDR_VERSION_FIELD_SHIFT);
+}
+
+static inline bier_hdr_len_id_t
+bier_hdr_get_len_id (const bier_hdr_t *bier_hdr)
+{
+    return ((bier_hdr->bh_first_word &
+             BIER_HDR_LEN_FIELD_MASK) >>
+            BIER_HDR_LEN_FIELD_SHIFT);
+}
+
+static inline bier_hdr_entropy_t
+bier_hdr_get_entropy (const bier_hdr_t *bier_hdr)
+{
+    return ((bier_hdr->bh_first_word &
+             BIER_HDR_ENTROPY_FIELD_MASK) >>
+            BIER_HDR_ENTROPY_FIELD_SHIFT);
+}
+
+static inline void
+bier_hdr_1st_nibble (bier_hdr_t *hdr)
+{
+    hdr->bh_first_word &= ~(BIER_HDR_1ST_NIBBLE_MASK);
+    hdr->bh_first_word |= (BIER_HDR_1ST_NIBBLE_VALUE <<
+                           BIER_HDR_1ST_NIBBLE_SHIFT);
+}
+
+static inline u8
+bier_hdr_get_1st_nibble (bier_hdr_t *hdr)
+{
+    return ((hdr->bh_first_word & BIER_HDR_1ST_NIBBLE_MASK) >>
+            BIER_HDR_1ST_NIBBLE_SHIFT);
+}
+
+static inline void
+bier_hdr_set_version (bier_hdr_t *hdr,
+                      bier_hdr_version_t version)
+{
+    hdr->bh_first_word &= ~(BIER_HDR_VERSION_FIELD_MASK);
+    hdr->bh_first_word |= (version << BIER_HDR_VERSION_FIELD_SHIFT);
+}
+
+static inline void
+bier_hdr_set_len_id (bier_hdr_t *hdr,
+                     bier_hdr_len_id_t len)
+{
+    hdr->bh_first_word &= ~(BIER_HDR_LEN_FIELD_MASK);
+    hdr->bh_first_word |= (len << BIER_HDR_LEN_FIELD_SHIFT);
+}
+
+static inline void
+bier_hdr_set_entropy (bier_hdr_t *hdr,
+                      bier_hdr_entropy_t entropy)
+{
+    entropy = entropy & BIER_HDR_ENTROPY_FIELD_MASK;
+    hdr->bh_first_word &= ~(BIER_HDR_ENTROPY_FIELD_MASK);
+    hdr->bh_first_word |= (entropy << BIER_HDR_ENTROPY_FIELD_SHIFT);
+}
+
+#define BIER_HDR_FIRST_WORD(version, len, entropy)          \
+    ((BIER_HDR_1ST_NIBBLE_VALUE <<                          \
+      BIER_HDR_1ST_NIBBLE_SHIFT) |                          \
+     (version << BIER_HDR_VERSION_FIELD_SHIFT) |            \
+     (len     << BIER_HDR_LEN_FIELD_SHIFT)     |            \
+     ((entropy & BIER_HDR_ENTROPY_FIELD_MASK)               \
+      << BIER_HDR_ENTROPY_FIELD_SHIFT))
+
+static inline void
+bier_hdr_ntoh (bier_hdr_t *bier_hdr)
+{
+    bier_hdr->bh_first_word = clib_net_to_host_u32(bier_hdr->bh_first_word);
+    bier_hdr->bh_oam_dscp_proto = clib_net_to_host_u16(bier_hdr->bh_oam_dscp_proto);
+    bier_hdr->bh_bfr_id = clib_net_to_host_u16(bier_hdr->bh_bfr_id);
+}
+
+static inline void
+bier_hdr_hton (bier_hdr_t *bier_hdr)
+{
+    bier_hdr->bh_first_word = clib_host_to_net_u32(bier_hdr->bh_first_word);
+    bier_hdr->bh_oam_dscp_proto = clib_host_to_net_u16(bier_hdr->bh_oam_dscp_proto);
+    bier_hdr->bh_bfr_id = clib_host_to_net_u16(bier_hdr->bh_bfr_id);
+}
+
+static inline bier_hdr_src_id_t
+bier_hdr_get_src_id (const bier_hdr_t *bier_hdr)
+{
+    return (bier_hdr->bh_bfr_id);
+}
+
+static inline void
+bier_hdr_set_src_id (bier_hdr_t *bier_hdr,
+                     bier_hdr_src_id_t src_id)
+{
+    bier_hdr->bh_bfr_id = src_id;
+}
+static inline void
+bier_hdr_set_proto_id (bier_hdr_t *bier_hdr,
+                       bier_hdr_proto_id_t proto)
+{
+    bier_hdr->bh_oam_dscp_proto &= ~(BIER_HDR_PROTO_FIELD_MASK);
+    bier_hdr->bh_oam_dscp_proto |= (proto << BIER_HDR_PROTO_FIELD_SHIFT);
+}
+
+static inline bier_hdr_proto_id_t
+bier_hdr_get_proto_id (const bier_hdr_t *bier_hdr)
+{
+    return ((bier_hdr->bh_oam_dscp_proto & BIER_HDR_PROTO_FIELD_MASK) >>
+            BIER_HDR_PROTO_FIELD_SHIFT);
+}
+
+static inline void
+bier_hdr_clear (bier_hdr_t *bier_hdr)
+{
+    memset(&bier_hdr->bh_bit_string, 0,
+           bier_hdr_len_id_to_num_buckets(
+               bier_hdr_get_len_id(bier_hdr)));
+}
+
+static inline void
+bier_hdr_init (bier_hdr_t *bier_hdr,
+               bier_hdr_version_t version,
+               bier_hdr_proto_id_t proto,
+               bier_hdr_len_id_t len,
+               bier_hdr_entropy_t entropy,
+               bier_bp_t src)
+{
+    bier_hdr_1st_nibble(bier_hdr);
+    bier_hdr_set_version(bier_hdr, version);
+    bier_hdr_set_len_id(bier_hdr, len);
+    bier_hdr_set_entropy(bier_hdr, entropy);
+    bier_hdr_set_proto_id(bier_hdr, proto);
+    bier_hdr_set_src_id(bier_hdr, src);
+    bier_hdr_clear(bier_hdr);
+}
+
+static inline size_t
+bier_hdr_str_num_bytes (const bier_hdr_t *bier_hdr)
+{
+    return (bier_hdr_len_id_to_num_bytes(
+                bier_hdr_get_len_id(bier_hdr)));
+}
+
+static inline size_t
+bier_hdr_num_bytes (const bier_hdr_t *bier_hdr)
+{
+    return (sizeof(bier_hdr_t) +
+            bier_hdr_str_num_bytes(bier_hdr));
+}
+
+static inline void
+bier_bit_string_init_from_hdr (bier_hdr_t *bier_hdr,
+                               bier_bit_string_t *bit_string)
+{
+    bit_string->bbs_len = bier_hdr_str_num_bytes(bier_hdr);
+    bit_string->bbs_buckets = bier_hdr->bh_bit_string;
+}
+
+#endif
diff --git a/src/vnet/bier/bier_imp.c b/src/vnet/bier/bier_imp.c
new file mode 100644
index 00000000000..c51dede9c0a
--- /dev/null
+++ b/src/vnet/bier/bier_imp.c
@@ -0,0 +1,288 @@
+/*
+ * Copyright (c) 2017 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * bier_imposition : The BIER imposition object
+ *
+ * A BIER imposition object is present in the IP mcast output list
+ * and represents the imposition of a BIER bitmask. After BIER header
+ * imposition the packet is forward within the appropriate/specifid
+ * BIER table
+ */
+
+#include <vnet/bier/bier_imp.h>
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+#include <vnet/fib/fib_node.h>
+#include <vnet/mpls/mpls_types.h>
+
+/**
+ * The memory pool of all imp objects
+ */
+bier_imp_t *bier_imp_pool;
+
+/**
+ * When constructing the BIER imp ID from an index and BSL, shift
+ * the BSL this far
+ */
+#define BIER_IMP_ID_HLEN_SHIFT 24
+
+static void
+bier_imp_lock_i (bier_imp_t *bi)
+{
+    bi->bi_locks++;
+}
+
+void
+bier_imp_lock (index_t bii)
+{
+    bier_imp_lock_i(bier_imp_get(bii));
+}
+
+static index_t
+bier_imp_get_index(bier_imp_t *bi)
+{
+    return (bi - bier_imp_pool);
+}
+
+index_t
+bier_imp_add_or_lock (const bier_table_id_t *bti,
+                      bier_bp_t sender,
+                      const bier_bit_string_t *bs)
+{
+    bier_imp_t *bi = NULL;
+    fib_protocol_t fproto;
+    index_t btii;
+
+    pool_get_aligned(bier_imp_pool, bi, CLIB_CACHE_LINE_BYTES);
+
+    bi->bi_tbl = *bti;
+    btii = bier_table_add_or_lock(bti, MPLS_LABEL_INVALID);
+
+    /*
+     * init the BIER header we will paint on in the data plane
+     */
+    bier_hdr_init(&bi->bi_hdr,
+                  BIER_HDR_VERSION_1,
+                  BIER_HDR_PROTO_INVALID, // filled in later
+                  bti->bti_hdr_len,
+                  0, // entropy
+                  sender);
+    bier_hdr_hton(&bi->bi_hdr);
+    clib_memcpy(&bi->bi_bits, bs->bbs_buckets, bs->bbs_len);
+
+    bier_imp_lock_i(bi);
+
+    /*
+     * get and stack on the forwarding info from the table
+     */
+    FOR_EACH_FIB_IP_PROTOCOL(fproto)
+    {
+        /*
+         * initialise to invalid first, lest we pick up garbage
+         * from the pool alloc
+         */
+        dpo_id_t dpo = DPO_INVALID;
+        bi->bi_dpo[fproto] = dpo;
+
+        bier_table_contribute_forwarding(btii, &dpo);
+        dpo_stack(DPO_BIER_IMP, fib_proto_to_dpo(fproto),
+                  &bi->bi_dpo[fproto],
+                  &dpo);
+        dpo_reset(&dpo);
+    }
+
+    return (bier_imp_get_index(bi));
+}
+
+void
+bier_imp_unlock (index_t bii)
+{
+    fib_protocol_t fproto;
+    bier_imp_t *bi;
+
+    if (INDEX_INVALID == bii)
+    {
+        return;
+    }
+
+    bi = bier_imp_get(bii);
+
+    bi->bi_locks--;
+
+    if (0 == bi->bi_locks)
+    {
+        bier_table_unlock(&bi->bi_tbl);
+
+        FOR_EACH_FIB_IP_PROTOCOL(fproto)
+        {
+            dpo_reset(&bi->bi_dpo[fproto]);
+        }
+        pool_put(bier_imp_pool, bi);
+    }
+}
+
+u8*
+format_bier_imp (u8* s, va_list *args)
+{
+    index_t bii = va_arg (*args, index_t);
+    u32 indent = va_arg(*args, u32);
+    bier_show_flags_t flags = va_arg(*args, bier_show_flags_t);
+    bier_imp_t *bi;
+
+    bi = bier_imp_get(bii);
+
+    s = format(s, "bier-imp:[%d]: tbl:[%U] hdr:[%U]",
+               bier_imp_get_index(bi),
+               format_bier_table_id, &bi->bi_tbl,
+               format_bier_hdr, &bi->bi_hdr);
+
+    if (BIER_SHOW_DETAIL & flags)
+    {
+        bier_bit_string_t bbs;
+        bier_hdr_t copy;
+
+        copy = bi->bi_hdr;
+        bier_hdr_ntoh(&copy);
+        bier_bit_string_init(&bbs,
+                             bier_hdr_get_len_id(&copy),
+                             bi->bi_bits.bits);
+
+        s = format(s, "\n%U%U",
+                   format_white_space, indent,
+                   format_bier_bit_string, &bbs);
+        s = format(s, "\n%U%U",
+                   format_white_space, indent,
+                   format_dpo_id, &bi->bi_dpo, indent+2);
+    }
+
+    return (s);
+}
+
+void
+bier_imp_contribute_forwarding (index_t bii,
+                                dpo_proto_t proto,
+                                dpo_id_t *dpo)
+{
+    dpo_set(dpo, DPO_BIER_IMP, proto, bii);
+}
+
+const static char* const bier_imp_ip4_nodes[] =
+{
+    "bier-imp-ip4",
+    NULL,
+};
+const static char* const bier_imp_ip6_nodes[] =
+{
+    "bier-imp-ip6",
+    NULL,
+};
+
+const static char* const * const bier_imp_nodes[DPO_PROTO_NUM] =
+{
+    [DPO_PROTO_IP4]  = bier_imp_ip4_nodes,
+    [DPO_PROTO_IP6]  = bier_imp_ip6_nodes,
+};
+
+static void
+bier_imp_dpo_lock (dpo_id_t *dpo)
+{
+    bier_imp_lock(dpo->dpoi_index);
+}
+
+static void
+bier_imp_dpo_unlock (dpo_id_t *dpo)
+{
+    bier_imp_unlock(dpo->dpoi_index);
+}
+
+static void
+bier_imp_dpo_mem_show (void)
+{
+    fib_show_memory_usage("BIER imposition",
+                          pool_elts(bier_imp_pool),
+                          pool_len(bier_imp_pool),
+                          sizeof(bier_imp_t));
+}
+
+static u8*
+format_bier_imp_dpo (u8* s, va_list *ap)
+{
+    index_t index = va_arg(*ap, index_t);
+    u32 indent = va_arg(*ap, u32);
+
+    s = format(s, "%U", format_bier_imp, index, indent, BIER_SHOW_DETAIL);
+
+    return (s);
+}
+
+const static dpo_vft_t bier_imp_vft = {
+    .dv_lock = bier_imp_dpo_lock,
+    .dv_unlock = bier_imp_dpo_unlock,
+    .dv_format = format_bier_imp_dpo,
+    .dv_mem_show = bier_imp_dpo_mem_show,
+};
+
+clib_error_t *
+bier_imp_db_module_init (vlib_main_t *vm)
+{
+    dpo_register(DPO_BIER_IMP, &bier_imp_vft, bier_imp_nodes);
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_imp_db_module_init);
+
+static clib_error_t *
+show_bier_imp (vlib_main_t * vm,
+               unformat_input_t * input,
+               vlib_cli_command_t * cmd)
+{
+    bier_imp_t *bi;
+    index_t bii;
+
+    bii = INDEX_INVALID;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "%d", &bii))
+            ;
+        else
+        {
+            break;
+        }
+    }
+
+    if (INDEX_INVALID == bii)
+    {
+        pool_foreach(bi, bier_imp_pool,
+        ({
+            vlib_cli_output(vm, "%U", format_bier_imp,
+                            bier_imp_get_index(bi),
+                            1,
+                            BIER_SHOW_BRIEF);
+        }));
+    }
+    else
+    {
+        vlib_cli_output(vm, "%U", format_bier_imp, bii, 1,
+                        BIER_SHOW_DETAIL);
+    }
+    return (NULL);
+}
+
+VLIB_CLI_COMMAND (show_bier_imp_node, static) = {
+    .path = "show bier imp",
+    .short_help = "show bier imp [index]",
+    .function = show_bier_imp,
+};
diff --git a/src/vnet/bier/bier_imp.h b/src/vnet/bier/bier_imp.h
new file mode 100644
index 00000000000..fa53989fc93
--- /dev/null
+++ b/src/vnet/bier/bier_imp.h
@@ -0,0 +1,87 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * bier_imposition : The BIER imposition object
+ *
+ * A BIER imposition object is present in the IP mcast output list
+ * and represents the imposition of a BIER bitmask. After BIER header
+ * imposition the packet is forward within the appropriate/specifid
+ * BIER table
+ */
+
+#ifndef __BIER_IMPOSITION_H__
+#define __BIER_IMPOSITION_H__
+
+#include <vnet/bier/bier_types.h>
+#include <vnet/fib/fib_types.h>
+#include <vnet/dpo/dpo.h>
+
+/**
+ * The BIER imposition object
+ */
+typedef struct bier_imp_t_ {
+    /**
+     * The BIER table into which to forward the post imposed packet
+     */
+    bier_table_id_t bi_tbl;
+
+    /**
+     * number of locks
+     */
+    u32 bi_locks;
+
+    /**
+     * The DPO contirubted from the resolving BIER table.
+     * One per-IP protocol. This allows us to share a BIER imposition
+     * object for a IPv4 and IPv6 mfib path.
+     */
+    dpo_id_t bi_dpo[FIB_PROTOCOL_IP_MAX];
+
+    /**
+     * The Header to impose.
+     */
+    bier_hdr_t bi_hdr;
+
+    /**
+     * The bit string.
+     *  This is a memory v. speed tradeoff. We inline here the
+     *  largest header type so as the bitstring is on the same
+     *  cacheline as the header.
+     */
+    bier_bit_mask_4096_t bi_bits;
+} bier_imp_t;
+
+extern index_t bier_imp_add_or_lock(const bier_table_id_t *bt,
+                                    bier_bp_t sender,
+                                    const bier_bit_string_t *bs);
+
+extern void bier_imp_unlock(index_t bii);
+extern void bier_imp_lock(index_t bii);
+
+extern u8* format_bier_imp(u8* s, va_list *ap);
+
+extern void bier_imp_contribute_forwarding(index_t bii,
+                                           dpo_proto_t proto,
+                                           dpo_id_t *dpo);
+
+extern bier_imp_t *bier_imp_pool;
+
+always_inline bier_imp_t*
+bier_imp_get (index_t bii)
+{
+    return (pool_elt_at_index(bier_imp_pool, bii));
+}
+
+#endif
diff --git a/src/vnet/bier/bier_imp_node.c b/src/vnet/bier/bier_imp_node.c
new file mode 100644
index 00000000000..e9aae93b460
--- /dev/null
+++ b/src/vnet/bier/bier_imp_node.c
@@ -0,0 +1,217 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/bier/bier_imp.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+#include <vnet/ip/ip.h>
+
+/**
+ * @brief A struct to hold tracing information for the BIER imposition
+ * node.
+ */
+typedef struct bier_imp_trace_t_
+{
+    /**
+     * BIER imposition object hit
+     */
+    index_t imp;
+
+    /**
+     * BIER hdr applied
+     */
+    bier_hdr_t hdr;
+} bier_imp_trace_t;
+
+always_inline uword
+bier_imp_dpo_inline (vlib_main_t * vm,
+                     vlib_node_runtime_t * node,
+                     vlib_frame_t * from_frame,
+                     fib_protocol_t fproto,
+                     bier_hdr_proto_id_t bproto)
+{
+    u32 n_left_from, next_index, * from, * to_next;
+
+    from = vlib_frame_vector_args (from_frame);
+    n_left_from = from_frame->n_vectors;
+
+    next_index = node->cached_next_index;
+
+    while (n_left_from > 0)
+    {
+        u32 n_left_to_next;
+
+        vlib_get_next_frame(vm, node, next_index, to_next, n_left_to_next);
+
+        while (n_left_from > 0 && n_left_to_next > 0)
+        {
+            vlib_buffer_t * b0;
+            bier_imp_t *bimp0;
+            bier_hdr_t *hdr0;
+            u32 bi0, bii0;
+            u32 next0;
+
+            bi0 = from[0];
+            to_next[0] = bi0;
+            from += 1;
+            to_next += 1;
+            n_left_from -= 1;
+            n_left_to_next -= 1;
+
+            b0 = vlib_get_buffer (vm, bi0);
+
+            bii0 = vnet_buffer(b0)->ip.adj_index[VLIB_TX];
+            bimp0 = bier_imp_get(bii0);
+
+            if (FIB_PROTOCOL_IP4 == fproto)
+            {
+                /*
+                 * decrement the TTL on ingress to the BIER domain
+                 */
+                ip4_header_t * ip0 = vlib_buffer_get_current(b0);
+                u32 checksum0;
+
+                checksum0 = ip0->checksum + clib_host_to_net_u16 (0x0100);
+                checksum0 += checksum0 >= 0xffff;
+
+                ip0->checksum = checksum0;
+                ip0->ttl -= 1;
+
+                /*
+                 * calculate an entropy
+                 */
+                if (0 == vnet_buffer(b0)->ip.flow_hash)
+                {
+                    vnet_buffer(b0)->ip.flow_hash =
+                        ip4_compute_flow_hash (ip0, IP_FLOW_HASH_DEFAULT);
+                }
+            }
+            if (FIB_PROTOCOL_IP6 == fproto)
+            {
+                /*
+                 * decrement the TTL on ingress to the BIER domain
+                 */
+                ip6_header_t * ip0 = vlib_buffer_get_current(b0);
+
+                ip0->hop_limit -= 1;
+
+                /*
+                 * calculate an entropy
+                 */
+                if (0 == vnet_buffer(b0)->ip.flow_hash)
+                {
+                    vnet_buffer(b0)->ip.flow_hash =
+                        ip6_compute_flow_hash (ip0, IP_FLOW_HASH_DEFAULT);
+                }
+            }
+
+            /* Paint the BIER header */
+            vlib_buffer_advance(b0, -(sizeof(bier_hdr_t) +
+                                      bier_hdr_len_id_to_num_bytes(bimp0->bi_tbl.bti_hdr_len)));
+            hdr0 = vlib_buffer_get_current(b0);
+            clib_memcpy(hdr0, &bimp0->bi_hdr,
+                        (sizeof(bier_hdr_t) +
+                         bier_hdr_len_id_to_num_bytes(bimp0->bi_tbl.bti_hdr_len)));
+            /*
+             * Fixup the entropy and protocol, both of which have a
+             * zero value post the paint job
+             */
+            hdr0->bh_oam_dscp_proto |=
+                clib_host_to_net_u16(bproto << BIER_HDR_PROTO_FIELD_SHIFT);
+            hdr0->bh_first_word |=
+                clib_host_to_net_u32((vnet_buffer(b0)->ip.flow_hash &
+                                      BIER_HDR_ENTROPY_FIELD_MASK) <<
+                                     BIER_HDR_ENTROPY_FIELD_SHIFT);
+
+            /* next node */
+            next0 = bimp0->bi_dpo[fproto].dpoi_next_node;
+            vnet_buffer(b0)->ip.adj_index[VLIB_TX] =
+                bimp0->bi_dpo[fproto].dpoi_index;
+
+            if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+            {
+                bier_imp_trace_t *tr =
+                    vlib_add_trace (vm, node, b0, sizeof (*tr));
+                tr->imp = bii0;
+                tr->hdr = *hdr0;
+            }
+
+            vlib_validate_buffer_enqueue_x1(vm, node, next_index, to_next,
+                                            n_left_to_next, bi0, next0);
+        }
+        vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+    return from_frame->n_vectors;
+}
+
+static u8 *
+format_bier_imp_trace (u8 * s, va_list * args)
+{
+    CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+    CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+    bier_imp_trace_t * t;
+    u32 indent;
+
+    t = va_arg (*args, bier_imp_trace_t *);
+    indent = format_get_indent (s);
+
+    s = format (s, "%U", format_bier_imp, t->imp, indent, BIER_SHOW_BRIEF);
+    return (s);
+}
+
+static uword
+bier_imp_ip4 (vlib_main_t * vm,
+              vlib_node_runtime_t * node,
+              vlib_frame_t * frame)
+{
+    return (bier_imp_dpo_inline(vm, node, frame,
+                                FIB_PROTOCOL_IP4,
+                                BIER_HDR_PROTO_IPV4));
+}
+
+VLIB_REGISTER_NODE (bier_imp_ip4_node) = {
+    .function = bier_imp_ip4,
+    .name = "bier-imp-ip4",
+    .vector_size = sizeof (u32),
+
+    .format_trace = format_bier_imp_trace,
+    .n_next_nodes = 1,
+    .next_nodes = {
+        [0] = "error-drop",
+    }
+};
+VLIB_NODE_FUNCTION_MULTIARCH (bier_imp_ip4_node, bier_imp_ip4)
+
+static uword
+bier_imp_ip6 (vlib_main_t * vm,
+              vlib_node_runtime_t * node,
+              vlib_frame_t * frame)
+{
+    return (bier_imp_dpo_inline(vm, node, frame,
+                                FIB_PROTOCOL_IP6,
+                                BIER_HDR_PROTO_IPV6));
+}
+
+VLIB_REGISTER_NODE (bier_imp_ip6_node) = {
+    .function = bier_imp_ip6,
+    .name = "bier-imp-ip6",
+    .vector_size = sizeof (u32),
+
+    .format_trace = format_bier_imp_trace,
+    .n_next_nodes = 1,
+    .next_nodes = {
+        [0] = "error-drop",
+    }
+};
+VLIB_NODE_FUNCTION_MULTIARCH (bier_imp_ip6_node, bier_imp_ip6)
diff --git a/src/vnet/bier/bier_input.c b/src/vnet/bier/bier_input.c
new file mode 100644
index 00000000000..88b37fc80c0
--- /dev/null
+++ b/src/vnet/bier/bier_input.c
@@ -0,0 +1,175 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/buffer.h>
+
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+
+typedef enum {
+#define bier_error(n,s) BIER_INPUT_ERROR_##n,
+#include <vnet/bier/bier_input_error.def>
+#undef bier_error
+    BIER_INPUT_N_ERROR,
+} bier_input_error_t;
+
+static char * bier_error_strings[] = {
+#define bier_error(n,s) s,
+#include <vnet/bier/bier_input_error.def>
+#undef bier_error
+};
+
+typedef enum bier_input_next_t_ {
+    BIER_INPUT_NEXT_BIER_LOOKUP,
+    BIER_INPUT_NEXT_DROP,
+    BIER_INPUT_N_NEXT,
+} bier_input_next_t;
+
+vlib_node_registration_t bier_input_node;
+
+/**
+ * @brief Packet trace recoed for a BIER output
+ */
+typedef struct bier_input_trace_t_
+{
+    u32 next_index;
+    u32 bt_index;
+} bier_input_trace_t;
+
+static int
+bier_hdr_validate (bier_hdr_t *bier_hdr,
+                   bier_hdr_len_id_t expected_length)
+{
+    /*
+     * checks:
+     *  - the version field must be 1
+     *  - the header length matches the length expected
+     */
+    if (PREDICT_FALSE((BIER_HDR_VERSION_1 != bier_hdr_get_version(bier_hdr)) ||
+                      (expected_length != bier_hdr_get_len_id(bier_hdr)))) {
+        return (0);
+    }
+
+    return (1);
+}
+
+static uword
+bier_input (vlib_main_t * vm,
+            vlib_node_runtime_t * node,
+            vlib_frame_t * from_frame)
+{
+    u32 n_left_from, next_index, * from, * to_next;
+
+    from = vlib_frame_vector_args (from_frame);
+    n_left_from = from_frame->n_vectors;
+
+    /*
+     * objection your honour! speculation!
+     */
+    next_index = node->cached_next_index;
+
+    while (n_left_from > 0)
+    {
+        u32 n_left_to_next;
+
+        vlib_get_next_frame (vm, node, next_index,
+                             to_next, n_left_to_next);
+
+        while (n_left_from > 0 && n_left_to_next > 0)
+        {
+            const bier_table_t *bt0;
+            vlib_buffer_t * b0;
+            bier_hdr_t * bh0;
+            u32 bi0, next0;
+            u32 bt_index0;
+
+            bi0 = from[0];
+            to_next[0] = bi0;
+            from += 1;
+            to_next += 1;
+            n_left_from -= 1;
+            n_left_to_next -= 1;
+
+            b0 = vlib_get_buffer (vm, bi0);
+            bh0 = vlib_buffer_get_current (b0);
+            bier_hdr_ntoh(bh0);
+
+            /*
+             * In the MPLS decap node we squirelled away the
+             * index for the BIER table as the tx adjacency
+             */
+            bt_index0 = vnet_buffer(b0)->ip.adj_index[VLIB_TX];
+            bt0 = bier_table_get(bt_index0);
+
+            if (PREDICT_TRUE(bier_hdr_validate(bh0, bt0->bt_id.bti_hdr_len)))
+            {
+                next0 = BIER_INPUT_NEXT_BIER_LOOKUP;
+            } else {
+                next0 = BIER_INPUT_NEXT_DROP;
+                b0->error = node->errors[BIER_INPUT_ERROR_INVALID_HEADER];
+            }
+
+            if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+            {
+                bier_input_trace_t *tr;
+
+                tr = vlib_add_trace (vm, node, b0, sizeof (*tr));
+                tr->next_index = next0;
+                tr->bt_index = bt_index0;
+            }
+
+            vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
+                                             to_next, n_left_to_next,
+                                             bi0, next0);
+        }
+
+        vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+
+    vlib_node_increment_counter (vm, bier_input_node.index,
+                                 BIER_INPUT_ERROR_PKTS_VALID,
+                                 from_frame->n_vectors);
+    return (from_frame->n_vectors);
+}
+
+static u8 *
+format_bier_input_trace (u8 * s, va_list * args)
+{
+    CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+    CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+    bier_input_trace_t * t = va_arg (*args, bier_input_trace_t *);
+
+    s = format (s, " next [%d], BIER Table index %d",
+                t->next_index, t->bt_index);
+    return s;
+}
+
+VLIB_REGISTER_NODE (bier_input_node) = {
+    .function = bier_input,
+    .name = "bier-input",
+    /* Takes a vector of packets. */
+    .vector_size = sizeof (u32),
+
+    .n_errors = BIER_INPUT_N_ERROR,
+    .error_strings = bier_error_strings,
+
+    .n_next_nodes = BIER_INPUT_N_NEXT,
+    .next_nodes = {
+        [BIER_INPUT_NEXT_BIER_LOOKUP] = "bier-lookup",
+        [BIER_INPUT_NEXT_DROP] = "bier-drop",
+    },
+
+    .format_trace = format_bier_input_trace,
+};
diff --git a/src/vnet/bier/bier_input_error.def b/src/vnet/bier/bier_input_error.def
new file mode 100644
index 00000000000..6f0034316f0
--- /dev/null
+++ b/src/vnet/bier/bier_input_error.def
@@ -0,0 +1,19 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+bier_error (NONE, "no error")
+bier_error (TABLE_NOT_FOUND, "table not found")
+bier_error (INVALID_HEADER, "invalid header")
+bier_error (PKTS_VALID, "BIER input packets")
diff --git a/src/vnet/bier/bier_lookup.c b/src/vnet/bier/bier_lookup.c
new file mode 100644
index 00000000000..4cf29f886e4
--- /dev/null
+++ b/src/vnet/bier/bier_lookup.c
@@ -0,0 +1,367 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/buffer.h>
+#include <vnet/vnet.h>
+
+#include <vnet/bier/bier_fmask.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_fmask.h>
+
+/**
+ * Struct maintining the per-worker thread data for BIER lookups
+ */
+typedef struct bier_lookup_main_t_
+{
+    /* per-cpu vector of cloned packets */
+    u32 **blm_clones;
+    /* per-cpu vector of BIER fmasks */
+    u32 **blm_fmasks;
+} bier_lookup_main_t;
+
+/**
+ * Single instance of the lookup main
+ */
+static bier_lookup_main_t bier_lookup_main;
+
+static char * bier_lookup_error_strings[] = {
+#define bier_error(n,s) s,
+#include <vnet/bier/bier_lookup_error.def>
+#undef bier_error
+};
+
+/*
+ * Keep these values sematically the same as BIER lookup
+ */
+#define foreach_bier_lookup_next                \
+    _(DROP, "bier-drop")                        \
+    _(OUTPUT, "bier-output")
+
+typedef enum {
+#define _(s,n) BIER_LOOKUP_NEXT_##s,
+    foreach_bier_lookup_next
+#undef _
+    BIER_LOOKUP_N_NEXT,
+} bier_lookup_next_t;
+
+typedef enum {
+#define bier_error(n,s) BIER_LOOKUP_ERROR_##n,
+#include <vnet/bier/bier_lookup_error.def>
+#undef bier_error
+    BIER_LOOKUP_N_ERROR,
+} bier_lookup_error_t;
+
+vlib_node_registration_t bier_lookup_node;
+
+/**
+ * @brief Packet trace recoed for a BIER lookup
+ */
+typedef struct bier_lookup_trace_t_
+{
+    u32 next_index;
+    index_t bt_index;
+    index_t bfm_index;
+} bier_lookup_trace_t;
+
+static uword
+bier_lookup (vlib_main_t * vm,
+             vlib_node_runtime_t * node,
+             vlib_frame_t * from_frame)
+{
+    u32 n_left_from, next_index, * from, * to_next;
+    bier_lookup_main_t *blm = &bier_lookup_main;
+    u32 thread_index = vlib_get_thread_index();
+
+    from = vlib_frame_vector_args (from_frame);
+    n_left_from = from_frame->n_vectors;
+    next_index = BIER_LOOKUP_NEXT_DROP;
+
+    while (n_left_from > 0)
+    {
+        u32 n_left_to_next;
+
+        vlib_get_next_frame (vm, node, next_index,
+                             to_next, n_left_to_next);
+
+        while (n_left_from > 0 && n_left_to_next > 0)
+        {
+            bier_bit_mask_bucket_t buckets_copy[BIER_HDR_BUCKETS_256];
+            u32 next0, bi0, n_bytes, bti0, bfmi0;
+            const bier_fmask_t *bfm0;
+            const bier_table_t *bt0;
+            u16 index, num_buckets;
+            const bier_hdr_t *bh0;
+            bier_bit_string_t bbs;
+            vlib_buffer_t *b0;
+            bier_bp_t fbs;
+            int bucket;
+
+            bi0 = from[0];
+            from += 1;
+            n_left_from -= 1;
+
+            b0 = vlib_get_buffer (vm, bi0);
+            bh0 = vlib_buffer_get_current (b0);
+            bti0 = vnet_buffer(b0)->ip.adj_index[VLIB_TX];
+
+            /*
+             * default to drop so that if no bits are matched then
+             * that is where we go - DROP.
+             */
+            next0 = BIER_LOOKUP_NEXT_DROP;
+
+            /*
+             * At the imposition or input node,
+             * we stored the BIER Table index in the TX adjacency
+             */
+            bt0 = bier_table_get(vnet_buffer(b0)->ip.adj_index[VLIB_TX]);
+
+            /*
+             * we should only forward via one for the ECMP tables
+             */
+            ASSERT(!bier_table_is_main(bt0));
+
+            /*
+             * number of integer sized buckets
+             */
+            n_bytes = bier_hdr_len_id_to_num_buckets(bt0->bt_id.bti_hdr_len);
+            vnet_buffer(b0)->bier.n_bytes = n_bytes;
+            vnet_buffer(b0)->sw_if_index[VLIB_TX] = ~0;
+            num_buckets = n_bytes / sizeof(int);
+            bier_bit_string_init(&bbs,
+                                 bt0->bt_id.bti_hdr_len,
+                                 buckets_copy);
+            memcpy(bbs.bbs_buckets, bh0->bh_bit_string, bbs.bbs_len);
+
+            /*
+             * reset the fmask and clone storage vectors
+             */
+            vec_reset_length (blm->blm_fmasks[thread_index]);
+            vec_reset_length (blm->blm_clones[thread_index]);
+
+            /*
+             * Loop through the buckets in the header
+             */
+            for (index = 0; index < num_buckets; index++) {
+                /*
+                 * loop through each bit in the bucket
+                 */
+                bucket = ((int*)bbs.bbs_buckets)[index];
+
+                while (bucket) {
+                    fbs  = bier_find_first_bit_string_set(bucket);
+                    fbs += (((num_buckets - 1) - index) *
+                            BIER_BIT_MASK_BITS_PER_INT);
+
+                    bfmi0 = bier_table_fwd_lookup(bt0, fbs);
+
+                    /*
+                     * whatever happens, the bit we just looked for
+                     * MUST be cleared from the packet
+                     * otherwise we could be in this loop a while ...
+                     */
+                    bier_bit_string_clear_bit(&bbs, fbs);
+
+                    if (PREDICT_TRUE(INDEX_INVALID != bfmi0))
+                    {
+                        bfm0 = bier_fmask_get(bfmi0);
+                        vnet_buffer (b0)->ip.adj_index[VLIB_TX] = bfmi0;
+
+                        /*
+                         * use the bit-string on the fmask to reset
+                         * the bits in the header we are walking
+                         */
+                        bier_bit_string_clear_string(
+                            &bfm0->bfm_bits.bfmb_input_reset_string,
+                            &bbs);
+                        bucket = ((int*)bbs.bbs_buckets)[index];
+
+                        /*
+                         * the fmask is resolved so replicate a
+                         * packet its way
+                         */
+                        next0 = BIER_LOOKUP_NEXT_OUTPUT;
+
+                        vec_add1 (blm->blm_fmasks[thread_index], bfmi0);
+                    } else {
+                        /*
+                         * go to the next bit-position set
+                         */
+                        bucket = ((int*)bbs.bbs_buckets)[index];
+                        continue;
+                    }
+                }
+            }
+
+            /*
+             * Full mask now processed.
+             * Create the number of clones we need based on the number
+             * of fmasks we are sending to.
+             */
+            u8 num_cloned, clone;
+            u32 n_clones;
+
+            n_clones = vec_len(blm->blm_fmasks[thread_index]);
+
+            if (PREDICT_TRUE(0 != n_clones))
+            {
+                ASSERT(n_clones < 256);
+                num_cloned = vlib_buffer_clone(vm, bi0,
+                                               blm->blm_clones[thread_index],
+                                               n_clones, 128);
+
+                if (num_cloned != n_clones)
+                {
+                    vlib_node_increment_counter
+                        (vm, node->node_index,
+                         BIER_LOOKUP_ERROR_BUFFER_ALLOCATION_FAILURE, 1);
+                }
+
+                for (clone = 0; clone < num_cloned; clone++)
+                {
+                    vlib_buffer_t *c0;
+                    u32 ci0;
+
+                    ci0 = blm->blm_clones[thread_index][clone];
+                    c0 = vlib_get_buffer(vm, ci0);
+
+                    to_next[0] = ci0;
+                    to_next += 1;
+                    n_left_to_next -= 1;
+
+                    if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+                    {
+                        bier_lookup_trace_t *tr;
+
+                        vlib_trace_buffer (vm, node, next0, c0, 0);
+                        tr = vlib_add_trace (vm, node, c0, sizeof (*tr));
+                        tr->bt_index = bti0;
+                        tr->bfm_index = blm->blm_fmasks[thread_index][clone];
+
+                        c0->flags |= VLIB_BUFFER_IS_TRACED;
+                    }
+
+                    vlib_validate_buffer_enqueue_x1(vm, node, next_index,
+                                                    to_next, n_left_to_next,
+                                                    ci0, next0);
+
+                    /*
+                     * After the enqueue it is possible that we over-flow the
+                     * frame of the to-next node. When this happens we need to
+                     * 'put' that full frame to the node and get a fresh empty
+                     * one. Note that these are macros with side effects that
+                     * change to_next & n_left_to_next
+                     */
+                    if (PREDICT_FALSE(0 == n_left_to_next))
+                    {
+                        vlib_put_next_frame (vm, node, next_index,
+                                             n_left_to_next);
+                        vlib_get_next_frame (vm, node, next_index,
+                                             to_next, n_left_to_next);
+                    }
+                }
+            }
+            else
+            {
+                /*
+                 * no clones/replications required. drop this packet
+                 */
+                next0 = BIER_LOOKUP_NEXT_DROP;
+                to_next[0] = bi0;
+                to_next += 1;
+                n_left_to_next -= 1;
+
+                if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+                {
+                    bier_lookup_trace_t *tr;
+
+                    tr = vlib_add_trace (vm, node, b0, sizeof (*tr));
+
+                    tr->bt_index = bti0;
+                    tr->bfm_index = ~0;
+                }
+
+                vlib_validate_buffer_enqueue_x1(vm, node, next_index,
+                                                to_next, n_left_to_next,
+                                                bi0, next0);
+            }
+        }
+
+        vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+
+    vlib_node_increment_counter (vm, bier_lookup_node.index,
+                                 BIER_LOOKUP_ERROR_NONE,
+                                 from_frame->n_vectors);
+    return (from_frame->n_vectors);
+}
+
+static u8 *
+format_bier_lookup_trace (u8 * s, va_list * args)
+{
+    CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+    CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+    bier_lookup_trace_t * t = va_arg (*args, bier_lookup_trace_t *);
+
+    s = format (s, "BIER: next [%d], tbl:%d BFM:%d",
+                t->next_index,
+                t->bt_index,
+                t->bfm_index);
+    return s;
+}
+
+VLIB_REGISTER_NODE (bier_lookup_node) = {
+    .function = bier_lookup,
+    .name = "bier-lookup",
+    /* Takes a vector of packets. */
+    .vector_size = sizeof (u32),
+
+    .n_errors = BIER_LOOKUP_N_ERROR,
+    .error_strings = bier_lookup_error_strings,
+
+    .format_trace = format_bier_lookup_trace,
+    .n_next_nodes = BIER_LOOKUP_N_NEXT,
+    .next_nodes = {
+        [BIER_LOOKUP_NEXT_DROP] = "bier-drop",
+        [BIER_LOOKUP_NEXT_OUTPUT] = "bier-output",
+    },
+};
+
+clib_error_t *
+bier_lookup_module_init (vlib_main_t * vm)
+{
+    bier_lookup_main_t *blm = &bier_lookup_main;
+    u32 thread_index;
+
+    vec_validate (blm->blm_clones, vlib_num_workers());
+    vec_validate (blm->blm_fmasks, vlib_num_workers());
+
+    for (thread_index = 0;
+         thread_index <= vlib_num_workers();
+         thread_index++)
+    {
+        /*
+         *  4096 is the most we will ever need to support
+         * a Bit-Mask length of 4096
+         */
+        vec_validate(blm->blm_fmasks[thread_index], 4095);
+        vec_validate(blm->blm_clones[thread_index], 4095);
+    }
+
+    return 0;
+}
+
+VLIB_INIT_FUNCTION (bier_lookup_module_init);
diff --git a/src/vnet/bier/bier_lookup_error.def b/src/vnet/bier/bier_lookup_error.def
new file mode 100644
index 00000000000..a0510b40ca3
--- /dev/null
+++ b/src/vnet/bier/bier_lookup_error.def
@@ -0,0 +1,18 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+bier_error (NONE, "no error")
+bier_error (FMASK_UNRES, "fmask unresolved")
+bier_error (BUFFER_ALLOCATION_FAILURE, "Buffer Allocation Failure")
diff --git a/src/vnet/bier/bier_output.c b/src/vnet/bier/bier_output.c
new file mode 100644
index 00000000000..fce6c50b309
--- /dev/null
+++ b/src/vnet/bier/bier_output.c
@@ -0,0 +1,195 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/buffer.h>
+
+#include <vnet/bier/bier_fmask.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+#include <vlib/vlib.h>
+
+static char * bier_output_error_strings[] = {
+#define bier_error(n,s) s,
+#include <vnet/bier/bier_output_error.def>
+#undef bier_error
+};
+
+/*
+ * Keep these values sematically the same as BIER output
+ */
+#define foreach_bier_output_next                \
+    _(DROP, "bier-drop")
+
+typedef enum {
+#define _(s,n) BIER_OUTPUT_NEXT_##s,
+    foreach_bier_output_next
+#undef _
+    BIER_OUTPUT_N_NEXT,
+} bier_output_next_t;
+
+typedef enum {
+#define bier_error(n,s) BIER_OUTPUT_ERROR_##n,
+#include <vnet/bier/bier_output_error.def>
+#undef bier_error
+    BIER_OUTPUT_N_ERROR,
+} bier_output_error_t;
+
+/**
+ * Forward declaration
+ */
+vlib_node_registration_t bier_output_node;
+
+/**
+ * @brief Packet trace recoed for a BIER output
+ */
+typedef struct bier_output_trace_t_
+{
+    u32 next_index;
+    index_t bfm_index;
+} bier_output_trace_t;
+
+static uword
+bier_output (vlib_main_t * vm,
+             vlib_node_runtime_t * node,
+             vlib_frame_t * from_frame)
+{
+    u32 n_left_from, next_index, * from, * to_next;
+
+    from = vlib_frame_vector_args (from_frame);
+    n_left_from = from_frame->n_vectors;
+
+    // vnet_buffer(b0)->sw_if_index[VLIB_TX] = d0->tx_fib_index;
+
+    /*
+     * objection your honour! speculation!
+     */
+    next_index = node->cached_next_index;
+
+    while (n_left_from > 0)
+    {
+        u32 n_left_to_next;
+
+        vlib_get_next_frame (vm, node, next_index,
+                             to_next, n_left_to_next);
+
+        while (n_left_from > 0 && n_left_to_next > 0)
+        {
+            bier_output_next_t next0;
+            bier_bit_string_t bbs;
+            vlib_buffer_t * b0;
+            bier_fmask_t *bfm0;
+            bier_hdr_t *bh0;
+            u32 bi0, *h0;
+            u32 bfmi0;
+
+            bi0 = from[0];
+            to_next[0] = bi0;
+            from += 1;
+            to_next += 1;
+            n_left_from -= 1;
+            n_left_to_next -= 1;
+
+            b0 = vlib_get_buffer (vm, bi0);
+            bh0 = vlib_buffer_get_current (b0);
+            bier_bit_string_init_from_hdr(bh0, &bbs);
+
+            /*
+             * In the BIER Lookup node we squirelled away the
+             * BIER fmask index as the adj index
+             */
+            bfmi0 = vnet_buffer (b0)->ip.adj_index[VLIB_TX];
+            bfm0 = bier_fmask_get(bfmi0);
+
+            /*
+             * perform the logical AND of the packet's mask with
+             * that of the fmask objects, to reset the bits that
+             * are only on the shortest path the the fmask NH.
+             */
+            bier_bit_string_logical_and_string(
+                &bfm0->bfm_bits.bfmb_input_reset_string,
+                &bbs);
+
+            /*
+             * this is the last time we touch the BIER header
+             * so flip to network order
+             */
+            bier_hdr_hton(bh0);
+
+            /*
+             * paint the BIER peer's label
+             */
+            if (!(bfm0->bfm_flags & BIER_FMASK_FLAG_DISP))
+            {
+                vlib_buffer_advance(b0, -(word)sizeof(mpls_label_t));
+                h0 = vlib_buffer_get_current(b0);
+                h0[0] = bfm0->bfm_label;
+            }
+
+            /*
+             * setup next graph node
+             */
+            next0 = bfm0->bfm_dpo.dpoi_next_node;
+            vnet_buffer(b0)->ip.adj_index[VLIB_TX] = bfm0->bfm_dpo.dpoi_index;
+
+            if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
+            {
+                bier_output_trace_t *tr;
+
+                tr = vlib_add_trace (vm, node, b0, sizeof (*tr));
+                tr->next_index = next0;
+                tr->bfm_index = bfmi0;
+            }
+
+            vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
+                                             to_next, n_left_to_next,
+                                             bi0, next0);
+        }
+
+        vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+
+    vlib_node_increment_counter (vm, bier_output_node.index,
+                                 BIER_OUTPUT_ERROR_NONE,
+                                 from_frame->n_vectors);
+    return (from_frame->n_vectors);
+}
+
+static u8 *
+format_bier_output_trace (u8 * s, va_list * args)
+{
+    CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+    CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+    bier_output_trace_t * t = va_arg (*args, bier_output_trace_t *);
+
+    s = format (s, " next [%d], BFM index %d",
+                t->next_index, t->bfm_index);
+    return s;
+}
+
+VLIB_REGISTER_NODE (bier_output_node) = {
+    .function = bier_output,
+    .name = "bier-output",
+    /* Takes a vector of packets. */
+    .vector_size = sizeof (u32),
+
+    .n_errors = BIER_OUTPUT_N_ERROR,
+    .error_strings = bier_output_error_strings,
+
+    .n_next_nodes = BIER_OUTPUT_N_NEXT,
+    .next_nodes = {
+        [BIER_OUTPUT_NEXT_DROP] = "bier-drop",
+    },
+
+    .format_trace = format_bier_output_trace,
+};
diff --git a/src/vnet/bier/bier_output_error.def b/src/vnet/bier/bier_output_error.def
new file mode 100644
index 00000000000..983c13c0c2d
--- /dev/null
+++ b/src/vnet/bier/bier_output_error.def
@@ -0,0 +1,17 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+bier_error (NONE, "no error")
+bier_error (FMASK_UNRES, "fmask unresolved")
diff --git a/src/vnet/bier/bier_table.c b/src/vnet/bier/bier_table.c
new file mode 100644
index 00000000000..74a09910601
--- /dev/null
+++ b/src/vnet/bier/bier_table.c
@@ -0,0 +1,744 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vppinfra/vec.h>
+
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_entry.h>
+#include <vnet/bier/bier_update.h>
+#include <vnet/bier/bier_fmask_db.h>
+#include <vnet/bier/bier_fmask.h>
+
+#include <vnet/fib/mpls_fib.h>
+#include <vnet/mpls/mpls.h>
+#include <vnet/fib/fib_path_list.h>
+
+/**
+ * Memory pool of all the allocated tables
+ */
+bier_table_t *bier_table_pool;
+
+/**
+ * DB store of all BIER tables index by SD/set/hdr-len
+ */
+static uword *bier_tables_by_key;
+
+/**
+ * The magic number of BIER ECMP tables to create.
+ * The load-balance distribution algorithm will use a power of 2
+ * for the number of buckets, which constrains the choice.
+ */
+#define BIER_N_ECMP_TABLES 16
+
+static inline index_t
+bier_table_get_index (const bier_table_t *bt)
+{
+    return (bt - bier_table_pool);
+}
+
+int
+bier_table_is_main (const bier_table_t *bt)
+{
+    return (BIER_ECMP_TABLE_ID_MAIN == bt->bt_id.bti_ecmp);
+}
+
+/*
+ * Construct the key to use to find a BIER table
+ * in the global hash map
+ */
+static u32
+bier_table_mk_key (const bier_table_id_t *id)
+{
+    /*
+     * the set and sub-domain Ids are 8 bit values.
+     * we have space for ECMP table ID and talbe type (SPF/TE)
+     * for later
+     */
+    u32 key = ((id->bti_sub_domain << 24)  |
+               (id->bti_set << 16) |
+               (id->bti_ecmp << 8) |
+               (id->bti_hdr_len << 4) |
+               (id->bti_type));
+
+    return (key);
+}
+
+static void
+bier_table_init (bier_table_t *bt,
+                 const bier_table_id_t *id,
+                 mpls_label_t ll)
+{
+    u32 num_entries;
+
+    bt->bt_lfei = FIB_NODE_INDEX_INVALID;
+    bt->bt_id = *id;
+    bt->bt_ll = ll;
+    num_entries = bier_hdr_len_id_to_num_bits(bt->bt_id.bti_hdr_len);
+
+    /*
+     * create the lookup table of entries.
+     */
+    if (bier_table_is_main(bt))
+    {
+        vec_validate_init_empty_aligned(bt->bt_entries,
+                                        num_entries,
+                                        INDEX_INVALID,
+                                        CLIB_CACHE_LINE_BYTES);
+        fib_table_find_or_create_and_lock(FIB_PROTOCOL_MPLS,
+                                          MPLS_FIB_DEFAULT_TABLE_ID,
+                                          FIB_SOURCE_BIER);
+    }
+    else
+    {
+        vec_validate_init_empty_aligned(bt->bt_fmasks,
+                                        num_entries,
+                                        INDEX_INVALID,
+                                        CLIB_CACHE_LINE_BYTES);
+    }
+}
+
+static void
+bier_table_rm_lfib (bier_table_t *bt)
+{
+    if (FIB_NODE_INDEX_INVALID != bt->bt_lfei)
+    {
+        fib_table_entry_delete_index(bt->bt_lfei,
+                                     FIB_SOURCE_BIER);
+    }
+    bt->bt_lfei = FIB_NODE_INDEX_INVALID;
+}
+
+static void
+bier_table_destroy (bier_table_t *bt)
+{
+    if (bier_table_is_main(bt))
+    {
+        index_t *bei;
+
+        fib_path_list_unlock(bt->bt_pl);
+        bt->bt_pl = FIB_NODE_INDEX_INVALID;
+        /*
+         * unresolve/remove all entries from the table
+         */
+        vec_foreach (bei, bt->bt_entries)
+        {
+            if (INDEX_INVALID != *bei)
+            {
+                bier_entry_delete(*bei);
+            }
+        }
+        vec_free (bt->bt_entries);
+        fib_table_unlock(fib_table_find(FIB_PROTOCOL_MPLS,
+                                        MPLS_FIB_DEFAULT_TABLE_ID),
+                         FIB_PROTOCOL_MPLS,
+                         FIB_SOURCE_BIER);
+    }
+    else
+    {
+        index_t *bfmi;
+
+        /*
+         * unlock any fmasks
+         */
+        vec_foreach (bfmi, bt->bt_fmasks)
+        {
+            bier_fmask_unlock(*bfmi);
+        }
+        vec_free(bt->bt_fmasks);
+    }
+
+    hash_unset(bier_tables_by_key,
+               bier_table_mk_key(&bt->bt_id));
+    pool_put(bier_table_pool, bt);
+}
+
+static void
+bier_table_lock_i (bier_table_t *bt)
+{
+    bt->bt_locks++;
+}
+
+static void
+bier_table_unlock_i (bier_table_t *bt)
+{
+    bt->bt_locks--;
+
+    if (0 == bt->bt_locks)
+    {
+        bier_table_rm_lfib(bt);
+        bier_table_destroy(bt);
+    }
+}
+
+void
+bier_table_unlock (const bier_table_id_t *bti)
+{
+    uword *p;
+    u32 key;
+
+    key = bier_table_mk_key(bti);
+
+    p = hash_get (bier_tables_by_key, key);
+
+    if (NULL != p) {
+        bier_table_unlock_i(bier_table_get(p[0]));
+    }
+}
+
+static void
+bier_table_mk_lfib (bier_table_t *bt)
+{
+    /*
+     * Add a new MPLS lfib entry
+     */
+    if (MPLS_LABEL_INVALID != bt->bt_ll) {
+        fib_prefix_t pfx = {
+            .fp_proto = FIB_PROTOCOL_MPLS,
+            .fp_len = 21,
+            .fp_label = bt->bt_ll,
+            .fp_eos = MPLS_EOS,
+            .fp_payload_proto = DPO_PROTO_BIER,
+        };
+        u32 mpls_fib_index;
+        dpo_id_t dpo = DPO_INVALID;
+
+        /*
+         * stack the entry on the forwarding chain prodcued by the
+         * path-list via the ECMP tables.
+         */
+        fib_path_list_contribute_forwarding(bt->bt_pl,
+                                            FIB_FORW_CHAIN_TYPE_BIER,
+                                            &dpo);
+
+        mpls_fib_index = fib_table_find(FIB_PROTOCOL_MPLS,
+                                        MPLS_FIB_DEFAULT_TABLE_ID);
+        bt->bt_lfei = fib_table_entry_special_dpo_add(mpls_fib_index,
+                                                      &pfx,
+                                                      FIB_SOURCE_BIER,
+                                                      FIB_ENTRY_FLAG_EXCLUSIVE,
+                                                      &dpo);
+        dpo_reset(&dpo);
+    }
+}
+
+static bier_table_t *
+bier_table_find (const bier_table_id_t *bti)
+{
+    uword *p;
+    u32 key;
+
+    key = bier_table_mk_key(bti);
+
+    p = hash_get(bier_tables_by_key, key);
+
+    if (NULL != p)
+    {
+        return (bier_table_get(p[0]));
+    }
+
+    return (NULL);
+}
+
+static bier_table_t *
+bier_table_mk_ecmp (index_t bti)
+{
+    fib_route_path_t *rpaths;
+    fib_node_index_t pli;
+    bier_table_t *bt;
+    int ii;
+
+    rpaths = NULL;
+    bt = bier_table_get(bti);
+
+    vec_validate(rpaths, BIER_N_ECMP_TABLES-1);
+
+    vec_foreach_index(ii, rpaths)
+    {
+        rpaths[ii].frp_bier_tbl = bt->bt_id;
+        rpaths[ii].frp_bier_tbl.bti_ecmp = ii;
+        rpaths[ii].frp_flags = FIB_ROUTE_PATH_BIER_TABLE;
+    }
+
+    /*
+     * no oppotunity to share, this the resolving ECMP tables are unique
+     * to this table.
+     * no need to be a child of the path list, we can do nothing with any
+     * notifications it would generate [not that it will].
+     */
+    pli = fib_path_list_create(FIB_PATH_LIST_FLAG_NO_URPF, rpaths);
+    fib_path_list_lock(pli);
+
+    /*
+     * constructing the path-list will have created many more BIER tables,
+     * so this main table will no doubt have re-alloc.
+     */
+    bt = bier_table_get(bti);
+    bt->bt_pl = pli;
+
+    vec_free(rpaths);
+
+    return (bt);
+}
+
+index_t
+bier_table_add_or_lock (const bier_table_id_t *btid,
+                        mpls_label_t local_label)
+{
+    bier_table_t *bt;
+    index_t bti;
+
+    bt = bier_table_find(btid);
+
+    if (NULL != bt) {
+        /*
+         * modify an existing table.
+         * change the lfib entry to the new local label
+         */
+        if (bier_table_is_main(bt) &&
+            (local_label != MPLS_LABEL_INVALID))
+        {
+            bier_table_rm_lfib(bt);
+
+            bt->bt_ll = local_label;
+            bier_table_mk_lfib(bt);
+        }
+        bti = bier_table_get_index(bt);
+    }
+    else
+    {
+        /*
+         * add a new table
+         */
+        u32 key;
+
+        key = bier_table_mk_key(btid);
+
+        pool_get_aligned(bier_table_pool, bt, CLIB_CACHE_LINE_BYTES);
+        bier_table_init(bt, btid, local_label);
+
+        hash_set(bier_tables_by_key, key, bier_table_get_index(bt));
+        bti = bier_table_get_index(bt);
+
+        if (bier_table_is_main(bt))
+        {
+            bt = bier_table_mk_ecmp(bti);
+            bier_table_mk_lfib(bt);
+        }
+    }
+
+    bier_table_lock_i(bt);
+
+    return (bti);
+}
+
+index_t
+bier_table_ecmp_create_and_lock (const bier_table_id_t *btid)
+{
+    return (bier_table_add_or_lock(btid, MPLS_LABEL_INVALID));
+}
+
+void
+bier_table_ecmp_unlock (index_t bti)
+{
+    bier_table_unlock_i(bier_table_get(bti));
+}
+
+static void
+bier_table_dpo_lock (dpo_id_t *dpo)
+{
+}
+
+static void
+bier_table_dpo_unlock (dpo_id_t *dpo)
+{
+}
+
+static void
+bier_table_dpo_mem_show (void)
+{
+    fib_show_memory_usage("BIER-table",
+                          pool_elts(bier_table_pool),
+                          pool_len(bier_table_pool),
+                          sizeof(bier_table_t));
+}
+static u8 *
+format_bier_table_dpo (u8 *s, va_list *ap)
+{
+    index_t bti = va_arg(*ap, index_t);
+    bier_table_t *bt;
+
+    bt = bier_table_get(bti);
+
+    return (format(s, "[%U]", format_bier_table_id, &bt->bt_id));
+}
+
+const static dpo_vft_t bier_table_dpo_vft = {
+    .dv_lock = bier_table_dpo_lock,
+    .dv_unlock = bier_table_dpo_unlock,
+    .dv_format = format_bier_table_dpo,
+    .dv_mem_show = bier_table_dpo_mem_show,
+};
+
+const static char *const bier_table_mpls_nodes[] =
+{
+    "bier-input"
+};
+const static char * const * const bier_table_nodes[DPO_PROTO_NUM] =
+{
+    [DPO_PROTO_BIER] = bier_table_mpls_nodes,
+};
+
+static clib_error_t *
+bier_table_module_init (vlib_main_t *vm)
+{
+    dpo_register(DPO_BIER_TABLE, &bier_table_dpo_vft, bier_table_nodes);
+
+    return (NULL);
+}
+
+VLIB_INIT_FUNCTION (bier_table_module_init);
+
+const bier_table_id_t *
+bier_table_get_id (index_t bti)
+{
+    bier_table_t *bt;
+
+    bt = bier_table_get(bti);
+
+    return (&bt->bt_id);
+}
+
+static void
+bier_table_insert (bier_table_t *bt,
+                   bier_bp_t bp,
+                   index_t bei)
+{
+    bt->bt_entries[BIER_BP_TO_INDEX(bp)] = bei;
+}
+
+static void
+bier_table_remove (bier_table_t *bt,
+                   bier_bp_t bp)
+{
+    bt->bt_entries[BIER_BP_TO_INDEX(bp)] = INDEX_INVALID;
+}
+
+void
+bier_table_route_add (const bier_table_id_t *btid,
+                      bier_bp_t bp,
+                      fib_route_path_t *brps)
+{
+    index_t bfmi, bti, bei, *bfmip, *bfmis = NULL;
+    fib_route_path_t *brp;
+    bier_table_t *bt;
+
+    bt = bier_table_find(btid);
+
+    if (NULL == bt) {
+        return;
+    }
+
+    bti = bier_table_get_index(bt);
+    bei = bier_table_lookup(bt, bp);
+
+    /*
+     * set the FIB index in the path to the BIER table index
+     */
+    vec_foreach(brp, brps)
+    {
+        bier_fmask_id_t fmid = {
+            .bfmi_nh = brp->frp_addr,
+            .bfmi_hdr_type = BIER_HDR_O_MPLS,
+        };
+        bfmi = bier_fmask_db_find_or_create_and_lock(bier_table_get_index(bt),
+                                                     &fmid,
+                                                     brp);
+
+        brp->frp_bier_fib_index = bti;
+        vec_add1(bfmis, bfmi);
+    }
+
+    if (INDEX_INVALID == bei)
+    {
+        bei = bier_entry_create(bti, bp);
+        bier_table_insert(bt, bp, bei);
+    }
+    bier_entry_path_add(bei, brps);
+
+    vec_foreach(bfmip, bfmis)
+    {
+        bier_fmask_unlock(*bfmip);
+    }
+    vec_free(bfmis);
+}
+
+void
+bier_table_route_remove (const bier_table_id_t *bti,
+                         bier_bp_t bp,
+                         fib_route_path_t *brps)
+{
+    fib_route_path_t *brp = NULL;
+    bier_table_t *bt;
+    index_t bei;
+
+    bt = bier_table_find(bti);
+
+    if (NULL == bt) {
+        return;
+    }
+
+    bei = bier_table_lookup(bt, bp);
+
+    if (INDEX_INVALID == bei)
+    {
+        /* no such entry */
+        return;
+    }
+
+    vec_foreach(brp, brps)
+    {
+        brp->frp_bier_fib_index = bier_table_get_index(bt);
+    }
+
+    if (0 == bier_entry_path_remove(bei, brps))
+    {
+        /* 0 remaining paths */
+        bier_table_remove(bt, bp);
+        bier_entry_delete(bei);
+    }
+}
+
+void
+bier_table_contribute_forwarding (index_t bti,
+                                  dpo_id_t *dpo)
+{
+    bier_table_t *bt;
+
+    bt = bier_table_get(bti);
+
+    if (BIER_ECMP_TABLE_ID_MAIN == bt->bt_id.bti_ecmp)
+    {
+        /*
+         * return the load-balance for the ECMP tables
+         */
+        fib_path_list_contribute_forwarding(bt->bt_pl,
+                                            FIB_FORW_CHAIN_TYPE_BIER,
+                                            dpo);
+    }
+    else
+    {
+        dpo_set(dpo, DPO_BIER_TABLE, DPO_PROTO_BIER, bti);
+    }
+}
+
+typedef struct bier_table_ecmp_walk_ctx_t_
+{
+    bier_table_ecmp_walk_fn_t fn;
+    void *ctx;
+} bier_table_ecmp_walk_ctx_t;
+
+static fib_path_list_walk_rc_t
+bier_table_ecmp_walk_path_list (fib_node_index_t pl_index,
+                                fib_node_index_t path_index,
+                                void *arg)
+{
+    bier_table_ecmp_walk_ctx_t *ctx = arg;
+
+    ctx->fn(fib_path_get_resolving_index(path_index), ctx->ctx);
+    /* continue */
+    return (FIB_PATH_LIST_WALK_CONTINUE);
+}
+
+void
+bier_table_ecmp_walk (index_t bti,
+                      bier_table_ecmp_walk_fn_t fn,
+                      void *ctx)
+{
+    bier_table_ecmp_walk_ctx_t ewc = {
+        .fn = fn,
+        .ctx = ctx,
+    };
+    bier_table_t *bt;
+
+    bt = bier_table_get(bti);
+
+    fib_path_list_walk(bt->bt_pl,
+                       bier_table_ecmp_walk_path_list,
+                       &ewc);
+}
+
+void
+bier_table_ecmp_set_fmask (index_t bti,
+                           bier_bp_t bp,
+                           index_t bfmi)
+{
+    bier_table_t *bt;
+
+    bt = bier_table_get(bti);
+
+    /*
+     * we hold a lock for fmasks in the table
+     */
+    bier_fmask_lock(bfmi);
+    bier_fmask_unlock(bt->bt_fmasks[BIER_BP_TO_INDEX(bp)]);
+
+    bt->bt_fmasks[BIER_BP_TO_INDEX(bp)] = bfmi;
+}
+
+u8 *
+format_bier_table_entry (u8 *s, va_list *ap)
+{
+    index_t bti = va_arg(*ap, index_t);
+    bier_bp_t bp = va_arg(*ap, bier_bp_t);
+    bier_table_t *bt;
+    bt = bier_table_get(bti);
+
+    if (bier_table_is_main(bt))
+    {
+        index_t bei;
+
+        bei = bier_table_lookup(bier_table_get(bti), bp);
+
+        if (INDEX_INVALID != bei)
+        {
+            s = format(s, "%U", format_bier_entry, bei,
+                       BIER_SHOW_DETAIL);
+        }
+    }
+    else
+    {
+        index_t bfmi;
+
+        bfmi = bier_table_fwd_lookup(bier_table_get(bti), bp);
+
+        if (INDEX_INVALID != bfmi)
+        {
+            s = format(s, "%U", format_bier_fmask, bfmi,
+                       BIER_SHOW_DETAIL);
+        }
+    }
+    return (s);
+}
+
+u8 *
+format_bier_table (u8 *s, va_list *ap)
+{
+    index_t bti = va_arg(*ap, index_t);
+    bier_show_flags_t flags = va_arg(*ap, bier_show_flags_t);
+    bier_table_t *bt;
+
+    if (pool_is_free_index(bier_table_pool, bti))
+    {
+        return (format(s, "No BIER f-mask %d", bti));
+    }
+
+    bt = bier_table_get(bti);
+
+    s = format(s, "[@%d] bier-table:[%U local-label:%U]",
+               bti,
+               format_bier_table_id, &bt->bt_id,
+               format_mpls_unicast_label, bt->bt_ll);
+
+    if (flags & BIER_SHOW_DETAIL)
+    {
+        s = format(s, " locks:%d", bt->bt_locks);
+    }
+    s = format(s, "]");
+
+    if (flags & BIER_SHOW_DETAIL)
+    {
+        if (bier_table_is_main(bt))
+        {
+            index_t *bei;
+
+            vec_foreach (bei, bt->bt_entries)
+            {
+                if (INDEX_INVALID != *bei)
+                {
+                    s = format(s, "\n%U", format_bier_entry, *bei, 2);
+                }
+            }
+        }
+        else
+        {
+            u32 ii;
+
+            vec_foreach_index (ii, bt->bt_fmasks)
+            {
+                if (INDEX_INVALID != bt->bt_fmasks[ii])
+                {
+                    s = format(s, "\n bp:%d\n %U", ii,
+                               format_bier_fmask, bt->bt_fmasks[ii], 2);
+                }
+            }
+        }
+    }
+
+    return (s);
+}
+
+void
+bier_table_show_all (vlib_main_t * vm,
+                     bier_show_flags_t flags)
+{
+    if (!pool_elts(bier_table_pool))
+    {
+        vlib_cli_output (vm, "No BIER tables");
+    }
+    else
+    {
+        int ii;
+
+        pool_foreach_index(ii, bier_table_pool,
+        ({
+            vlib_cli_output (vm, "%U", format_bier_table, ii, flags);
+        }));
+    }
+}
+
+void
+bier_tables_walk (bier_tables_walk_fn_t fn,
+                  void *ctx)
+{
+    ASSERT(0);
+}
+
+
+void
+bier_table_walk (const bier_table_id_t *bti,
+                 bier_table_walk_fn_t fn,
+                 void *ctx)
+{
+    bier_table_t *bt;
+    bier_entry_t *be;
+    index_t *bei;
+
+    bt = bier_table_find(bti);
+
+    if (NULL == bt)
+    {
+        return;
+    }
+
+    vec_foreach (bei, bt->bt_entries)
+    {
+        if (INDEX_INVALID != *bei)
+        {
+            be = bier_entry_get(*bei);
+
+            fn(bt, be, ctx);
+        }
+    }
+}
diff --git a/src/vnet/bier/bier_table.h b/src/vnet/bier/bier_table.h
new file mode 100644
index 00000000000..a22e2e335b2
--- /dev/null
+++ b/src/vnet/bier/bier_table.h
@@ -0,0 +1,172 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef __BIER_TABLE_H__
+#define __BIER_TABLE_H__
+
+#include <vlib/vlib.h>
+#include <vnet/fib/fib_types.h>
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_entry.h>
+
+#include <vnet/dpo/dpo.h>
+
+/**
+ * Forward declarations
+ */
+struct bier_route_update_t_;
+
+/**
+ * A BIER Table is the bit-indexed forwarding table.
+ * Each entry (bit-position) represents one destination, and its reachability
+ *
+ * The number of entries in a table is thus the maximum supported
+ * bit-position. Since this is smal <4096, the table is a flat arry
+ */
+typedef struct bier_table_t_ {
+    /**
+     * Save the MPLS local label associated with the table
+     */
+    mpls_label_t bt_ll;
+
+    /**
+     * The path-list used for the ECMP-tables
+     */
+    fib_node_index_t bt_pl;
+
+    /**
+     * The index of the lfib entry created for this table.
+     * Only the EOS is required.
+     */
+    fib_node_index_t bt_lfei;
+
+    /**
+     * Number of locks on the table
+     */
+    u16 bt_locks;
+
+    /**
+     * Entries in the table
+     * This is a vector sized to the appropriate number of entries
+     * given the table's supported Bit-string length
+     */
+    index_t *bt_entries;
+
+    /**
+     * Everything before this declaration is unused in the switch path
+     */
+    CLIB_CACHE_LINE_ALIGN_MARK(cacheline0);
+
+    /**
+     * The identity/key or the table. we need the hdr_len in the data-path
+     */
+    bier_table_id_t bt_id;
+
+    /**
+     * f-masks in the ECMP table
+     * This is a vector sized to the appropriate number of entries
+     * given the table's supported Bit-string length.
+     * In the ECMP table the LB choice has been pre-resolved, so each entry
+     * links to just one f-mask, i.e. there is a 1:1 mapping of bit-position to
+     * fmask. For efficient forwarding we collapse the fmasks up to the table.
+     */
+    index_t *bt_fmasks;
+} bier_table_t;
+
+STATIC_ASSERT((sizeof(bier_table_t) <= 2*CLIB_CACHE_LINE_BYTES),
+              "BIER table fits on 2 cache lines");
+
+extern index_t bier_table_add_or_lock(const bier_table_id_t *id,
+                                      mpls_label_t ll);
+extern void bier_table_unlock(const bier_table_id_t *id);
+
+extern void bier_table_route_add(const bier_table_id_t *bti,
+                                 bier_bp_t bp,
+                                 fib_route_path_t *brp);
+extern void bier_table_route_remove(const bier_table_id_t *bti,
+                                    bier_bp_t bp,
+                                    fib_route_path_t *brp);
+
+extern void bier_table_show_all(vlib_main_t * vm,
+                                bier_show_flags_t flags);
+
+extern const bier_table_id_t *bier_table_get_id(index_t bti);
+
+extern u8 *format_bier_table (u8 *s, va_list *args);
+extern u8 *format_bier_table_entry (u8 *s, va_list *args);
+
+extern index_t bier_table_ecmp_create_and_lock(const bier_table_id_t *id);
+extern void bier_table_ecmp_unlock(index_t bti);
+extern void bier_table_ecmp_set_fmask(index_t bti,
+                                      bier_bp_t bp,
+                                      index_t bfmi);
+
+extern void bier_table_contribute_forwarding(index_t bti,
+                                             dpo_id_t *dpo);
+
+/**
+ * Types and functions to walk the ECMP tables of a main table
+ */
+typedef void (*bier_table_ecmp_walk_fn_t)(index_t btei,
+                                          void *ctx);
+extern void bier_table_ecmp_walk(index_t bti,
+                                 bier_table_ecmp_walk_fn_t fn,
+                                 void *ctx);
+extern int bier_table_is_main (const bier_table_t *bt);
+
+/**
+ * Types and functions to walk all the BIER Tables
+ */
+typedef void (*bier_tables_walk_fn_t)(const bier_table_t *bt,
+                                      void *ctx);
+extern void bier_tables_walk(bier_tables_walk_fn_t fn,
+                             void *ctx);
+
+/**
+ * Types and functions to walk all the entries in one BIER Table
+ */
+typedef void (*bier_table_walk_fn_t)(const bier_table_t *bt,
+                                     const bier_entry_t *be,
+                                     void *ctx);
+extern void bier_table_walk(const bier_table_id_t *id,
+                            bier_table_walk_fn_t fn,
+                            void *ctx);
+
+/*
+ * provided for fast data plane access.
+ */
+extern bier_table_t *bier_table_pool;
+
+static inline bier_table_t *
+bier_table_get (index_t bti)
+{
+    return (pool_elt_at_index(bier_table_pool, bti));
+}
+
+static inline const index_t
+bier_table_lookup (const bier_table_t *bt,
+                   bier_bp_t bp)
+{
+    return (bt->bt_entries[BIER_BP_TO_INDEX(bp)]);
+}
+
+static inline const index_t
+bier_table_fwd_lookup (const bier_table_t *bt,
+                       bier_bp_t bp)
+{
+    return (bt->bt_fmasks[BIER_BP_TO_INDEX(bp)]);
+}
+
+#endif
diff --git a/src/vnet/bier/bier_test.c b/src/vnet/bier/bier_test.c
new file mode 100644
index 00000000000..6119410edfa
--- /dev/null
+++ b/src/vnet/bier/bier_test.c
@@ -0,0 +1,951 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+
+#include <vnet/mpls/mpls.h>
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_entry.h>
+#include <vnet/bier/bier_fmask.h>
+#include <vnet/bier/bier_bit_string.h>
+#include <vnet/bier/bier_imp.h>
+#include <vnet/bier/bier_disp_table.h>
+#include <vnet/bier/bier_disp_entry.h>
+#include <vnet/fib/fib_entry.h>
+#include <vnet/fib/fib_table.h>
+#include <vnet/fib/mpls_fib.h>
+#include <vnet/dpo/load_balance.h>
+#include <vnet/dpo/drop_dpo.h>
+#include <vnet/dpo/lookup_dpo.h>
+#include <vnet/mfib/mfib_table.h>
+
+#include <vnet/fib/fib_test.h>
+
+/*
+ * Add debugs for passing tests
+ */
+static int bier_test_do_debug;
+
+#define BIER_TEST_I(_cond, _comment, _args...)			\
+({								\
+    int _evald = (_cond);					\
+    if (!(_evald)) {						\
+        fformat(stderr, "FAIL:%d: " _comment "\n",		\
+                __LINE__, ##_args);				\
+    } else {							\
+        if (bier_test_do_debug)                                 \
+            fformat(stderr, "PASS:%d: " _comment "\n",          \
+                    __LINE__, ##_args);				\
+    }								\
+    _evald;							\
+})
+#define BIER_TEST(_cond, _comment, _args...)			\
+{								\
+    if (!BIER_TEST_I(_cond, _comment, ##_args)) {		\
+        return 1;                                               \
+        ASSERT(!("FAIL: " _comment));				\
+    }								\
+}
+
+/**
+ * A 'i'm not fussed is this is not efficient' store of test data
+ */
+typedef struct test_main_t_ {
+    /**
+     * HW if indicies
+     */
+    u32 hw_if_indicies[4];
+    /**
+     * HW interfaces
+     */
+    vnet_hw_interface_t * hw[4];
+
+} test_main_t;
+static test_main_t test_main;
+
+/* fake ethernet device class, distinct from "fake-ethX" */
+static u8 * format_test_interface_name (u8 * s, va_list * args)
+{
+  u32 dev_instance = va_arg (*args, u32);
+  return format (s, "test-eth%d", dev_instance);
+}
+
+static uword dummy_interface_tx (vlib_main_t * vm,
+                                 vlib_node_runtime_t * node,
+                                 vlib_frame_t * frame)
+{
+  clib_warning ("you shouldn't be here, leaking buffers...");
+  return frame->n_vectors;
+}
+
+VNET_DEVICE_CLASS (test_interface_device_class,static) = {
+  .name = "Test interface",
+  .format_device_name = format_test_interface_name,
+  .tx_function = dummy_interface_tx,
+};
+
+static u8 *hw_address;
+
+static int
+bier_test_mk_intf (u32 ninterfaces)
+{
+    clib_error_t * error = NULL;
+    test_main_t *tm = &test_main;
+    u8 byte;
+    u32 i;
+
+    ASSERT(ninterfaces <= ARRAY_LEN(tm->hw_if_indicies));
+
+    for (i=0; i<6; i++)
+    {
+        byte = 0xd0+i;
+        vec_add1(hw_address, byte);
+    }
+
+    for (i = 0; i < ninterfaces; i++)
+    {
+        hw_address[5] = i;
+
+        error = ethernet_register_interface(vnet_get_main(),
+                                            test_interface_device_class.index,
+                                            i /* instance */,
+                                            hw_address,
+                                            &tm->hw_if_indicies[i],
+                                            /* flag change */ 0);
+
+        BIER_TEST((NULL == error), "ADD interface %d", i);
+
+        tm->hw[i] = vnet_get_hw_interface(vnet_get_main(),
+                                          tm->hw_if_indicies[i]);
+        vec_validate (ip4_main.fib_index_by_sw_if_index, tm->hw[i]->sw_if_index);
+        vec_validate (ip6_main.fib_index_by_sw_if_index, tm->hw[i]->sw_if_index);
+        ip4_main.fib_index_by_sw_if_index[tm->hw[i]->sw_if_index] = 0;
+        ip6_main.fib_index_by_sw_if_index[tm->hw[i]->sw_if_index] = 0;
+        error = vnet_sw_interface_set_flags(vnet_get_main(),
+                                            tm->hw[i]->sw_if_index,
+                                            VNET_SW_INTERFACE_FLAG_ADMIN_UP);
+        BIER_TEST((NULL == error), "UP interface %d", i);
+    }
+    /*
+     * re-eval after the inevitable realloc
+     */
+    for (i = 0; i < ninterfaces; i++)
+    {
+        tm->hw[i] = vnet_get_hw_interface(vnet_get_main(),
+                                          tm->hw_if_indicies[i]);
+    }
+
+    return (0);
+}
+
+#define BIER_TEST_LB(_cond, _comment, _args...)			\
+{								\
+    if (!BIER_TEST_I(_cond, _comment, ##_args)) {		\
+        return (0);						\
+    }								\
+}
+
+static int
+bier_test_validate_entry (index_t bei,
+                          u16 n_buckets,
+                          ...)
+{
+    dpo_id_t dpo = DPO_INVALID;
+    const load_balance_t *lb;
+    va_list ap;
+    int res;
+
+    va_start(ap, n_buckets);
+
+    bier_entry_contribute_forwarding(bei, &dpo);
+
+    BIER_TEST_LB((DPO_LOAD_BALANCE == dpo.dpoi_type),
+                 "Entry links to %U",
+                 format_dpo_type, dpo.dpoi_type);
+
+    lb = load_balance_get(dpo.dpoi_index);
+    res = fib_test_validate_lb_v(lb, n_buckets, &ap);
+
+    dpo_reset(&dpo);
+
+    va_end(ap);
+
+    return (res);
+}
+
+static int
+bier_test_mpls_spf (void)
+{
+    fib_node_index_t lfei, fei, bti;
+    u32 mpls_fib_index;
+    test_main_t *tm;
+    int lb_count;
+
+    lb_count = pool_elts(load_balance_pool);
+    tm = &test_main;
+#define N_BIER_ECMP_TABLES 16
+    int ii;
+
+    /*
+     * Add the BIER Main table
+     */
+    const bier_table_id_t bt_0_0_0_256 = {
+        .bti_set = 0,
+        .bti_sub_domain = 0,
+        .bti_hdr_len = BIER_HDR_LEN_256,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+
+    bti = bier_table_add_or_lock(&bt_0_0_0_256, 1600);
+
+    fib_test_lb_bucket_t l_o_bt[N_BIER_ECMP_TABLES];
+    bier_table_id_t bt_ecmp_0_0_0_256 = bt_0_0_0_256;
+
+    for (ii = 0; ii < N_BIER_ECMP_TABLES; ii++)
+    {
+        bt_ecmp_0_0_0_256.bti_ecmp = ii;
+
+        l_o_bt[ii].type = FT_LB_BIER_TABLE;
+        l_o_bt[ii].bier.table =
+            bier_table_ecmp_create_and_lock(&bt_ecmp_0_0_0_256);
+    };
+    const fib_prefix_t pfx_1600_neos = {
+        .fp_len = 21,
+        .fp_proto = FIB_PROTOCOL_MPLS,
+        .fp_label = 1600,
+        .fp_eos = MPLS_NON_EOS,
+        .fp_payload_proto = DPO_PROTO_BIER,
+    };
+    const fib_prefix_t pfx_1600_eos = {
+        .fp_len = 21,
+        .fp_proto = FIB_PROTOCOL_MPLS,
+        .fp_label = 1600,
+        .fp_eos = MPLS_EOS,
+        .fp_payload_proto = DPO_PROTO_BIER,
+    };
+
+    mpls_fib_index = fib_table_find(FIB_PROTOCOL_MPLS,
+                                    MPLS_FIB_DEFAULT_TABLE_ID);
+
+    lfei = fib_table_lookup(mpls_fib_index, &pfx_1600_neos);
+    BIER_TEST(FIB_NODE_INDEX_INVALID == lfei, "1600/0 is not present");
+
+    lfei = fib_table_lookup(mpls_fib_index, &pfx_1600_eos);
+    BIER_TEST(fib_test_validate_entry(lfei, FIB_FORW_CHAIN_TYPE_MPLS_EOS,
+                                      16,
+                                      &l_o_bt[0],
+                                      &l_o_bt[1],
+                                      &l_o_bt[2],
+                                      &l_o_bt[3],
+                                      &l_o_bt[4],
+                                      &l_o_bt[5],
+                                      &l_o_bt[6],
+                                      &l_o_bt[7],
+                                      &l_o_bt[8],
+                                      &l_o_bt[9],
+                                      &l_o_bt[10],
+                                      &l_o_bt[11],
+                                      &l_o_bt[12],
+                                      &l_o_bt[13],
+                                      &l_o_bt[14],
+                                      &l_o_bt[15]),
+              "1600/1 LB stacks on BIER table %d", bti);
+
+    /*
+     * modify the table's local label - keep the lock count accurate
+     */
+    const fib_prefix_t pfx_1601_eos = {
+        .fp_len = 21,
+        .fp_proto = FIB_PROTOCOL_MPLS,
+        .fp_label = 1601,
+        .fp_eos = MPLS_EOS,
+        .fp_payload_proto = DPO_PROTO_BIER,
+    };
+    bti = bier_table_add_or_lock(&bt_0_0_0_256, 1601);
+    bier_table_unlock(&bt_0_0_0_256);
+
+    lfei = fib_table_lookup(mpls_fib_index, &pfx_1600_eos);
+    BIER_TEST(FIB_NODE_INDEX_INVALID == lfei, "1600/1 is deleted");
+
+    lfei = fib_table_lookup(mpls_fib_index, &pfx_1601_eos);
+    BIER_TEST(fib_test_validate_entry(lfei, FIB_FORW_CHAIN_TYPE_MPLS_EOS,
+                                      16,
+                                      &l_o_bt[0],
+                                      &l_o_bt[1],
+                                      &l_o_bt[2],
+                                      &l_o_bt[3],
+                                      &l_o_bt[4],
+                                      &l_o_bt[5],
+                                      &l_o_bt[6],
+                                      &l_o_bt[7],
+                                      &l_o_bt[8],
+                                      &l_o_bt[9],
+                                      &l_o_bt[10],
+                                      &l_o_bt[11],
+                                      &l_o_bt[12],
+                                      &l_o_bt[13],
+                                      &l_o_bt[14],
+                                      &l_o_bt[15]),
+              "1601/1 LB stacks on BIER table %d", bti);
+
+    /*
+     * add a route to the table. the via IP route does not exist.
+     */
+    const ip46_address_t nh_1_1_1_1 = {
+        .ip4 = {
+            .as_u32 = clib_host_to_net_u32(0x01010101),
+        },
+    };
+    fib_route_path_t *paths_1_1_1_1 = NULL;
+    fib_route_path_t path_1_1_1_1 = {
+        .frp_addr = nh_1_1_1_1,
+        .frp_bier_fib_index = bti,
+        .frp_flags = FIB_ROUTE_PATH_BIER_FMASK,
+    };
+    vec_add1(path_1_1_1_1.frp_label_stack, 500);
+    vec_add1(paths_1_1_1_1, path_1_1_1_1);
+    const fib_prefix_t pfx_1_1_1_1_s_32 = {
+        .fp_addr = nh_1_1_1_1,
+        .fp_len = 32,
+        .fp_proto = FIB_PROTOCOL_IP4,
+    };
+    const bier_fmask_id_t bfm_id_1_1_1_1 = {
+        .bfmi_hdr_type = BIER_HDR_O_MPLS,
+        .bfmi_nh = nh_1_1_1_1,
+    };
+    index_t bei_1;
+
+    bier_table_route_add(&bt_0_0_0_256, 1, paths_1_1_1_1);
+    bei_1 = bier_table_lookup(bier_table_get(bti), 1);
+
+    BIER_TEST((INDEX_INVALID != bei_1), "BP:1 present");
+
+    /*
+     * the newly created fmask should stack on the non-eos chain
+     * of the via-fib-entry
+     */
+    dpo_id_t neos_dpo_1_1_1_1 = DPO_INVALID;
+    bier_fmask_t *bfm_1_1_1_1;
+    index_t bfmi_1_1_1_1;
+
+    fei = fib_table_lookup_exact_match(0, &pfx_1_1_1_1_s_32);
+    fib_entry_contribute_forwarding(fei,
+                                    FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS,
+                                    &neos_dpo_1_1_1_1);
+
+    bfmi_1_1_1_1 = bier_fmask_db_find(bti, &bfm_id_1_1_1_1);
+    bfm_1_1_1_1 = bier_fmask_get(bfmi_1_1_1_1);
+
+    BIER_TEST(!dpo_cmp(&neos_dpo_1_1_1_1, &bfm_1_1_1_1->bfm_dpo),
+              "Fmask via 1.1.1.1 stacks on neos from 1.1.1.1/32");
+
+    /*
+     * and that n-eos LB at this stage is a drop..
+     */
+    const fib_test_lb_bucket_t bucket_drop = {
+        .type = FT_LB_DROP,
+    };
+    BIER_TEST(fib_test_validate_lb(&neos_dpo_1_1_1_1, 1, &bucket_drop),
+             "1.1.1.1/32 n-eos LB 1 buckets via: DROP");
+
+    /*
+     * The BIER entry should stack on the forwarding chain of the fmask
+     */
+    const fib_test_lb_bucket_t dpo_o_bfm_1_1_1_1 = {
+        .type = FT_LB_BIER_FMASK,
+        .bier = {
+            .fmask = bfmi_1_1_1_1,
+        },
+    };
+    BIER_TEST(bier_test_validate_entry(bei_1, 1, &bucket_drop),
+              "BP:1  stacks on bier drop");
+
+    /*
+     * give 1.1.1.1/32 a path and hence a interesting n-eos chain
+     */
+    ip46_address_t nh_10_10_10_1 = {
+        .ip4 = {
+            .as_u32 = clib_host_to_net_u32(0x0a0a0a01),
+        },
+    };
+    adj_index_t ai_mpls_10_10_10_1;
+    ai_mpls_10_10_10_1 = adj_nbr_add_or_lock(FIB_PROTOCOL_IP4,
+                                             VNET_LINK_MPLS,
+                                             &nh_10_10_10_1,
+                                             tm->hw[0]->sw_if_index);
+
+    fib_test_lb_bucket_t bucket_neos_99_via_10_10_10_1 = {
+        .type = FT_LB_LABEL_O_ADJ,
+        .label_o_adj = {
+            .label = 99,
+            .eos = MPLS_NON_EOS,
+            .adj = ai_mpls_10_10_10_1,
+            .ttl = 255,
+        },
+    };
+    mpls_label_t *out_lbl_99 = NULL;
+    vec_add1(out_lbl_99, 99);
+
+    fei = fib_table_entry_update_one_path(0,
+                                          &pfx_1_1_1_1_s_32,
+                                          FIB_SOURCE_API,
+                                          FIB_ENTRY_FLAG_NONE,
+                                          DPO_PROTO_IP4,
+                                          &nh_10_10_10_1,
+                                          tm->hw[0]->sw_if_index,
+                                          ~0, // invalid fib index
+                                          1,
+                                          out_lbl_99,
+                                          FIB_ROUTE_PATH_FLAG_NONE);
+    fib_entry_contribute_forwarding(fei,
+                                    FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS,
+                                    &neos_dpo_1_1_1_1);
+    BIER_TEST(fib_test_validate_lb(&neos_dpo_1_1_1_1, 1,
+                                   &bucket_neos_99_via_10_10_10_1),
+              "1.1.1.1/32 n-eos LB 1 buckets via: 99 + 10.10.10.1");
+    BIER_TEST(!dpo_cmp(&neos_dpo_1_1_1_1,
+                       &bfm_1_1_1_1->bfm_dpo),
+              "Fmask via 1.1.1.1 stacks on updated non-eos of 1.1.1.1/32");
+    BIER_TEST(bier_test_validate_entry(bei_1, 1, &dpo_o_bfm_1_1_1_1),
+              "BP:1  stacks on fmask 1.1.1.1");
+
+    /*
+     * add another path to the via entry.
+     * this makes the via-entry instantiate a new load-balance with
+     * 2 buckets. and the back-walk to the BIER entry will need to
+     * re-stack on it.
+     */
+    ip46_address_t nh_10_10_10_2 = {
+        .ip4 = {
+            .as_u32 = clib_host_to_net_u32(0x0a0a0a02),
+        },
+    };
+    adj_index_t ai_mpls_10_10_10_2;
+
+    ai_mpls_10_10_10_2 = adj_nbr_add_or_lock(FIB_PROTOCOL_IP4,
+                                             VNET_LINK_MPLS,
+                                             &nh_10_10_10_2,
+                                             tm->hw[0]->sw_if_index);
+
+    fib_test_lb_bucket_t bucket_neos_100_via_10_10_10_2 = {
+        .type = FT_LB_LABEL_O_ADJ,
+        .label_o_adj = {
+            .label = 100,
+            .eos = MPLS_NON_EOS,
+            .adj = ai_mpls_10_10_10_2,
+            .ttl = 255,
+        },
+    };
+    mpls_label_t *out_lbl_100 = NULL;
+    vec_add1(out_lbl_100, 100);
+
+    fei = fib_table_entry_path_add(0,
+                                   &pfx_1_1_1_1_s_32,
+                                   FIB_SOURCE_API,
+                                   FIB_ENTRY_FLAG_NONE,
+                                   DPO_PROTO_IP4,
+                                   &nh_10_10_10_2,
+                                   tm->hw[0]->sw_if_index,
+                                   ~0, // invalid fib index
+                                   1,
+                                   out_lbl_100,
+                                   FIB_ROUTE_PATH_FLAG_NONE);
+
+    fib_entry_contribute_forwarding(fei,
+                                    FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS,
+                                    &neos_dpo_1_1_1_1);
+    BIER_TEST(fib_test_validate_lb(&neos_dpo_1_1_1_1, 2,
+                                   &bucket_neos_99_via_10_10_10_1,
+                                   &bucket_neos_100_via_10_10_10_2),
+              "1.1.1.1/32 n-eos LB 2 buckets "
+              "via: 99 + 10.10.10.1, "
+              "via: 100 + 10.10.10.2");
+    BIER_TEST(!dpo_cmp(&neos_dpo_1_1_1_1,
+                       &bfm_1_1_1_1->bfm_dpo),
+              "Fmask via 1.1.1.1 stacks on updated non-eos of 1.1.1.1/32");
+
+    /*
+     * add another bier bit-position via the same next-hop
+     * since its the same next hop, the two bit-positions should link
+     * to the same fmask
+     */
+    index_t bei_2;
+
+    bier_table_route_add(&bt_0_0_0_256, 2, paths_1_1_1_1);
+    bei_2 = bier_table_lookup(bier_table_get(bti), 2);
+
+    BIER_TEST(bier_test_validate_entry(bei_2, 1, &dpo_o_bfm_1_1_1_1),
+              "BP:2 stacks on fmask 1.1.1.1");
+
+    /*
+     * now add a bit-position via a different next hop and expect to
+     * link via a different fmask
+     */
+    const ip46_address_t nh_1_1_1_2 = {
+        .ip4 = {
+            .as_u32 = clib_host_to_net_u32(0x01010102),
+        },
+    };
+    const fib_prefix_t pfx_1_1_1_2_s_32 = {
+        .fp_addr = nh_1_1_1_2,
+        .fp_len = 32,
+        .fp_proto = FIB_PROTOCOL_IP4,
+    };
+    fib_route_path_t *paths_1_1_1_2 = NULL, path_1_1_1_2 = {
+        .frp_addr = nh_1_1_1_2,
+        .frp_bier_fib_index = bti,
+        .frp_flags = FIB_ROUTE_PATH_BIER_FMASK,
+    };
+    vec_add1(path_1_1_1_2.frp_label_stack, 501);
+    vec_add1(paths_1_1_1_2, path_1_1_1_2);
+    const bier_fmask_id_t bfm_id_1_1_1_2 = {
+        .bfmi_hdr_type = BIER_HDR_O_MPLS,
+        .bfmi_nh = nh_1_1_1_2,
+    };
+    index_t bei_3;
+
+    mpls_label_t *out_lbl_101 = NULL;
+    vec_add1(out_lbl_101, 101);
+    fei = fib_table_entry_path_add(0,
+                                   &pfx_1_1_1_2_s_32,
+                                   FIB_SOURCE_API,
+                                   FIB_ENTRY_FLAG_NONE,
+                                   DPO_PROTO_IP4,
+                                   &nh_10_10_10_2,
+                                   tm->hw[0]->sw_if_index,
+                                   ~0, // invalid fib index
+                                   1,
+                                   out_lbl_101,
+                                   FIB_ROUTE_PATH_FLAG_NONE);
+    bier_table_route_add(&bt_0_0_0_256, 3, paths_1_1_1_2);
+    bei_3 = bier_table_lookup(bier_table_get(bti), 3);
+
+    BIER_TEST((INDEX_INVALID != bei_3), "BP:3 present");
+
+    /*
+     * the newly created fmask should stack on the non-eos chain
+     * of the via-fib-entry
+     */
+    dpo_id_t neos_dpo_1_1_1_2 = DPO_INVALID;
+    bier_fmask_t *bfm_1_1_1_2;
+    index_t bfmi_1_1_1_2;
+
+    fei = fib_table_lookup_exact_match(0, &pfx_1_1_1_2_s_32);
+    fib_entry_contribute_forwarding(fei,
+                                    FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS,
+                                    &neos_dpo_1_1_1_2);
+
+    bfmi_1_1_1_2 = bier_fmask_db_find(bti, &bfm_id_1_1_1_2);
+    bfm_1_1_1_2 = bier_fmask_get(bfmi_1_1_1_2);
+
+    BIER_TEST(!dpo_cmp(&neos_dpo_1_1_1_2,
+                       &bfm_1_1_1_2->bfm_dpo),
+              "Fmask via 1.1.1.2 stacks on non-eos of 1.1.1.2/32");
+
+    /*
+     * The BIER entry should stack on the forwarding chain of the fmask
+     */
+    const fib_test_lb_bucket_t dpo_o_bfm_1_1_1_2 = {
+        .type = FT_LB_BIER_FMASK,
+        .bier = {
+            .fmask = bfmi_1_1_1_2,
+        },
+    };
+    BIER_TEST(bier_test_validate_entry(bei_3, 1, &dpo_o_bfm_1_1_1_2),
+              "BP:3 stacks on fmask 1.1.1.2");
+
+    /*
+     * Load-balance BP:3 over both next-hops
+     */
+    bier_table_route_add(&bt_0_0_0_256, 3, paths_1_1_1_1);
+
+    BIER_TEST(bier_test_validate_entry(bei_3, 2,
+                                       &dpo_o_bfm_1_1_1_1,
+                                       &dpo_o_bfm_1_1_1_2),
+              "BP:3 stacks on fmask 1.1.1.2 & 1.1.1.1");
+
+    /*
+     * test that the ECMP choices for BP:3 have been spread over the
+     * ECMP tables
+     */
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[0].bier.table), 3) ==
+               bfmi_1_1_1_1),
+              "fwd lookup for BP:3 ECMP:0 is 1.1.1.1");
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[1].bier.table), 3) ==
+               bfmi_1_1_1_2),
+              "fwd lookup for BP:3 ECMP:1 is 1.1.1.2");
+
+    /*
+     * Withdraw one of the via FIB and thus bring down the fmask
+     * expect the bier0entry forwarding to remove this from the set
+     */
+    fib_table_entry_delete(0, &pfx_1_1_1_2_s_32, FIB_SOURCE_API);
+
+    BIER_TEST(bier_test_validate_entry(bei_3, 1,
+                                       &dpo_o_bfm_1_1_1_1),
+              "BP:3 post 1.1.1.2 removal stacks on fmask 1.1.1.1");
+
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[0].bier.table), 3) ==
+               bfmi_1_1_1_1),
+              "fwd lookup for BP:3 ECMP:0 is 1.1.1.1");
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[1].bier.table), 3) ==
+               bfmi_1_1_1_1),
+              "fwd lookup for BP:3 ECMP:1 is 1.1.1.1");
+
+    /*
+     * add the via back
+     */
+    out_lbl_101 = NULL;
+    vec_add1(out_lbl_101, 101);
+    fei = fib_table_entry_path_add(0,
+                                   &pfx_1_1_1_2_s_32,
+                                   FIB_SOURCE_API,
+                                   FIB_ENTRY_FLAG_NONE,
+                                   DPO_PROTO_IP4,
+                                   &nh_10_10_10_2,
+                                   tm->hw[0]->sw_if_index,
+                                   ~0, // invalid fib index
+                                   1,
+                                   out_lbl_101,
+                                   FIB_ROUTE_PATH_FLAG_NONE);
+    /* suspend so the update walk kicks int */
+    vlib_process_suspend(vlib_get_main(), 1e-5);
+
+    BIER_TEST(bier_test_validate_entry(bei_3, 2,
+                                       &dpo_o_bfm_1_1_1_1,
+                                       &dpo_o_bfm_1_1_1_2),
+              "BP:3 stacks on fmask 1.1.1.2 & 1.1.1.1");
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[0].bier.table), 3) ==
+               bfmi_1_1_1_1),
+              "fwd lookup for BP:3 ECMP:0 is 1.1.1.1");
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[1].bier.table), 3) ==
+               bfmi_1_1_1_2),
+              "fwd lookup for BP:3 ECMP:1 is 1.1.1.2");
+
+    /*
+     * remove the original 1.1.1.2 fmask from BP:3
+     */
+    bier_table_route_remove(&bt_0_0_0_256, 3, paths_1_1_1_2);
+    BIER_TEST(bier_test_validate_entry(bei_3, 1,
+                                       &dpo_o_bfm_1_1_1_1),
+              "BP:3 stacks on fmask 1.1.1.1");
+    /*
+     * test that the ECMP choices for BP:3 have been updated
+     */
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[0].bier.table), 3) ==
+               bfmi_1_1_1_1),
+              "fwd lookup for BP:3 ECMP:0 is 1.1.1.1");
+    BIER_TEST((bier_table_fwd_lookup(bier_table_get(l_o_bt[1].bier.table), 3) ==
+               bfmi_1_1_1_1),
+              "fwd lookup for BP:3 ECMP:1 is 1.1.1.1");
+
+    /*
+     * remove the routes added
+     */
+    bier_table_route_remove(&bt_0_0_0_256, 2, paths_1_1_1_1);
+    bier_table_route_remove(&bt_0_0_0_256, 3, paths_1_1_1_2);
+    bier_table_route_remove(&bt_0_0_0_256, 3, paths_1_1_1_1);
+    bier_table_route_remove(&bt_0_0_0_256, 1, paths_1_1_1_1);
+
+
+    /*
+     * delete the table
+     */
+    bier_table_unlock(&bt_0_0_0_256);
+
+    /*
+     * test resources are freed
+     */
+    for (ii = 0; ii < N_BIER_ECMP_TABLES; ii++)
+    {
+        bier_table_ecmp_unlock(l_o_bt[ii].bier.table);
+    };
+    BIER_TEST(0 == pool_elts(bier_table_pool), "BIER table pool empty");
+    BIER_TEST(0 == pool_elts(bier_fmask_pool), "BIER fmask pool empty");
+    BIER_TEST(0 == pool_elts(bier_entry_pool), "BIER entry pool empty");
+
+    adj_unlock(ai_mpls_10_10_10_1);
+    adj_unlock(ai_mpls_10_10_10_2);
+    dpo_reset(&neos_dpo_1_1_1_1);
+    dpo_reset(&neos_dpo_1_1_1_2);
+    fib_table_entry_delete(0, &pfx_1_1_1_1_s_32, FIB_SOURCE_API);
+    fib_table_entry_delete(0, &pfx_1_1_1_2_s_32, FIB_SOURCE_API);
+
+    /* +1 to account for the one time alloc'd drop LB in the MPLS fibs */
+    BIER_TEST(lb_count+1 == pool_elts(load_balance_pool),
+              "Load-balance resources freed ");
+    BIER_TEST((0 == adj_nbr_db_size()), "ADJ DB size is %d",
+             adj_nbr_db_size());
+
+    vec_free(paths_1_1_1_1);
+    vec_free(paths_1_1_1_2);
+
+    return (0);
+}
+
+static int
+bier_test_mpls_imp (void)
+{
+    fib_node_index_t bii;
+    /* test_main_t *tm; */
+
+    /* tm = &test_main; */
+
+    /*
+     * Add the BIER Main table
+     */
+    const bier_table_id_t bt_0_0_0_256 = {
+        .bti_set = 0,
+        .bti_sub_domain = 0,
+        .bti_hdr_len = BIER_HDR_LEN_256,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+
+    bier_table_add_or_lock(&bt_0_0_0_256, 1600);
+
+    /*
+     * A bit-string for imp 1.
+     */
+    bier_bit_string_t bbs_256;
+    u8 buckets[BIER_HDR_BUCKETS_256];
+    memset(buckets, 0x5, BIER_HDR_BUCKETS_256);
+
+    bier_bit_string_init(&bbs_256, BIER_HDR_LEN_256, buckets);
+
+    bii = bier_imp_add_or_lock(&bt_0_0_0_256, 1, &bbs_256);
+
+    /*
+     * An mfib entry that resolves via the BIER imposition
+     */
+    const mfib_prefix_t pfx_1_1_1_1_c_239_1_1_1 = {
+        .fp_len = 64,
+        .fp_proto = FIB_PROTOCOL_IP4,
+        .fp_grp_addr = {
+            .ip4.as_u32 = clib_host_to_net_u32(0xef010101),
+        },
+        .fp_src_addr = {
+            .ip4.as_u32 = clib_host_to_net_u32(0x01010101),
+        },
+    };
+    fib_route_path_t path_via_bier_imp_1 = {
+        .frp_proto = DPO_PROTO_BIER,
+        .frp_bier_imp = bii,
+        .frp_weight = 0,
+        .frp_flags = FIB_ROUTE_PATH_BIER_IMP,
+    };
+    mfib_table_entry_path_update(0, // default table
+                                 &pfx_1_1_1_1_c_239_1_1_1 ,
+                                 MFIB_SOURCE_API,
+                                 &path_via_bier_imp_1,
+                                 MFIB_ITF_FLAG_FORWARD);
+    mfib_table_entry_delete(0,
+                            &pfx_1_1_1_1_c_239_1_1_1 ,
+                            MFIB_SOURCE_API);
+
+    bier_imp_unlock(bii);
+    bier_table_unlock(&bt_0_0_0_256);
+
+    BIER_TEST(0 == pool_elts(bier_imp_pool),
+              "BIER imposition resources freed ");
+    BIER_TEST(0 == pool_elts(bier_table_pool),
+              "BIER table resources freed ");
+
+    return (0);
+}
+
+static int
+bier_test_mpls_disp (void)
+{
+    /* test_main_t *tm; */
+
+    /* tm = &test_main; */
+
+    /*
+     * Add the BIER Main table
+     */
+    const bier_table_id_t bt_0_0_0_256 = {
+        .bti_set = 0,
+        .bti_sub_domain = 0,
+        .bti_hdr_len = BIER_HDR_LEN_256,
+        .bti_type = BIER_TABLE_MPLS_SPF,
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+    index_t bti;
+
+    bti = bier_table_add_or_lock(&bt_0_0_0_256, 1600);
+
+    /*
+     * Add a BIER dispoition table
+     */
+    const u32 bier_disp_tbl_id = 1;
+    index_t bdti1;
+
+    bdti1 = bier_disp_table_add_or_lock(bier_disp_tbl_id);
+
+    /*
+     * add a bit-poistion in the table that resolves via
+     * DISP table, i.e. a for-us bit-position
+     */
+    fib_route_path_t *paths_via_disp = NULL, path_via_disp = {
+        // .frp_addr = all-zeros
+        .frp_bier_fib_index = bdti1,
+        .frp_flags = FIB_ROUTE_PATH_BIER_FMASK,
+    };
+    vec_add1(paths_via_disp, path_via_disp);
+
+    bier_table_route_add(&bt_0_0_0_256, 3, paths_via_disp);
+
+    /*
+     * the fmask should stack on the BIER disp table
+     */
+    const bier_fmask_id_t bfm_id_0_0_0_0 = {
+        .bfmi_hdr_type = BIER_HDR_O_MPLS,
+    };
+    bier_fmask_t *bfm_0_0_0_0;
+    index_t bfmi_0_0_0_0;
+    dpo_id_t dpo_disp_tbl_1 = DPO_INVALID;
+
+    bier_disp_table_contribute_forwarding(bdti1, &dpo_disp_tbl_1);
+
+    bfmi_0_0_0_0 = bier_fmask_db_find(bti, &bfm_id_0_0_0_0);
+    bfm_0_0_0_0 = bier_fmask_get(bfmi_0_0_0_0);
+
+    BIER_TEST(!dpo_cmp(&dpo_disp_tbl_1, &bfm_0_0_0_0->bfm_dpo),
+              "Fmask via 0.0.0.0 stacks on BIER disp table 1");
+
+    /*
+     * and a deag entry into the disposition table
+     */
+    fib_route_path_t *rpaths = NULL, path_via_mfib = {
+        .frp_proto = DPO_PROTO_IP4,
+        .frp_addr = zero_addr,
+        .frp_fib_index = 0, // default MFIB table
+        .frp_rpf_id = 9, // some non-zero value
+        .frp_flags = FIB_ROUTE_PATH_RPF_ID,
+    };
+    u16 src = 99;
+    vec_add1(rpaths, path_via_mfib);
+    bier_disp_table_entry_path_add(bier_disp_tbl_id, src,
+                                   BIER_HDR_PROTO_IPV4, rpaths);
+
+    /* which should stack on a lookup in the mfib table */
+    const dpo_id_t *dpo_disp_entry_lb;
+    const dpo_id_t *dpo_disp_entry_v4;
+    bier_disp_entry_t *bde_99;
+    index_t bdei;
+
+    bdei = bier_disp_table_lookup(bdti1, clib_host_to_net_u16(src));
+    bde_99 = bier_disp_entry_get(bdei);
+    dpo_disp_entry_lb = &bde_99->bde_fwd[BIER_HDR_PROTO_IPV4].bde_dpo;
+
+    BIER_TEST(dpo_disp_entry_lb->dpoi_type == DPO_LOAD_BALANCE,
+              "BIER Disp entry stacks on LB");
+
+    load_balance_t *lb;
+    lb = load_balance_get(dpo_disp_entry_lb->dpoi_index);
+    dpo_disp_entry_v4 = load_balance_get_bucket_i(lb, 0);
+
+    lookup_dpo_t *lkd = lookup_dpo_get(dpo_disp_entry_v4->dpoi_index);
+
+    BIER_TEST((bdti1 == lkd->lkd_fib_index),
+              "disp is deag in %d %U",
+              lkd->lkd_fib_index,
+              format_dpo_id, dpo_disp_entry_v4, 0);
+    BIER_TEST((LOOKUP_INPUT_DST_ADDR == lkd->lkd_input),
+              "disp is destination deag in %d %U",
+              lkd->lkd_input,
+              format_dpo_id, dpo_disp_entry_v4, 0);
+    BIER_TEST((LOOKUP_MULTICAST == lkd->lkd_cast),
+              "disp is multicast deag in %d %U",
+              lkd->lkd_input,
+              format_dpo_id, dpo_disp_entry_v4, 0);
+
+    /*
+     * cleanup
+     */
+    dpo_reset(&dpo_disp_tbl_1);
+
+    bier_disp_table_entry_path_remove(bier_disp_tbl_id, src,
+                                      BIER_HDR_PROTO_IPV4, rpaths);
+    bier_table_route_remove(&bt_0_0_0_256, 3, paths_via_disp);
+
+    bier_disp_table_unlock_w_table_id(bier_disp_tbl_id);
+
+    bier_table_unlock(&bt_0_0_0_256);
+
+    BIER_TEST(0 == pool_elts(bier_fmask_pool),
+              "BIER fmask resources freed ");
+    BIER_TEST(0 == pool_elts(bier_table_pool),
+              "BIER table resources freed ");
+    BIER_TEST(0 == pool_elts(bier_disp_table_pool),
+              "BIER Disposition table resources freed ");
+    BIER_TEST(0 == pool_elts(bier_disp_entry_pool),
+              "BIER Disposition entry resources freed ");
+
+    vec_free(paths_via_disp);
+    return (0);
+}
+
+static clib_error_t *
+bier_test (vlib_main_t * vm,
+           unformat_input_t * input,
+           vlib_cli_command_t * cmd_arg)
+{
+    int res = 0;
+
+    res += bier_test_mk_intf(4);
+
+    if (unformat (input, "debug"))
+    {
+        bier_test_do_debug = 1;
+    }
+
+    if (unformat (input, "mid"))
+        res += bier_test_mpls_spf();
+    else if (unformat (input, "head"))
+        res += bier_test_mpls_imp();
+    else if (unformat (input, "tail"))
+        res += bier_test_mpls_disp();
+    else
+    {
+        res += bier_test_mpls_spf();
+        res += bier_test_mpls_imp();
+        res += bier_test_mpls_disp();
+    }
+
+    if (res)
+    {
+        return clib_error_return(0, "BIER Unit Test Failed");
+    }
+    else
+    {
+        return (NULL);
+    }
+}
+
+VLIB_CLI_COMMAND (test_route_command, static) = {
+    .path = "test bier",
+    .short_help = "bier unit tests",
+    .function = bier_test,
+};
+
+clib_error_t *
+bier_test_init (vlib_main_t *vm)
+{
+    return 0;
+}
+
+VLIB_INIT_FUNCTION (bier_test_init);
diff --git a/src/vnet/bier/bier_types.c b/src/vnet/bier/bier_types.c
new file mode 100644
index 00000000000..0d27524aeb1
--- /dev/null
+++ b/src/vnet/bier/bier_types.c
@@ -0,0 +1,190 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vppinfra/types.h>
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_hdr_inlines.h>
+
+/*
+ * enum to string conversions
+ */
+const static char* const bier_hdr_len_id_names[] = BIER_HDR_LEN_IDS;
+const static char* const bier_hdr_proto_names[] = BIER_HDR_PROTO_ID_NAMES;
+
+const static u16 bier_hdr_len_num_buckets[] = {
+    [BIER_HDR_LEN_INVALID] = 0,
+    [BIER_HDR_LEN_64] = 8,
+    [BIER_HDR_LEN_128] = 16,
+    [BIER_HDR_LEN_256] = 32,
+    [BIER_HDR_LEN_512] = 64,
+    [BIER_HDR_LEN_1024] = 128,
+    [BIER_HDR_LEN_2048] = 256,
+    [BIER_HDR_LEN_4096] = 512,
+};
+
+const static u16 bier_hdr_len_num_bits[] = {
+    [BIER_HDR_LEN_INVALID] = 0,
+    [BIER_HDR_LEN_64] = 64,
+    [BIER_HDR_LEN_128] = 128,
+    [BIER_HDR_LEN_256] = 256,
+    [BIER_HDR_LEN_512] = 512,
+    [BIER_HDR_LEN_1024] = 1024,
+    [BIER_HDR_LEN_2048] = 2048,
+    [BIER_HDR_LEN_4096] = 4096,
+};
+
+const static u16 bier_hdr_len_prefix_len[] = {
+    [BIER_HDR_LEN_INVALID] = 0,
+    [BIER_HDR_LEN_64] = 7,
+    [BIER_HDR_LEN_128] = 8,
+    [BIER_HDR_LEN_256] = 9,
+    [BIER_HDR_LEN_512] = 10,
+    [BIER_HDR_LEN_1024] = 11,
+    [BIER_HDR_LEN_2048] = 12,
+    [BIER_HDR_LEN_4096] = 13,
+};
+
+uint32_t
+bier_hdr_len_id_to_num_buckets (bier_hdr_len_id_t id)
+{
+    return (bier_hdr_len_num_buckets[id]);
+}
+
+uint32_t
+bier_hdr_len_id_to_num_bytes (bier_hdr_len_id_t id)
+{
+    return (bier_hdr_len_id_to_num_buckets(id));
+}
+
+uint32_t
+bier_hdr_len_id_to_max_bucket (bier_hdr_len_id_t id)
+{
+    return (bier_hdr_len_id_to_num_buckets(id) - 1);
+}
+
+uint32_t
+bier_hdr_len_id_to_num_bits (bier_hdr_len_id_t id)
+{
+    return (bier_hdr_len_num_bits[id]);
+}
+
+uint32_t
+bier_hdr_len_id_to_max_bit (bier_hdr_len_id_t id)
+{
+    return (bier_hdr_len_id_to_num_bits(id));
+}
+
+uint32_t
+bier_hdr_len_id_to_prefix_len (bier_hdr_len_id_t id)
+{
+    return (bier_hdr_len_prefix_len[id]);
+}
+
+u8 *
+format_bier_hdr_len_id (u8 *s, va_list *ap)
+{
+    bier_hdr_len_id_t hli = va_arg(*ap, int); // int promotion of bier_hdr_len_id_t
+
+    return (format(s, "%s", bier_hdr_len_id_names[hli]));
+}
+
+u8 *
+format_bier_hdr_proto (u8 *s, va_list *ap)
+{
+    bier_hdr_proto_id_t pi = va_arg(*ap, int);
+
+    return (format(s, "%s", bier_hdr_proto_names[pi]));
+}
+
+int
+bier_table_id_cmp (const bier_table_id_t *btid1,
+                   const bier_table_id_t *btid2)
+{
+    int res;
+
+    res = (btid1->bti_set - btid2->bti_set);
+
+    if (0 == res)
+    {
+        res  = (btid1->bti_sub_domain - btid2->bti_sub_domain);
+    }
+    if (0 == res)
+    {
+        res = (btid1->bti_ecmp - btid2->bti_ecmp);
+    }
+    if (0 == res)
+    {
+        res = (btid1->bti_hdr_len - btid2->bti_hdr_len);
+    }
+    if (0 == res)
+    {
+        res = (btid1->bti_type - btid2->bti_type);
+    }
+    return (res);
+}
+
+dpo_proto_t
+bier_hdr_proto_to_dpo (bier_hdr_proto_id_t bproto)
+{
+    switch (bproto)
+    {
+    case BIER_HDR_PROTO_INVALID:
+    case BIER_HDR_PROTO_CTRL:
+    case BIER_HDR_PROTO_OAM:
+        ASSERT(0);
+        break;
+    case BIER_HDR_PROTO_MPLS_DOWN_STREAM:
+    case BIER_HDR_PROTO_MPLS_UP_STREAM:
+        return (DPO_PROTO_MPLS);
+    case BIER_HDR_PROTO_ETHERNET:
+    case BIER_HDR_PROTO_VXLAN:
+        return (DPO_PROTO_ETHERNET);
+    case BIER_HDR_PROTO_IPV4:
+        return (DPO_PROTO_IP4);
+    case BIER_HDR_PROTO_IPV6:
+        return (DPO_PROTO_IP4);
+    }
+
+    return (DPO_PROTO_NUM);
+}
+
+u8 *
+format_bier_table_id (u8 *s, va_list *ap)
+{
+    bier_table_id_t *btid = va_arg(*ap, bier_table_id_t *);
+
+    return (format(s, "sub-domain:%d set:%d ecmp:%d bsl:%U",
+                   btid->bti_sub_domain,
+                   btid->bti_set,
+                   btid->bti_ecmp,
+                   format_bier_hdr_len_id, btid->bti_hdr_len));
+}
+
+u8 *
+format_bier_hdr (u8 *s, va_list *ap)
+{
+    bier_hdr_t *bh = va_arg(*ap, bier_hdr_t *);
+    bier_hdr_t copy = *bh;
+
+    bier_hdr_ntoh(&copy);
+
+    return (format(s, "nibble:%d version:%d hdr-len:%U entropy:%d proto:%U src:%d",
+                   bier_hdr_get_1st_nibble(&copy),
+                   bier_hdr_get_version(&copy),
+                   format_bier_hdr_len_id, bier_hdr_get_len_id(&copy),
+                   bier_hdr_get_entropy(&copy),
+                   format_bier_hdr_proto, bier_hdr_get_proto_id(&copy),
+                   bier_hdr_get_src_id(&copy)));
+}
diff --git a/src/vnet/bier/bier_types.h b/src/vnet/bier/bier_types.h
new file mode 100644
index 00000000000..fa1cd423278
--- /dev/null
+++ b/src/vnet/bier/bier_types.h
@@ -0,0 +1,507 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef __BIER_TYPES_H__
+#define __BIER_TYPES_H__
+
+#include <vlib/vlib.h>
+#include <vnet/dpo/dpo.h>
+
+/**
+ * @brief Flags to control show output
+ */
+typedef enum bier_show_flags_t_ {
+    BIER_SHOW_BRIEF,
+    BIER_SHOW_DETAIL = (1 << 0),
+} bier_show_flags_t;
+
+/**
+ * Types of BIER tables
+ */
+typedef enum bier_table_type_t_ {
+    /**
+     * BIER over MPLS with SPF
+     */
+    BIER_TABLE_MPLS_SPF,
+
+    /**
+     * BIER over MPLS for TE
+     */
+    BIER_TABLE_MPLS_TE,
+} __attribute__((packed)) bier_table_type_t;
+
+#define BIER_TABLE_TYPES {              \
+    [BIER_TABLE_MPLS_SPF] = "mpls-spf", \
+    [BIER_TABLE_MPLS_TE]  = "mpls-te",  \
+}
+
+/**
+ * bier_hdr_len_id_t enumerator
+ **/
+typedef enum bier_hdr_len_id_t_ {
+    BIER_HDR_LEN_64 = 0,
+    BIER_HDR_LEN_128,
+    BIER_HDR_LEN_256,
+    BIER_HDR_LEN_512,
+    BIER_HDR_LEN_1024,
+    BIER_HDR_LEN_2048,
+    BIER_HDR_LEN_4096,
+    BIER_HDR_LEN_INVALID,
+} __attribute__((packed)) bier_hdr_len_id_t;
+
+#define BIER_HDR_LEN_IDS {             \
+    [BIER_HDR_LEN_INVALID] = "invalid",\
+    [BIER_HDR_LEN_64]      = "64",     \
+    [BIER_HDR_LEN_128]     = "128",    \
+    [BIER_HDR_LEN_256]     = "256",    \
+    [BIER_HDR_LEN_512]     = "512",    \
+    [BIER_HDR_LEN_1024]    = "1024",   \
+    [BIER_HDR_LEN_2048]    = "2048",   \
+    [BIER_HDR_LEN_4096]    = "4096",   \
+}
+
+#define FOR_EACH_BIER_HDR_LEN(_len)    \
+    for (_item = BIER_HDR_LEN_64;      \
+         _item <= BIER_HDR_LEN_4096;   \
+         _item++)
+
+/**
+ * Format the header length field
+ */
+extern u8 *format_bier_hdr_len_id(u8 *s, va_list *ap);
+
+/*
+ * convert from prefix len to hdr ID
+ */
+static inline bier_hdr_len_id_t
+bier_prefix_len_to_hdr_id (u16 prfx_len) {
+
+    switch (prfx_len) {
+    case 7:
+        return (BIER_HDR_LEN_64);
+    case 8:
+        return (BIER_HDR_LEN_128);
+    case 9:
+        return (BIER_HDR_LEN_256);
+    case 10:
+        return (BIER_HDR_LEN_512);
+    case 11:
+        return (BIER_HDR_LEN_1024);
+    case 12:
+        return (BIER_HDR_LEN_2048);
+    case 13:
+        return (BIER_HDR_LEN_4096);
+    default:
+        break;
+    }
+
+    return (BIER_HDR_LEN_INVALID);
+}
+
+static inline bier_hdr_len_id_t
+bier_hdr_byte_len_to_id (u32 bytes)
+{
+    switch (bytes) {
+    case 8:
+        return (BIER_HDR_LEN_64);
+    case 16:
+        return (BIER_HDR_LEN_128);
+    case 32:
+        return (BIER_HDR_LEN_256);
+    case 64:
+        return (BIER_HDR_LEN_512);
+    case 128:
+        return (BIER_HDR_LEN_1024);
+    case 256:
+        return (BIER_HDR_LEN_2048);
+    case 512:
+        return (BIER_HDR_LEN_4096);
+    }
+
+    return (BIER_HDR_LEN_INVALID);
+}
+
+static inline bier_hdr_len_id_t
+bier_hdr_bit_len_to_id (u32 bytes)
+{
+    switch (bytes) {
+    case 64:
+        return (BIER_HDR_LEN_64);
+    case 128:
+        return (BIER_HDR_LEN_128);
+    case 256:
+        return (BIER_HDR_LEN_256);
+    case 512:
+        return (BIER_HDR_LEN_512);
+    case 1024:
+        return (BIER_HDR_LEN_1024);
+    case 2048:
+        return (BIER_HDR_LEN_2048);
+    case 4096:
+        return (BIER_HDR_LEN_4096);
+    }
+
+    return (BIER_HDR_LEN_INVALID);
+}
+
+/**
+ * bier_hdr_len_num_buckets_t enumerator
+ **/
+typedef enum bier_hdr_len_num_buckets_t_ {
+    BIER_HDR_BUCKETS_64 = 8,
+    BIER_HDR_BUCKETS_128 = 16,
+    BIER_HDR_BUCKETS_256 = 32,
+    BIER_HDR_BUCKETS_512 = 64,
+    BIER_HDR_BUCKETS_1024 = 128,
+    BIER_HDR_BUCKETS_2048 = 256,
+    BIER_HDR_BUCKETS_4096 = 512,
+} bier_hdr_len_num_buckets_t;
+
+/**
+ * BIER header protocol payload types
+ **/
+typedef enum bier_hdr_proto_id_t_ {
+    BIER_HDR_PROTO_INVALID = 0,
+    BIER_HDR_PROTO_MPLS_DOWN_STREAM,
+    BIER_HDR_PROTO_MPLS_UP_STREAM,
+    BIER_HDR_PROTO_ETHERNET,
+    BIER_HDR_PROTO_IPV4,
+    BIER_HDR_PROTO_IPV6,
+    BIER_HDR_PROTO_VXLAN,
+    BIER_HDR_PROTO_CTRL,
+    BIER_HDR_PROTO_OAM,
+} __attribute__((packed)) bier_hdr_proto_id_t;
+
+#define BIER_HDR_N_PROTO (BIER_HDR_PROTO_OAM + 1)
+
+#define BIER_HDR_PROTO_ID_NAMES {                               \
+    [BIER_HDR_PROTO_INVALID] = "invalid",			\
+    [BIER_HDR_PROTO_MPLS_DOWN_STREAM] = "mpls-down-stream",     \
+    [BIER_HDR_PROTO_MPLS_UP_STREAM] = "mpls-up-stream",         \
+    [BIER_HDR_PROTO_ETHERNET] = "ethernet",                     \
+    [BIER_HDR_PROTO_IPV4] = "ipv4",                             \
+    [BIER_HDR_PROTO_IPV6] = "ipv6",                             \
+    [BIER_HDR_PROTO_VXLAN] = "vxlan",                           \
+    [BIER_HDR_PROTO_CTRL] = "control-plane",                    \
+    [BIER_HDR_PROTO_OAM] = "oam",                               \
+}
+
+#define FOR_EACH_BIER_HDR_PROTO(_proto)                 \
+    for (_proto = BIER_HDR_PROTO_MPLS_DOWN_STREAM;      \
+         _proto <= BIER_HDR_PROTO_OAM;                  \
+         _proto++)
+
+/**
+ * Format the header length field
+ */
+extern u8 *format_bier_hdr_proto(u8 *s, va_list *ap);
+
+/**
+ * Convert from BIER next-hop proto to DPO proto
+ */
+extern dpo_proto_t bier_hdr_proto_to_dpo(bier_hdr_proto_id_t bproto);
+
+/**
+ * BIER header versions
+ **/
+typedef enum bier_hdr_version_t_ {
+    BIER_HDR_VERSION_1 = 0,
+} __attribute__((packed)) bier_hdr_version_t;
+
+/**
+ * bier_hdr_code_t enumerator
+ **/
+typedef enum bier_hdr_code_t_ {
+    BIER_HDR_CODE_OAM_IPV4 = 0,
+    BIER_HDR_CODE_OAM_IPV6 = 1,
+    BIER_HDR_CODE_CTRL_IPV4 = 2,
+    BIER_HDR_CODE_CTRL_IPV6 = 3,
+} __attribute__((packed)) bier_hdr_code_t;
+
+/**
+ * bier_hdr_oam_sub_code_t enumerator
+ */
+typedef enum bier_hdr_oam_sub_code_t_ {
+    BIER_HDR_SUB_CODE_OAM_PING_REQ = 0,
+    BIER_HDR_SUB_CODE_OAM_PING_RESP = 1,
+} __attribute__((packed)) bier_hdr_oam_sub_code_t;
+
+/**
+ * bier_hdr_ctrl_sub_code_t enumerator
+ */
+typedef enum bier_hdr_ctrl_sub_code_t_ {
+    BIER_HDR_SUB_CODE_CTRL_MEMBER_REQ = 0,
+    BIER_HDR_SUB_CODE_CTRL_ATTACHED_NET = 1,
+} __attribute__((packed)) bier_hdr_ctrl_sub_code_t;
+
+/**
+ * A bucket is a byte. The byte string is thus always in network byte order.
+ */
+typedef u8 bier_bit_mask_bucket_t;
+
+/**
+ * A BIER Bit-String value of length 64 bits.
+ */
+typedef struct bier_bit_mask_64_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_64];
+} bier_bit_mask_64_t;
+
+/**
+ * A BIER Bit-String value of length 128 bits.
+ */
+typedef struct bier_bit_mask_128_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_128];
+} bier_bit_mask_128_t;
+
+/**
+ * A BIER Bit-String value of length 256 bits.
+ */
+typedef struct bier_bit_mask_256_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_256];
+} bier_bit_mask_256_t;
+
+/**
+ * A BIER Bit-String value of length 512 bits.
+ */
+typedef struct bier_bit_mask_512_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_512];
+} bier_bit_mask_512_t;
+
+/**
+ * A BIER Bit-String value of length 1024 bits.
+ */
+typedef struct bier_bit_mask_1024_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_1024];
+} bier_bit_mask_1024_t;
+
+/**
+ * A BIER Bit-String value of length 2048 bits.
+ */
+typedef struct bier_bit_mask_2048_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_2048];
+} bier_bit_mask_2048_t;
+
+/**
+ * A BIER Bit-String value of length 4096 bits.
+ */
+typedef struct bier_bit_mask_4096_t_ {
+    bier_bit_mask_bucket_t bits[BIER_HDR_BUCKETS_4096];
+} bier_bit_mask_4096_t;
+
+
+/**
+ * 256 bits = 32 bytes
+ */
+#define BIER_BIT_MASK_NUM_BUCKETS 32
+#define BIER_BIT_MASK_MAX_BUCKET (BIER_BIT_MASK_NUM_BUCKETS - 1)
+
+/**
+ * number of bits in a bucket
+ */
+#define BIER_BIT_MASK_BITS_PER_BUCKET 8
+
+/**
+ * Supported bit-posiotn range
+ */
+#define BIER_BIT_MASK_MIN_POS (1)
+
+/**
+ * A Variable length BitString
+ */
+typedef struct bier_bit_string_t_ {
+    /**
+     * The length of the string in BYTES
+     */
+    u16 bbs_len;
+
+    /**
+     * The buckets in the string
+     */
+    bier_bit_mask_bucket_t *bbs_buckets;
+} bier_bit_string_t;
+
+/**
+ * A BIER Bit-mask value
+ *
+ * The size of this mask represents this platforms BIER capabilities
+ */
+typedef bier_bit_mask_256_t bier_bit_mask_t;
+
+/**
+ * A bit positon
+ *  as assigned to egress PEs
+ */
+typedef u32 bier_bp_t;
+
+#define BIER_BP_TO_INDEX(bp) (bp - 1)
+
+/**
+ * The maximum BP that can be assigned
+ */
+#define BIER_BP_MAX 0x10000
+
+/**
+ * An identifier of the sender of BIER packets
+ * this is the source of the 'tree' - the BFIR
+ */
+typedef u16 bier_hdr_src_id_t;
+
+/**
+ * An entropy value in a BIER header
+ */
+typedef u32 bier_hdr_entropy_t;
+
+#define BIER_BP_INVALID 0
+
+/**
+ * A BIER header of variable length
+ * The encoding follows:
+ *   https://tools.ietf.org/html/draft-ietf-bier-mpls-encapsulation-10
+ */
+typedef struct bier_hdr_t_ {
+    /**
+     * The first nibble is always set to 0101
+     * to ensure that when carried over MPLS, the BIER packet
+     * is not mistaken for IPv[46]:
+     *   type: bier_hdr_version_t
+     *
+     * The second nibble is the version - this is 0:
+     *   type: bier_hdr_version_t
+     *
+     * The third nibble is header length ID
+     *   type: bier_hdr_len_id_t
+     *
+     * The next 20 bits are entropy
+     * An entropy value, calculated by the head end, used
+     * at the head and mid-points for load-balance hash
+     *   type: bier_hdr_entropy_t
+     */
+    u32 bh_first_word;
+
+    /**
+     * The second word comprises:
+     *  2 bits of OAM for passive perf measurement
+     *  2 reserved bits;
+     *  6 bits of DSCP
+     *  6 bits for the next-proto field of type;
+     *     bier_hdr_proto_id_t
+     */
+    u16 bh_oam_dscp_proto;
+
+    /**
+     * The BFR-ID of the sender
+     */
+    u16 bh_bfr_id;
+
+    /**
+     * The variable length bit-string
+     */
+    bier_bit_mask_bucket_t bh_bit_string[0];
+} bier_hdr_t;
+
+/**
+ * Format a BIER header
+ */
+extern u8 *format_bier_hdr(u8 *s, va_list *ap);
+
+/**
+ * The BIER Set ID assigned to a BIER table
+ */
+typedef u32 bier_table_set_id_t;
+
+#define BIER_TABLE_SET_INVALID_ID 0xffffffff
+
+/**
+ * The BIER Sub-domain ID assigned to a BIER table
+ */
+typedef u32 bier_table_sub_domain_id_t;
+
+#define BIER_TABLE_SUB_DOMAIN_INVALID_ID 0xffffffff
+
+/**
+ * An ID or instance number of a BIER sub-table
+ */
+typedef u32 bier_table_ecmp_id_t;
+
+/**
+ * Definition of the ID of the BIER main table
+ */
+#define BIER_ECMP_TABLE_ID_MAIN 0xFFFF
+
+/**
+ * The ID of a table
+ */
+typedef struct bier_table_id_t_ {
+    /**
+     * The SET-ID
+     *  The control plane divdies the bit-position space
+     * into sets in the case the max bit-position is greater
+     * than the table's bit-string size
+     */
+    bier_table_set_id_t bti_set;
+
+    /**
+     * The Sub-Domain-ID
+     * The control plane has the configuration option to specify multiple
+     * domains or topologies.
+     */
+    bier_table_sub_domain_id_t bti_sub_domain;
+
+    /**
+     * The SUB/ECMP-ID
+     * Constructed by FIB to achieve ECMP between BFR-NBRs
+     */
+    bier_table_ecmp_id_t bti_ecmp;
+
+    /**
+     * The size of the bit string processed by this table.
+     */
+    bier_hdr_len_id_t bti_hdr_len;
+
+   /**
+     * The type of the table; SPF or TE, MPLS or IPv6
+     */
+    bier_table_type_t bti_type;
+} bier_table_id_t;
+
+/**
+ * Format a BIER table ID
+ */
+extern u8 *format_bier_table_id(u8 *s, va_list *ap);
+
+/**
+ * Compare to BIER table IDs for equality
+ */
+extern int bier_table_id_cmp(const bier_table_id_t *btid1,
+                             const bier_table_id_t *btid2);
+
+/**
+ * Conversion functions for the enumerated bit-string length
+ * values, to bit and bytes
+ */
+extern u32 bier_hdr_len_id_to_num_buckets(bier_hdr_len_id_t id);
+extern u32 bier_hdr_len_id_to_num_bytes(bier_hdr_len_id_t id);
+extern u32 bier_hdr_len_id_to_max_bucket(bier_hdr_len_id_t id);
+extern u32 bier_hdr_len_id_to_num_bits(bier_hdr_len_id_t id);
+extern u32 bier_hdr_len_id_to_max_bit(bier_hdr_len_id_t id);
+extern u32 bier_hdr_len_id_to_prefix_len(bier_hdr_len_id_t id);
+
+#define BIER_OK 0
+#define BIER_ERR_NO_TABLE 1
+#define BIER_ERR_DUPLICATE_TABLE 2
+#define BIER_ERR_PANIC 3
+typedef int bier_rc;
+
+#endif /* __BIER_TYPES_H__ */
diff --git a/src/vnet/bier/bier_update.c b/src/vnet/bier/bier_update.c
new file mode 100644
index 00000000000..ddbdd72f8ee
--- /dev/null
+++ b/src/vnet/bier/bier_update.c
@@ -0,0 +1,182 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vnet/vnet.h>
+#include <vnet/mpls/mpls.h>
+
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_types.h>
+#include <vnet/bier/bier_update.h>
+
+clib_error_t *
+vnet_bier_table_cmd (vlib_main_t * vm,
+                     unformat_input_t * input,
+                     vlib_cli_command_t * cmd)
+{
+    u32 hdr_len, local_label;
+    clib_error_t * error = 0;
+    bier_table_id_t bti = {
+        .bti_ecmp = BIER_ECMP_TABLE_ID_MAIN,
+    };
+    u32 is_add = 0;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "del")) {
+            is_add = 0;
+        } else if (unformat (input, "add")) {
+            is_add = 1;
+        } else if (unformat (input, "sd %d", &bti.bti_sub_domain)) {
+        } else if (unformat (input, "set %d", &bti.bti_set)) {
+        } else if (unformat (input, "bsl %d", &hdr_len)) {
+        } else if (unformat (input, "mpls %d", &local_label)) {
+        } else {
+            error = unformat_parse_error (input);
+            goto done;
+        }
+    }
+
+    bti.bti_hdr_len = bier_hdr_bit_len_to_id(hdr_len);
+    // FIXME
+    bti.bti_type = BIER_TABLE_MPLS_SPF;
+
+    if (is_add)
+    {
+        bier_table_add_or_lock(&bti, local_label);
+    }
+    else
+    {
+        bier_table_unlock(&bti);
+    }
+
+done:
+    return (error);
+}
+
+VLIB_CLI_COMMAND (bier_table_command) = {
+  .path = "bier table",
+  .short_help = "Add/delete BIER Tables",
+  .function = vnet_bier_table_cmd,
+};
+
+clib_error_t *
+vnet_bier_route_cmd (vlib_main_t * vm,
+                     unformat_input_t * input,
+                     vlib_cli_command_t * cmd)
+{
+    clib_error_t * error = NULL;
+    fib_route_path_t brp = {
+        .frp_flags = FIB_ROUTE_PATH_BIER_FMASK,
+    };
+    bier_table_id_t bti = {
+    };
+    mpls_label_t out_label;
+    bier_bp_t bp;
+    u32 hdr_len;
+    u32 add = 1;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "del")) {
+            add = 0;
+        } else if (unformat (input, "sd %d", &bti.bti_sub_domain)) {
+        } else if (unformat (input, "set %d", &bti.bti_set)) {
+        } else if (unformat (input, "bsl %d", &hdr_len)) {
+        } else if (unformat (input, "bp %d", &bp)) {
+        } else if (unformat (input, "v4-nh %U",
+                             unformat_ip46_address,
+                             &brp.frp_addr, 0)) {
+        } else if (unformat (input, "mpls %d", &out_label)) {
+            vec_add1(brp.frp_label_stack, out_label);
+        } else {
+            error = unformat_parse_error (input);
+            goto done;
+        }
+    }
+
+    bti.bti_hdr_len = bier_hdr_bit_len_to_id(hdr_len);
+    // FIXME
+    bti.bti_type    = BIER_TABLE_MPLS_SPF;
+
+    if (add)
+    {
+        bier_table_route_add(&bti, bp, &brp);
+    }
+    else
+    {
+        bier_table_route_remove(&bti, bp, &brp);
+    }
+
+done:
+    return (error);
+}
+
+VLIB_CLI_COMMAND (bier_route_command) = {
+  .path = "bier route",
+  .short_help = "Add/delete BIER Routes",
+  .function = vnet_bier_route_cmd,
+};
+
+static clib_error_t *
+show_bier_fib_command_fn (vlib_main_t * vm,
+                          unformat_input_t * input,
+                          vlib_cli_command_t * cmd)
+{
+    bier_show_flags_t flags;
+    index_t bti, bei;
+    bier_bp_t bp;
+
+    bp = BIER_BP_INVALID;
+    bti = bei = INDEX_INVALID;
+
+    while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) {
+        if (unformat (input, "%d %d", &bti, &bp))
+        {
+             flags = BIER_SHOW_DETAIL;
+        }
+        else if (unformat (input, "%d", &bti))
+        {
+              flags = BIER_SHOW_DETAIL;
+        }
+        else
+        {
+            break;
+        }
+    }
+
+    if (INDEX_INVALID == bti)
+    {
+        bier_table_show_all(vm, flags);
+    }
+    else
+    {
+        if (!pool_is_free_index(bier_table_pool, bti))
+        {
+            if (BIER_BP_INVALID == bp)
+            {
+                vlib_cli_output (vm, "%U", format_bier_table, bti, flags);
+            }
+            else
+            {
+                vlib_cli_output (vm, "%U", format_bier_table_entry, bti, bp);
+            }
+        }
+    }
+    return (NULL);
+}
+
+VLIB_CLI_COMMAND (show_bier_fib_command, static) = {
+    .path = "show bier fib",
+    .short_help = "show bier fib [table-index] [bit-position]",
+    .function = show_bier_fib_command_fn,
+};
diff --git a/src/vnet/bier/bier_update.h b/src/vnet/bier/bier_update.h
new file mode 100644
index 00000000000..b4bee9fef72
--- /dev/null
+++ b/src/vnet/bier/bier_update.h
@@ -0,0 +1,24 @@
+/*
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef __BIER_UPDATE_H__
+#define __BIER_UPDATE_H__
+
+#include <vppinfra/types.h>
+#include <vnet/bier/bier_types.h>
+#include <vnet/ip/ip.h>
+
+
+#endif
diff --git a/src/vnet/buffer.h b/src/vnet/buffer.h
index 045328816cd..6518fb67016 100644
--- a/src/vnet/buffer.h
+++ b/src/vnet/buffer.h
@@ -88,6 +88,8 @@ _(ipsec)					\
 _(map)						\
 _(map_t)					\
 _(ip_frag)					\
+_(mpls)					        \
+_(bier) 					\
 _(tcp)
 
 /*
@@ -165,6 +167,16 @@ typedef struct
       u8 first;
     } mpls;
 
+    /*
+     * BIER - the nubmer of bytes in the header.
+     *  the len field inthe header is not authoritative. It's the
+     * value in the table that counts.
+     */
+    struct
+    {
+      u8 n_bytes;
+    } bier;
+
     /* ip4-in-ip6 softwire termination, only valid there */
     struct
     {
diff --git a/src/vnet/dpo/dpo.c b/src/vnet/dpo/dpo.c
index 7658132d47a..1ee0727135d 100644
--- a/src/vnet/dpo/dpo.c
+++ b/src/vnet/dpo/dpo.c
@@ -122,6 +122,7 @@ dpo_proto_to_link (dpo_proto_t dp)
     case DPO_PROTO_IP4:
         return (VNET_LINK_IP4);
     case DPO_PROTO_MPLS:
+    case DPO_PROTO_BIER:
         return (VNET_LINK_MPLS);
     case DPO_PROTO_ETHERNET:
         return (VNET_LINK_ETHERNET);
diff --git a/src/vnet/dpo/dpo.h b/src/vnet/dpo/dpo.h
index 304b4331495..27c129ed444 100644
--- a/src/vnet/dpo/dpo.h
+++ b/src/vnet/dpo/dpo.h
@@ -63,6 +63,7 @@ typedef enum dpo_proto_t_
     DPO_PROTO_IP6,
     DPO_PROTO_MPLS,
     DPO_PROTO_ETHERNET,
+    DPO_PROTO_BIER,
     DPO_PROTO_NSH,
 } __attribute__((packed)) dpo_proto_t;
 
@@ -75,6 +76,7 @@ typedef enum dpo_proto_t_
     [DPO_PROTO_ETHERNET]  = "ethernet", \
     [DPO_PROTO_MPLS] = "mpls",	\
     [DPO_PROTO_NSH] = "nsh",    \
+    [DPO_PROTO_BIER] = "bier",	\
 }
 
 #define FOR_EACH_DPO_PROTO(_proto)    \
@@ -116,6 +118,11 @@ typedef enum dpo_type_t_ {
     DPO_INTERFACE_TX,
     DPO_L2_BRIDGE,
     DPO_L3_PROXY,
+    DPO_BIER_TABLE,
+    DPO_BIER_FMASK,
+    DPO_BIER_IMP,
+    DPO_BIER_DISP_TABLE,
+    DPO_BIER_DISP_ENTRY,
     DPO_LAST,
 } __attribute__((packed)) dpo_type_t;
 
@@ -145,6 +152,11 @@ typedef enum dpo_type_t_ {
     [DPO_INTERFACE_TX] = "dpo-interface-tx",	\
     [DPO_L2_BRIDGE] = "dpo-l2-bridge",	\
     [DPO_L3_PROXY] = "dpo-l3-proxy",	\
+    [DPO_BIER_TABLE] = "bier-table",	\
+    [DPO_BIER_FMASK] = "bier-fmask",	\
+    [DPO_BIER_IMP] = "bier-imposition",	\
+    [DPO_BIER_DISP_ENTRY] = "bier-disp-entry",	\
+    [DPO_BIER_DISP_TABLE] = "bier-disp-table",	\
 }
 
 /**
diff --git a/src/vnet/dpo/drop_dpo.c b/src/vnet/dpo/drop_dpo.c
index a1821dddfcc..6bb3ff7d5e0 100644
--- a/src/vnet/dpo/drop_dpo.c
+++ b/src/vnet/dpo/drop_dpo.c
@@ -96,6 +96,11 @@ const static char* const drop_nsh_nodes[] =
     "error-drop",
     NULL,
 };
+const static char* const drop_bier_nodes[] =
+{
+    "bier-drop",
+    NULL,
+};
 const static char* const * const drop_nodes[DPO_PROTO_NUM] =
 {
     [DPO_PROTO_IP4]  = drop_ip4_nodes,
@@ -103,6 +108,7 @@ const static char* const * const drop_nodes[DPO_PROTO_NUM] =
     [DPO_PROTO_MPLS] = drop_mpls_nodes,
     [DPO_PROTO_ETHERNET] = drop_ethernet_nodes,
     [DPO_PROTO_NSH] = drop_nsh_nodes,
+    [DPO_PROTO_BIER] = drop_bier_nodes,
 };
 
 void
diff --git a/src/vnet/dpo/load_balance.c b/src/vnet/dpo/load_balance.c
index af054f1c3f4..b48702ae3f2 100644
--- a/src/vnet/dpo/load_balance.c
+++ b/src/vnet/dpo/load_balance.c
@@ -21,6 +21,7 @@
 #include <vnet/adj/adj.h>
 #include <vnet/adj/adj_internal.h>
 #include <vnet/fib/fib_urpf_list.h>
+#include <vnet/bier/bier_hdr_inlines.h>
 
 /*
  * distribution error tolerance for load-balancing
@@ -814,6 +815,10 @@ const static char* const load_balance_l2_nodes[] =
 const static char* const load_balance_nsh_nodes[] =
 {
     "nsh-load-balance",
+};
+const static char* const load_balance_bier_nodes[] =
+{
+    "bier-load-balance",
     NULL,
 };
 const static char* const * const load_balance_nodes[DPO_PROTO_NUM] =
@@ -823,6 +828,7 @@ const static char* const * const load_balance_nodes[DPO_PROTO_NUM] =
     [DPO_PROTO_MPLS] = load_balance_mpls_nodes,
     [DPO_PROTO_ETHERNET] = load_balance_l2_nodes,
     [DPO_PROTO_NSH] = load_balance_nsh_nodes,
+    [DPO_PROTO_BIER] = load_balance_bier_nodes,
 };
 
 void
@@ -935,10 +941,11 @@ typedef struct load_balance_trace_t_
     index_t lb_index;
 } load_balance_trace_t;
 
-static uword
-l2_load_balance (vlib_main_t * vm,
-		 vlib_node_runtime_t * node,
-		 vlib_frame_t * frame)
+always_inline uword
+load_balance_inline (vlib_main_t * vm,
+		     vlib_node_runtime_t * node,
+		     vlib_frame_t * frame,
+		     int is_l2)
 {
   u32 n_left_from, next_index, *from, *to_next;
 
@@ -973,7 +980,16 @@ l2_load_balance (vlib_main_t * vm,
 	  lbi0 =  vnet_buffer (b0)->ip.adj_index[VLIB_TX];
 	  lb0 = load_balance_get(lbi0);
 
-	  vnet_buffer(b0)->ip.flow_hash = l2_flow_hash(b0);
+	  if (is_l2)
+	  {
+	      vnet_buffer(b0)->ip.flow_hash = l2_flow_hash(b0);
+	  }
+	  else
+	  {
+	      /* it's BIER */
+	      const bier_hdr_t *bh0 = vlib_buffer_get_current(b0);
+	      vnet_buffer(b0)->ip.flow_hash = bier_hdr_get_entropy(bh0);
+	  }
 
 	  dpo0 = load_balance_get_bucket_i(lb0, 
 					   vnet_buffer(b0)->ip.flow_hash &
@@ -998,6 +1014,14 @@ l2_load_balance (vlib_main_t * vm,
   return frame->n_vectors;
 }
 
+static uword
+l2_load_balance (vlib_main_t * vm,
+		 vlib_node_runtime_t * node,
+		 vlib_frame_t * frame)
+{
+    return (load_balance_inline(vm, node, frame, 1));
+}
+
 static u8 *
 format_l2_load_balance_trace (u8 * s, va_list * args)
 {
@@ -1113,3 +1137,34 @@ VLIB_REGISTER_NODE (nsh_load_balance_node) = {
       [0] = "error-drop",
   },
 };
+
+static u8 *
+format_bier_load_balance_trace (u8 * s, va_list * args)
+{
+  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+  load_balance_trace_t *t = va_arg (*args, load_balance_trace_t *);
+
+  s = format (s, "BIER-load-balance: index %d", t->lb_index);
+  return s;
+}
+
+static uword
+bier_load_balance (vlib_main_t * vm,
+		   vlib_node_runtime_t * node,
+		   vlib_frame_t * frame)
+{
+    return (load_balance_inline(vm, node, frame, 0));
+}
+
+/**
+ * @brief
+ */
+VLIB_REGISTER_NODE (bier_load_balance_node) = {
+  .function = bier_load_balance,
+  .name = "bier-load-balance",
+  .vector_size = sizeof (u32),
+
+  .format_trace = format_bier_load_balance_trace,
+  .sibling_of = "mpls-load-balance",
+};
diff --git a/src/vnet/dpo/replicate_dpo.c b/src/vnet/dpo/replicate_dpo.c
index 9fdb9a05071..ae045f6a431 100644
--- a/src/vnet/dpo/replicate_dpo.c
+++ b/src/vnet/dpo/replicate_dpo.c
@@ -261,7 +261,7 @@ replicate_fill_buckets (replicate_t *rep,
                         u32 n_buckets)
 {
     load_balance_path_t * nh;
-    u16 ii, bucket;
+    u16 bucket;
 
     bucket = 0;
 
@@ -271,11 +271,8 @@ replicate_fill_buckets (replicate_t *rep,
      */
     vec_foreach (nh, nhs)
     {
-        for (ii = 0; ii < nh->path_weight; ii++)
-        {
-            ASSERT(bucket < n_buckets);
-            replicate_set_bucket_i(rep, bucket++, buckets, &nh->path_dpo);
-        }
+        ASSERT(bucket < n_buckets);
+        replicate_set_bucket_i(rep, bucket++, buckets, &nh->path_dpo);
     }
 }
 
@@ -694,7 +691,10 @@ replicate_inline (vlib_main_t * vm,
 
                 if (PREDICT_FALSE(c0->flags & VLIB_BUFFER_IS_TRACED))
                 {
-                    replicate_trace_t *t = vlib_add_trace (vm, node, c0, sizeof (*t));
+                    replicate_trace_t *t;
+
+                    vlib_trace_buffer (vm, node, next0, c0, 0);
+                    t = vlib_add_trace (vm, node, c0, sizeof (*t));
                     t->rep_index = repi0;
                     t->dpo = *dpo0;
                 }
diff --git a/src/vnet/fib/fib_entry.c b/src/vnet/fib/fib_entry.c
index 4c9b1abd881..74c6a4a587b 100644
--- a/src/vnet/fib/fib_entry.c
+++ b/src/vnet/fib/fib_entry.c
@@ -394,6 +394,7 @@ fib_entry_chain_type_mcast_to_ucast (fib_forward_chain_type_t fct)
     case FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS:
     case FIB_FORW_CHAIN_TYPE_ETHERNET:
     case FIB_FORW_CHAIN_TYPE_NSH:
+    case FIB_FORW_CHAIN_TYPE_BIER:
         break;
     }
 
diff --git a/src/vnet/fib/fib_entry.h b/src/vnet/fib/fib_entry.h
index 7e4b52acffd..cd2a685b765 100644
--- a/src/vnet/fib/fib_entry.h
+++ b/src/vnet/fib/fib_entry.h
@@ -61,6 +61,10 @@ typedef enum fib_source_t_ {
      */
     FIB_SOURCE_PLUGIN_HI,
     /**
+     * From the BIER subsystem
+     */
+    FIB_SOURCE_BIER,
+    /**
      * From the control plane API
      */
     FIB_SOURCE_API,
@@ -141,6 +145,7 @@ STATIC_ASSERT (sizeof(fib_source_t) == 1,
     [FIB_SOURCE_SPECIAL] = "special",			\
     [FIB_SOURCE_INTERFACE] = "interface",		\
     [FIB_SOURCE_PROXY] = "proxy",                       \
+    [FIB_SOURCE_BIER] = "BIER",			        \
     [FIB_SOURCE_API] = "API",			        \
     [FIB_SOURCE_CLI] = "CLI",			        \
     [FIB_SOURCE_ADJ] = "adjacency",			\
diff --git a/src/vnet/fib/fib_entry_delegate.c b/src/vnet/fib/fib_entry_delegate.c
index 41af14f2245..4bf37df0a7e 100644
--- a/src/vnet/fib/fib_entry_delegate.c
+++ b/src/vnet/fib/fib_entry_delegate.c
@@ -122,6 +122,7 @@ fib_entry_chain_type_to_delegate_type (fib_forward_chain_type_t fct)
         return (FIB_ENTRY_DELEGATE_CHAIN_ETHERNET);
     case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
     case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
+    case FIB_FORW_CHAIN_TYPE_BIER:
         break;
     case FIB_FORW_CHAIN_TYPE_NSH:
         return (FIB_ENTRY_DELEGATE_CHAIN_NSH);
diff --git a/src/vnet/fib/fib_entry_src.c b/src/vnet/fib/fib_entry_src.c
index 173df74f0a4..667aa485f7c 100644
--- a/src/vnet/fib/fib_entry_src.c
+++ b/src/vnet/fib/fib_entry_src.c
@@ -275,6 +275,7 @@ fib_entry_src_get_path_forwarding (fib_node_index_t path_index,
     case FIB_FORW_CHAIN_TYPE_UNICAST_IP6:
     case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
     case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
+    case FIB_FORW_CHAIN_TYPE_BIER:
         /*
          * EOS traffic with no label to stack, we need the IP Adj
          */
@@ -480,8 +481,8 @@ fib_entry_src_mk_lb (fib_entry_t *fib_entry,
         }
         else
         {
+            fib_protocol_t flow_hash_proto;
             flow_hash_config_t fhc;
-            fib_protocol_t fp;
 
             /*
              * if the protocol for the LB we are building does not match that
@@ -489,16 +490,17 @@ fib_entry_src_mk_lb (fib_entry_t *fib_entry,
              * then the fib_index is not an index that relates to the table
              * type we need. So get the default flow-hash config instead.
              */
-            fp = dpo_proto_to_fib(lb_proto);
-
-            if (fib_entry->fe_prefix.fp_proto != fp)
+            flow_hash_proto = dpo_proto_to_fib(lb_proto);
+            if (fib_entry->fe_prefix.fp_proto != flow_hash_proto)
             {
-                fhc = fib_table_get_default_flow_hash_config(fp);
+                fhc = fib_table_get_default_flow_hash_config(flow_hash_proto);
             }
             else
             {
-                fhc = fib_table_get_flow_hash_config(fib_entry->fe_fib_index, fp);
+                fhc = fib_table_get_flow_hash_config(fib_entry->fe_fib_index,
+                                                     flow_hash_proto);
             }
+
             dpo_set(dpo_lb,
                     DPO_LOAD_BALANCE,
                     lb_proto,
diff --git a/src/vnet/fib/fib_entry_src_special.c b/src/vnet/fib/fib_entry_src_special.c
index c3e4fe5eba7..a2493bb5be9 100644
--- a/src/vnet/fib/fib_entry_src_special.c
+++ b/src/vnet/fib/fib_entry_src_special.c
@@ -68,4 +68,5 @@ fib_entry_src_special_register (void)
     fib_entry_src_register(FIB_SOURCE_CLASSIFY, &special_src_vft);
     fib_entry_src_register(FIB_SOURCE_AE, &special_src_vft);
     fib_entry_src_register(FIB_SOURCE_PROXY, &special_src_vft);
+    fib_entry_src_register(FIB_SOURCE_BIER, &special_src_vft);
 }
diff --git a/src/vnet/fib/fib_node.h b/src/vnet/fib/fib_node.h
index 6d26bdd3671..d4c96c90b77 100644
--- a/src/vnet/fib/fib_node.h
+++ b/src/vnet/fib/fib_node.h
@@ -45,6 +45,8 @@ typedef enum fib_node_type_t_ {
     FIB_NODE_TYPE_VXLAN_GPE_TUNNEL,
     FIB_NODE_TYPE_GENEVE_TUNNEL,
     FIB_NODE_TYPE_UDP_ENCAP,
+    FIB_NODE_TYPE_BIER_FMASK,
+    FIB_NODE_TYPE_BIER_ENTRY,
     /**
      * Marker. New types before this one. leave the test last.
      */
@@ -70,6 +72,8 @@ typedef enum fib_node_type_t_ {
     [FIB_NODE_TYPE_MAP_E] = "map-e", \
     [FIB_NODE_TYPE_VXLAN_GPE_TUNNEL] = "vxlan-gpe-tunnel", \
     [FIB_NODE_TYPE_UDP_ENCAP] = "udp-encap", \
+    [FIB_NODE_TYPE_BIER_FMASK] = "bier-fmask",	\
+    [FIB_NODE_TYPE_BIER_ENTRY] = "bier-entry",	\
 }
 
 /**
diff --git a/src/vnet/fib/fib_path.c b/src/vnet/fib/fib_path.c
index 4fccca80672..3e031929419 100644
--- a/src/vnet/fib/fib_path.c
+++ b/src/vnet/fib/fib_path.c
@@ -24,6 +24,7 @@
 #include <vnet/dpo/interface_rx_dpo.h>
 #include <vnet/dpo/mpls_disposition.h>
 #include <vnet/dpo/l2_bridge_dpo.h>
+#include <vnet/dpo/drop_dpo.h>
 
 #include <vnet/adj/adj.h>
 #include <vnet/adj/adj_mcast.h>
@@ -37,6 +38,9 @@
 #include <vnet/fib/fib_urpf_list.h>
 #include <vnet/fib/mpls_fib.h>
 #include <vnet/udp/udp_encap.h>
+#include <vnet/bier/bier_fmask.h>
+#include <vnet/bier/bier_table.h>
+#include <vnet/bier/bier_imp.h>
 
 /**
  * Enurmeration of path types
@@ -83,9 +87,21 @@ typedef enum fib_path_type_t_ {
      */
     FIB_PATH_TYPE_RECEIVE,
     /**
+     * bier-imp. it's via a BIER imposition.
+     */
+    FIB_PATH_TYPE_BIER_IMP,
+    /**
+     * bier-fmask. it's via a BIER ECMP-table.
+     */
+    FIB_PATH_TYPE_BIER_TABLE,
+    /**
+     * bier-fmask. it's via a BIER f-mask.
+     */
+    FIB_PATH_TYPE_BIER_FMASK,
+    /**
      * Marker. Add new types before this one, then update it.
      */
-    FIB_PATH_TYPE_LAST = FIB_PATH_TYPE_RECEIVE,
+    FIB_PATH_TYPE_LAST = FIB_PATH_TYPE_BIER_FMASK,
 } __attribute__ ((packed)) fib_path_type_t;
 
 /**
@@ -103,10 +119,15 @@ typedef enum fib_path_type_t_ {
     [FIB_PATH_TYPE_INTF_RX]           = "intf-rx",	        \
     [FIB_PATH_TYPE_UDP_ENCAP]         = "udp-encap",	        \
     [FIB_PATH_TYPE_RECEIVE]           = "receive",	        \
+    [FIB_PATH_TYPE_BIER_IMP]          = "bier-imp",	        \
+    [FIB_PATH_TYPE_BIER_TABLE]        = "bier-table",	        \
+    [FIB_PATH_TYPE_BIER_FMASK]        = "bier-fmask",	        \
 }
 
-#define FOR_EACH_FIB_PATH_TYPE(_item) \
-    for (_item = FIB_PATH_TYPE_FIRST; _item <= FIB_PATH_TYPE_LAST; _item++)
+#define FOR_EACH_FIB_PATH_TYPE(_item)           \
+    for (_item = FIB_PATH_TYPE_FIRST;           \
+         _item <= FIB_PATH_TYPE_LAST;           \
+         _item++)
 
 /**
  * Enurmeration of path operational (i.e. derived) attributes
@@ -252,12 +273,42 @@ typedef struct fib_path_t_ {
                     mpls_eos_bit_t fp_eos;
                 };
 	    } fp_nh;
-	    /**
-	     * The FIB table index in which to find the next-hop.
-	     */
-	    fib_node_index_t fp_tbl_id;
+            union {
+                /**
+                 * The FIB table index in which to find the next-hop.
+                 */
+                fib_node_index_t fp_tbl_id;
+                /**
+                 * The BIER FIB the fmask is in
+                 */
+                index_t fp_bier_fib;
+            };
 	} recursive;
 	struct {
+            /**
+             * The next-hop
+             */
+            ip46_address_t fp_nh;
+            /**
+             * The BIER FIB the fmask is in
+             */
+            index_t fp_bier_fib;
+	} bier_fmask;
+	struct {
+            /**
+             * The BIER table's ID
+             */
+            bier_table_id_t fp_bier_tbl;
+	} bier_table;
+	struct {
+            /**
+             * The BIER imposition object
+             * this is part of the path's key, since the index_t
+             * of an imposition object is the object's key.
+             */
+            index_t fp_bier_imp;
+	} bier_imp;
+	struct {
 	    /**
 	     * The FIB index in which to perfom the next lookup
 	     */
@@ -311,11 +362,21 @@ typedef struct fib_path_t_ {
      */
     fib_path_oper_flags_t fp_oper_flags;
 
-    /**
-     * the resolving via fib. not part of the union, since it it not part
-     * of the path's hash.
-     */
-    fib_node_index_t fp_via_fib;
+    union {
+        /**
+         * the resolving via fib. not part of the union, since it it not part
+         * of the path's hash.
+         */
+        fib_node_index_t fp_via_fib;
+        /**
+         * the resolving bier-fmask
+         */
+        index_t fp_via_bier_fmask;
+        /**
+         * the resolving bier-table
+         */
+        index_t fp_via_bier_tbl;
+    };
 
     /**
      * The Data-path objects through which this path resolves for IP.
@@ -494,6 +555,31 @@ format_fib_path (u8 * s, va_list * args)
     case FIB_PATH_TYPE_UDP_ENCAP:
         s = format (s, " UDP-encap ID:%d", path->udp_encap.fp_udp_encap_id);
         break;
+    case FIB_PATH_TYPE_BIER_TABLE:
+        s = format (s, "via bier-table:[%U}",
+                    format_bier_table_id,
+                    &path->bier_table.fp_bier_tbl);
+        s = format (s, " via-dpo:[%U:%d]",
+                    format_dpo_type, path->fp_dpo.dpoi_type,
+                    path->fp_dpo.dpoi_index);
+        break;
+    case FIB_PATH_TYPE_BIER_FMASK:
+        s = format (s, "via %U",
+                    format_ip46_address,
+                    &path->bier_fmask.fp_nh,
+                    IP46_TYPE_ANY);
+	s = format (s, " in BIER-fib:%d",
+		    path->bier_fmask.fp_bier_fib,
+		    path->fp_via_fib); 
+	s = format (s, " via-fmask:%d", path->fp_via_bier_fmask); 
+	s = format (s, " via-dpo:[%U:%d]",
+		    format_dpo_type, path->fp_dpo.dpoi_type, 
+		    path->fp_dpo.dpoi_index);
+	break;
+    case FIB_PATH_TYPE_BIER_IMP:
+        s = format (s, "via %U", format_bier_imp,
+                    path->bier_imp.fp_bier_imp, 0, BIER_SHOW_BRIEF);
+        break;
     case FIB_PATH_TYPE_RECEIVE:
     case FIB_PATH_TYPE_INTF_RX:
     case FIB_PATH_TYPE_SPECIAL:
@@ -742,6 +828,28 @@ fib_path_recursive_adj_update (fib_path_t *path,
 }
 
 /*
+ * re-evaulate the forwarding state for a via fmask path
+ */
+static void
+fib_path_bier_fmask_update (fib_path_t *path,
+                            dpo_id_t *dpo)
+{
+    bier_fmask_contribute_forwarding(path->fp_via_bier_fmask, dpo);
+
+    /*
+     * if we are stakcing on the drop, then the path is not resolved
+     */
+    if (dpo_is_drop(dpo))
+    {
+        path->fp_oper_flags &= ~FIB_PATH_OPER_FLAG_RESOLVED;
+    }
+    else
+    {
+        path->fp_oper_flags |= FIB_PATH_OPER_FLAG_RESOLVED;
+    }
+}
+
+/*
  * fib_path_is_permanent_drop
  *
  * Return !0 if the path is configured to permanently drop,
@@ -786,6 +894,20 @@ fib_path_unresolve (fib_path_t *path)
 	    path->fp_via_fib = FIB_NODE_INDEX_INVALID;
 	}
 	break;
+    case FIB_PATH_TYPE_BIER_FMASK:
+	if (FIB_NODE_INDEX_INVALID != path->fp_via_bier_fmask)
+	{
+	    bier_fmask_child_remove(path->fp_via_bier_fmask,
+                                    path->fp_sibling);
+	    path->fp_via_bier_fmask = FIB_NODE_INDEX_INVALID;
+	}
+	break;
+    case FIB_PATH_TYPE_BIER_IMP:
+        bier_imp_unlock(path->fp_dpo.dpoi_index);
+	break;
+    case FIB_PATH_TYPE_BIER_TABLE:
+        bier_table_ecmp_unlock(path->fp_via_bier_tbl);
+        break;
     case FIB_PATH_TYPE_ATTACHED_NEXT_HOP:
 	adj_child_remove(path->fp_dpo.dpoi_index,
 			 path->fp_sibling);
@@ -890,6 +1012,30 @@ fib_path_back_walk_notify (fib_node_t *node,
 	    return (FIB_NODE_BACK_WALK_CONTINUE);
 	}
 	break;
+    case FIB_PATH_TYPE_BIER_FMASK:
+	if (FIB_NODE_BW_REASON_FLAG_EVALUATE & ctx->fnbw_reason)
+	{
+	    /*
+	     * update to use the BIER fmask's new forwading
+	     */
+	    fib_path_bier_fmask_update(path, &path->fp_dpo);
+	}
+	if ((FIB_NODE_BW_REASON_FLAG_ADJ_UPDATE & ctx->fnbw_reason) ||
+            (FIB_NODE_BW_REASON_FLAG_ADJ_DOWN   & ctx->fnbw_reason))
+	{
+	    /*
+	     * ADJ updates (complete<->incomplete) do not need to propagate to
+	     * recursive entries.
+	     * The only reason its needed as far back as here, is that the adj
+	     * and the incomplete adj are a different DPO type, so the LBs need
+	     * to re-stack.
+	     * If this walk was quashed in the fib_entry, then any non-fib_path
+	     * children (like tunnels that collapse out the LB when they stack)
+	     * would not see the update.
+	     */
+	    return (FIB_NODE_BACK_WALK_CONTINUE);
+	}
+	break;
     case FIB_PATH_TYPE_ATTACHED_NEXT_HOP:
 	/*
 FIXME comment
@@ -1043,6 +1189,8 @@ FIXME comment
     case FIB_PATH_TYPE_SPECIAL:
     case FIB_PATH_TYPE_RECEIVE:
     case FIB_PATH_TYPE_EXCLUSIVE:
+    case FIB_PATH_TYPE_BIER_TABLE:
+    case FIB_PATH_TYPE_BIER_IMP:
 	/*
 	 * these path types have no parents. so to be
 	 * walked from one is unexpected.
@@ -1164,6 +1312,22 @@ fib_path_create (fib_node_index_t pl_index,
         path->deag.fp_tbl_id = rpath->frp_fib_index;
         path->deag.fp_rpf_id = rpath->frp_rpf_id;
     }
+    else if (rpath->frp_flags & FIB_ROUTE_PATH_BIER_FMASK)
+    {
+        path->fp_type = FIB_PATH_TYPE_BIER_FMASK;
+        path->bier_fmask.fp_nh = rpath->frp_addr;
+        path->bier_fmask.fp_bier_fib = rpath->frp_bier_fib_index;
+    }
+    else if (rpath->frp_flags & FIB_ROUTE_PATH_BIER_IMP)
+    {
+        path->fp_type = FIB_PATH_TYPE_BIER_IMP;
+        path->bier_imp.fp_bier_imp = rpath->frp_bier_imp;
+    }
+    else if (rpath->frp_flags & FIB_ROUTE_PATH_BIER_TABLE)
+    {
+        path->fp_type = FIB_PATH_TYPE_BIER_TABLE;
+        path->bier_table.fp_bier_tbl = rpath->frp_bier_tbl;
+    }
     else if (~0 != rpath->frp_sw_if_index)
     {
         if (ip46_address_is_zero(&rpath->frp_addr))
@@ -1204,7 +1368,7 @@ fib_path_create (fib_node_index_t pl_index,
 	    {
 		path->recursive.fp_nh.fp_ip = rpath->frp_addr;
 	    }
-	    path->recursive.fp_tbl_id = rpath->frp_fib_index;
+            path->recursive.fp_tbl_id = rpath->frp_fib_index;
 	}
     }
 
@@ -1386,6 +1550,24 @@ fib_path_cmp_i (const fib_path_t *path1,
 		res = (path1->recursive.fp_tbl_id - path2->recursive.fp_tbl_id);
 	    }
 	    break;
+	case FIB_PATH_TYPE_BIER_FMASK:
+	    res = ip46_address_cmp(&path1->bier_fmask.fp_nh,
+				   &path2->bier_fmask.fp_nh);
+ 
+	    if (0 == res)
+	    {
+		res = (path1->bier_fmask.fp_bier_fib -
+                       path2->bier_fmask.fp_bier_fib);
+	    }
+	    break;
+	case FIB_PATH_TYPE_BIER_IMP:
+            res = (path1->bier_imp.fp_bier_imp -
+                   path2->bier_imp.fp_bier_imp);
+	    break;
+        case FIB_PATH_TYPE_BIER_TABLE:
+            res = bier_table_id_cmp(&path1->bier_table.fp_bier_tbl,
+                                    &path2->bier_table.fp_bier_tbl);
+            break;
 	case FIB_PATH_TYPE_DEAG:
 	    res = (path1->deag.fp_tbl_id - path2->deag.fp_tbl_id);
 	    if (0 == res)
@@ -1510,6 +1692,22 @@ fib_path_cmp_w_route_path (fib_node_index_t path_index,
                 res = (path->recursive.fp_tbl_id - rpath->frp_fib_index);
             }
 	    break;
+	case FIB_PATH_TYPE_BIER_FMASK:
+            res = ip46_address_cmp(&path->bier_fmask.fp_nh,
+                                   &rpath->frp_addr);
+
+            if (0 == res)
+            {
+                res = (path->bier_fmask.fp_bier_fib - rpath->frp_bier_fib_index);
+            }
+	    break;
+	case FIB_PATH_TYPE_BIER_IMP:
+            res = (path->bier_imp.fp_bier_imp - rpath->frp_bier_imp);
+	    break;
+        case FIB_PATH_TYPE_BIER_TABLE:
+            res = bier_table_id_cmp(&path->bier_table.fp_bier_tbl,
+                                    &rpath->frp_bier_tbl);
+            break;
 	case FIB_PATH_TYPE_INTF_RX:
 	    res = (path->intf_rx.fp_interface - rpath->frp_sw_if_index);
             break;
@@ -1623,6 +1821,9 @@ fib_path_recursive_loop_detect (fib_node_index_t path_index,
     case FIB_PATH_TYPE_INTF_RX:
     case FIB_PATH_TYPE_UDP_ENCAP:
     case FIB_PATH_TYPE_EXCLUSIVE:
+    case FIB_PATH_TYPE_BIER_FMASK:
+    case FIB_PATH_TYPE_BIER_TABLE:
+    case FIB_PATH_TYPE_BIER_IMP:
 	/*
 	 * these path types cannot be part of a loop, since they are the leaves
 	 * of the graph.
@@ -1742,6 +1943,53 @@ fib_path_resolve (fib_node_index_t path_index)
 
 	break;
     }
+    case FIB_PATH_TYPE_BIER_FMASK:
+    {
+        /*
+         * Find the BIER f-mask to link to
+         */
+        bier_fmask_id_t fmid = {
+            .bfmi_nh = path->bier_fmask.fp_nh,
+            .bfmi_hdr_type = BIER_HDR_O_MPLS,
+        };
+
+        ASSERT(FIB_NODE_INDEX_INVALID == path->fp_via_bier_fmask);
+
+        path->fp_via_bier_fmask = bier_fmask_db_find(path->bier_fmask.fp_bier_fib,
+                                                &fmid);
+
+        /*
+         * become a dependent child of the entry so the path is
+         * informed when the forwarding for the entry changes.
+         */
+        path->fp_sibling = bier_fmask_child_add(path->fp_via_bier_fmask,
+                                                FIB_NODE_TYPE_PATH,
+                                                fib_path_get_index(path));
+
+        fib_path_bier_fmask_update(path, &path->fp_dpo);
+
+        break;
+    }
+    case FIB_PATH_TYPE_BIER_IMP:
+        bier_imp_lock(path->bier_imp.fp_bier_imp);
+        bier_imp_contribute_forwarding(path->bier_imp.fp_bier_imp,
+                                       DPO_PROTO_IP4,
+                                       &path->fp_dpo);
+        break;
+    case FIB_PATH_TYPE_BIER_TABLE:
+    {
+        /*
+         * Find/create the BIER table to link to
+         */
+        ASSERT(FIB_NODE_INDEX_INVALID == path->fp_via_bier_tbl);
+
+        path->fp_via_bier_tbl =
+            bier_table_ecmp_create_and_lock(&path->bier_table.fp_bier_tbl);
+
+        bier_table_contribute_forwarding(path->fp_via_bier_tbl,
+                                         &path->fp_dpo);
+        break;
+    }
     case FIB_PATH_TYPE_SPECIAL:
 	/*
 	 * Resolve via the drop
@@ -1833,11 +2081,45 @@ fib_path_get_resolving_interface (fib_node_index_t path_index)
     case FIB_PATH_TYPE_SPECIAL:
     case FIB_PATH_TYPE_DEAG:
     case FIB_PATH_TYPE_EXCLUSIVE:
+    case FIB_PATH_TYPE_BIER_FMASK:
+    case FIB_PATH_TYPE_BIER_TABLE:
+    case FIB_PATH_TYPE_BIER_IMP:
 	break;
     }
     return (~0);
 }
 
+index_t
+fib_path_get_resolving_index (fib_node_index_t path_index)
+{
+    fib_path_t *path;
+
+    path = fib_path_get(path_index);
+
+    switch (path->fp_type)
+    {
+    case FIB_PATH_TYPE_ATTACHED_NEXT_HOP:
+    case FIB_PATH_TYPE_ATTACHED:
+    case FIB_PATH_TYPE_RECEIVE:
+    case FIB_PATH_TYPE_INTF_RX:
+    case FIB_PATH_TYPE_SPECIAL:
+    case FIB_PATH_TYPE_DEAG:
+    case FIB_PATH_TYPE_EXCLUSIVE:
+        break;
+    case FIB_PATH_TYPE_UDP_ENCAP:
+	return (path->udp_encap.fp_udp_encap_id);
+    case FIB_PATH_TYPE_RECURSIVE:
+	return (path->fp_via_fib);
+    case FIB_PATH_TYPE_BIER_FMASK:
+ 	return (path->fp_via_bier_fmask);
+   case FIB_PATH_TYPE_BIER_TABLE:
+       return (path->fp_via_bier_tbl);
+   case FIB_PATH_TYPE_BIER_IMP:
+       return (path->bier_imp.fp_bier_imp);
+    }
+    return (~0);
+}
+
 adj_index_t
 fib_path_get_adj (fib_node_index_t path_index)
 {
@@ -1877,6 +2159,23 @@ fib_path_get_preference (fib_node_index_t path_index)
     return (path->fp_preference);
 }
 
+u32
+fib_path_get_rpf_id (fib_node_index_t path_index)
+{
+    fib_path_t *path;
+
+    path = fib_path_get(path_index);
+
+    ASSERT(path);
+
+    if (FIB_PATH_CFG_FLAG_RPF_ID & path->fp_cfg_flags)
+    {
+        return (path->deag.fp_rpf_id);
+    }
+
+    return (~0);
+}
+
 /**
  * @brief Contribute the path's adjacency to the list passed.
  * By calling this function over all paths, recursively, a child
@@ -1937,6 +2236,9 @@ fib_path_contribute_urpf (fib_node_index_t path_index,
     case FIB_PATH_TYPE_RECEIVE:
     case FIB_PATH_TYPE_INTF_RX:
     case FIB_PATH_TYPE_UDP_ENCAP:
+    case FIB_PATH_TYPE_BIER_FMASK:
+    case FIB_PATH_TYPE_BIER_TABLE:
+    case FIB_PATH_TYPE_BIER_IMP:
 	/*
 	 * these path types don't link to an adj
 	 */
@@ -1979,6 +2281,9 @@ fib_path_stack_mpls_disp (fib_node_index_t path_index,
     case FIB_PATH_TYPE_UDP_ENCAP:
     case FIB_PATH_TYPE_EXCLUSIVE:
     case FIB_PATH_TYPE_SPECIAL:
+    case FIB_PATH_TYPE_BIER_FMASK:
+    case FIB_PATH_TYPE_BIER_TABLE:
+    case FIB_PATH_TYPE_BIER_IMP:
         break;
     }
 }
@@ -2036,8 +2341,9 @@ fib_path_contribute_forwarding (fib_node_index_t path_index,
 	    }
 	    case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
 	    case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
-	    break;
-            }
+	    case FIB_FORW_CHAIN_TYPE_BIER:
+		break;
+	    }
             break;
 	case FIB_PATH_TYPE_RECURSIVE:
 	    switch (fct)
@@ -2048,6 +2354,7 @@ fib_path_contribute_forwarding (fib_node_index_t path_index,
 	    case FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS:
 	    case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
 	    case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
+	    case FIB_FORW_CHAIN_TYPE_BIER:
 		fib_path_recursive_adj_update(path, fct, dpo);
 		break;
 	    case FIB_FORW_CHAIN_TYPE_ETHERNET:
@@ -2056,6 +2363,47 @@ fib_path_contribute_forwarding (fib_node_index_t path_index,
 		break;
 	    }
 	    break;
+        case FIB_PATH_TYPE_BIER_TABLE:
+            switch (fct)
+            {
+            case FIB_FORW_CHAIN_TYPE_BIER:
+                bier_table_contribute_forwarding(path->fp_via_bier_tbl, dpo);
+                break;
+            case FIB_FORW_CHAIN_TYPE_MPLS_EOS:
+            case FIB_FORW_CHAIN_TYPE_UNICAST_IP4:
+            case FIB_FORW_CHAIN_TYPE_UNICAST_IP6:
+            case FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS:
+            case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
+            case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
+            case FIB_FORW_CHAIN_TYPE_ETHERNET:
+            case FIB_FORW_CHAIN_TYPE_NSH:
+                ASSERT(0);
+                break;
+            }
+            break;
+	case FIB_PATH_TYPE_BIER_FMASK:
+	    switch (fct)
+	    {
+	    case FIB_FORW_CHAIN_TYPE_BIER:
+		fib_path_bier_fmask_update(path, dpo);
+		break;
+	    case FIB_FORW_CHAIN_TYPE_MPLS_EOS:
+	    case FIB_FORW_CHAIN_TYPE_UNICAST_IP4:
+	    case FIB_FORW_CHAIN_TYPE_UNICAST_IP6:
+	    case FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS:
+	    case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
+	    case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
+	    case FIB_FORW_CHAIN_TYPE_ETHERNET:
+	    case FIB_FORW_CHAIN_TYPE_NSH:
+		ASSERT(0);
+		break;
+	    }
+	    break;
+	case FIB_PATH_TYPE_BIER_IMP:
+            bier_imp_contribute_forwarding(path->bier_imp.fp_bier_imp,
+                                           fib_forw_chain_type_to_dpo_proto(fct),
+                                           dpo);
+	    break;
 	case FIB_PATH_TYPE_DEAG:
             switch (fct)
 	    {
@@ -2074,6 +2422,8 @@ fib_path_contribute_forwarding (fib_node_index_t path_index,
 		break;
 	    case FIB_FORW_CHAIN_TYPE_MCAST_IP4:
 	    case FIB_FORW_CHAIN_TYPE_MCAST_IP6:
+	    case FIB_FORW_CHAIN_TYPE_BIER:
+		break;
 	    case FIB_FORW_CHAIN_TYPE_ETHERNET:
 	    case FIB_FORW_CHAIN_TYPE_NSH:
 		ASSERT(0);
@@ -2097,6 +2447,7 @@ fib_path_contribute_forwarding (fib_node_index_t path_index,
 	    case FIB_FORW_CHAIN_TYPE_MPLS_EOS:
 	    case FIB_FORW_CHAIN_TYPE_ETHERNET:
 	    case FIB_FORW_CHAIN_TYPE_NSH:
+            case FIB_FORW_CHAIN_TYPE_BIER:
                 {
                     adj_index_t ai;
 
@@ -2263,6 +2614,10 @@ fib_path_encode (fib_node_index_t path_list_index,
         api_rpath->rpath.frp_sw_if_index = path->attached_next_hop.fp_interface;
         api_rpath->rpath.frp_addr = path->attached_next_hop.fp_nh;
         break;
+      case FIB_PATH_TYPE_BIER_FMASK:
+        api_rpath->rpath.frp_fib_index = path->bier_fmask.fp_bier_fib;
+        api_rpath->rpath.frp_addr = path->bier_fmask.fp_nh;
+        break;
       case FIB_PATH_TYPE_SPECIAL:
         break;
       case FIB_PATH_TYPE_DEAG:
diff --git a/src/vnet/fib/fib_path.h b/src/vnet/fib/fib_path.h
index 3a0544c1c42..861bda95933 100644
--- a/src/vnet/fib/fib_path.h
+++ b/src/vnet/fib/fib_path.h
@@ -35,6 +35,7 @@
 
 #include <vnet/fib/fib_types.h>
 #include <vnet/adj/adj_types.h>
+#include <vnet/bier/bier_types.h>
 
 /**
  * Enurmeration of path configuration attributes
@@ -173,8 +174,10 @@ extern adj_index_t fib_path_get_adj(fib_node_index_t path_index);
 extern int fib_path_recursive_loop_detect(fib_node_index_t path_index,
 					  fib_node_index_t **entry_indicies);
 extern u32 fib_path_get_resolving_interface(fib_node_index_t fib_entry_index);
+extern index_t fib_path_get_resolving_index(fib_node_index_t path_index);
 extern u16 fib_path_get_weight(fib_node_index_t path_index);
 extern u16 fib_path_get_preference(fib_node_index_t path_index);
+extern u32 fib_path_get_rpf_id(fib_node_index_t path_index);
 
 extern void fib_path_module_init(void);
 extern fib_path_list_walk_rc_t fib_path_encode(fib_node_index_t path_list_index,
diff --git a/src/vnet/fib/fib_path_list.c b/src/vnet/fib/fib_path_list.c
index f30fd7ea0a4..f9477406ed3 100644
--- a/src/vnet/fib/fib_path_list.c
+++ b/src/vnet/fib/fib_path_list.c
@@ -583,6 +583,11 @@ fib_path_list_get_n_paths (fib_node_index_t path_list_index)
 {
     fib_path_list_t *path_list;
 
+    if (FIB_NODE_INDEX_INVALID == path_list_index)
+    {
+        return (0);
+    }
+
     path_list = fib_path_list_get(path_list_index);
 
     return (vec_len(path_list->fpl_paths));
diff --git a/src/vnet/fib/fib_table.c b/src/vnet/fib/fib_table.c
index d5625d83673..4dd6e7c5421 100644
--- a/src/vnet/fib/fib_table.c
+++ b/src/vnet/fib/fib_table.c
@@ -953,6 +953,7 @@ fib_table_get_flow_hash_config (u32 fib_index,
 
     return (fib->ft_flow_hash_config);
 }
+
 flow_hash_config_t
 fib_table_get_default_flow_hash_config (fib_protocol_t proto)
 {
diff --git a/src/vnet/fib/fib_table.h b/src/vnet/fib/fib_table.h
index 6c2640458b9..d2522aa50c8 100644
--- a/src/vnet/fib/fib_table.h
+++ b/src/vnet/fib/fib_table.h
@@ -697,7 +697,7 @@ extern u32 fib_table_create_and_lock(fib_protocol_t proto,
  *  The index of the FIB
  *
  * @paran proto
- *  The protocol of the FIB (and thus the entries therein)
+ *  The protocol the packets the flow hash will be calculated for.
  *
  * @return The flow hash config
  */
diff --git a/src/vnet/fib/fib_test.c b/src/vnet/fib/fib_test.c
index 66a38236216..2658eb27811 100644
--- a/src/vnet/fib/fib_test.c
+++ b/src/vnet/fib/fib_test.c
@@ -13,6 +13,7 @@
  * limitations under the License.
  */
 
+#include <vnet/fib/fib_test.h>
 #include <vnet/fib/ip6_fib.h>
 #include <vnet/fib/ip4_fib.h>
 #include <vnet/fib/mpls_fib.h>
@@ -544,15 +545,31 @@ fib_test_validate_lb_v (const load_balance_t *lb,
 			dpo->dpoi_index,
                         exp->lb.lb);
 	    break;
-	case FT_LB_SPECIAL:
-	    FIB_TEST_I((DPO_DROP == dpo->dpoi_type),
-		       "bucket %d stacks on %U",
-		       bucket,
-		       format_dpo_type, dpo->dpoi_type);
-	    FIB_TEST_LB((exp->special.adj == dpo->dpoi_index),
-			"bucket %d stacks on drop %d",
+	case FT_LB_BIER_TABLE:
+	    FIB_TEST_LB((DPO_BIER_TABLE == dpo->dpoi_type),
+                        "bucket %d stacks on %U",
+                        bucket,
+                        format_dpo_type, dpo->dpoi_type);
+	    FIB_TEST_LB((exp->bier.table == dpo->dpoi_index),
+			"bucket %d stacks on lb %d",
+			bucket,
+			exp->bier.table);
+            break;
+	case FT_LB_BIER_FMASK:
+	    FIB_TEST_LB((DPO_BIER_FMASK == dpo->dpoi_type),
+                        "bucket %d stacks on %U",
+                        bucket,
+                        format_dpo_type, dpo->dpoi_type);
+	    FIB_TEST_LB((exp->bier.fmask == dpo->dpoi_index),
+			"bucket %d stacks on lb %d",
 			bucket,
-			exp->special.adj);
+			exp->bier.fmask);
+            break;
+	case FT_LB_DROP:
+	    FIB_TEST_LB((DPO_DROP == dpo->dpoi_type),
+                        "bucket %d stacks on %U",
+                        bucket,
+                        format_dpo_type, dpo->dpoi_type);
 	    break;
 	}
     }
@@ -560,6 +577,29 @@ fib_test_validate_lb_v (const load_balance_t *lb,
 }
 
 int
+fib_test_validate_lb (const dpo_id_t *dpo,
+                     u16 n_buckets,
+                     ...)
+{
+    const load_balance_t *lb;
+    va_list ap;
+    int res;
+
+    va_start(ap, n_buckets);
+
+    FIB_TEST_LB((DPO_LOAD_BALANCE == dpo->dpoi_type),
+               "Entry links to %U",
+               format_dpo_type, dpo->dpoi_type);
+    lb = load_balance_get(dpo->dpoi_index);
+
+    res = fib_test_validate_lb_v(lb, n_buckets, &ap);
+
+    va_end(ap);
+
+    return (res);
+}
+
+int
 fib_test_validate_entry (fib_node_index_t fei,
 			 fib_forward_chain_type_t fct,
 			 int n_buckets,
@@ -6678,13 +6718,10 @@ fib_test_label (void)
      * remove the other path with a valid label
      */
     fib_test_lb_bucket_t bucket_drop = {
-	.type = FT_LB_SPECIAL,
-	.special = {
-	    .adj = DPO_PROTO_IP4,
-	},
+	.type = FT_LB_DROP,
     };
     fib_test_lb_bucket_t mpls_bucket_drop = {
-	.type = FT_LB_SPECIAL,
+	.type = FT_LB_DROP,
 	.special = {
 	    .adj = DPO_PROTO_MPLS,
 	},
@@ -8497,13 +8534,13 @@ lfib_test (void)
      * A recursive via a label that does not exist
      */
     fib_test_lb_bucket_t bucket_drop = {
-	.type = FT_LB_SPECIAL,
+	.type = FT_LB_DROP,
 	.special = {
 	    .adj = DPO_PROTO_IP4,
 	},
     };
     fib_test_lb_bucket_t mpls_bucket_drop = {
-	.type = FT_LB_SPECIAL,
+	.type = FT_LB_DROP,
 	.special = {
 	    .adj = DPO_PROTO_MPLS,
 	},
diff --git a/src/vnet/fib/fib_test.h b/src/vnet/fib/fib_test.h
index a9463d60830..f3d8346aab3 100644
--- a/src/vnet/fib/fib_test.h
+++ b/src/vnet/fib/fib_test.h
@@ -17,16 +17,24 @@
 #define __FIB_TEST_H__
 
 #include <vnet/fib/fib_types.h>
+#include <vnet/mpls/mpls_types.h>
+#include <vnet/fib/fib_types.h>
+#include <vnet/mpls/packet.h>
+#include <vnet/dpo/load_balance.h>
+#include <vnet/adj/adj_types.h>
+#include <vnet/dpo/replicate_dpo.h>
 
 typedef enum fib_test_lb_bucket_type_t_ {
     FT_LB_LABEL_O_ADJ,
     FT_LB_LABEL_STACK_O_ADJ,
     FT_LB_LABEL_O_LB,
     FT_LB_O_LB,
-    FT_LB_SPECIAL,
-    FT_LB_ADJ,
     FT_LB_INTF,
     FT_LB_L2,
+    FT_LB_BIER_TABLE,
+    FT_LB_BIER_FMASK,
+    FT_LB_DROP,
+    FT_LB_ADJ,
 } fib_test_lb_bucket_type_t;
 
 typedef struct fib_test_lb_bucket_t_ {
@@ -68,6 +76,13 @@ typedef struct fib_test_lb_bucket_t_ {
 	{
 	    index_t adj;
 	} special;
+        struct
+        {
+            union {
+                index_t table;
+                index_t fmask;
+            };
+	} bier;
     };
 } fib_test_lb_bucket_t;
 
@@ -105,9 +120,14 @@ extern int fib_test_validate_lb_v(const load_balance_t *lb,
                                   u16 n_buckets,
                                   va_list *ap);
 
+extern int fib_test_validate_lb(const dpo_id_t *dpo,
+				u16 n_buckets,
+				...);
+
 extern int fib_test_validate_entry(fib_node_index_t fei,
                                    fib_forward_chain_type_t fct,
                                    int n_buckets,
                                    ...);
 
+
 #endif
diff --git a/src/vnet/fib/fib_types.c b/src/vnet/fib/fib_types.c
index d84642d29da..656966b2331 100644
--- a/src/vnet/fib/fib_types.c
+++ b/src/vnet/fib/fib_types.c
@@ -272,6 +272,8 @@ fib_forw_chain_type_from_dpo_proto (dpo_proto_t proto)
 	return (FIB_FORW_CHAIN_TYPE_ETHERNET);
     case DPO_PROTO_NSH:
         return (FIB_FORW_CHAIN_TYPE_NSH);
+    case DPO_PROTO_BIER:
+	return (FIB_FORW_CHAIN_TYPE_BIER);
     }
     ASSERT(0);
     return (FIB_FORW_CHAIN_TYPE_UNICAST_IP4);
@@ -293,6 +295,7 @@ fib_forw_chain_type_to_link_type (fib_forward_chain_type_t fct)
     case FIB_FORW_CHAIN_TYPE_NSH:
         return (VNET_LINK_NSH);
     case FIB_FORW_CHAIN_TYPE_MPLS_EOS:
+    case FIB_FORW_CHAIN_TYPE_BIER:
 	/*
 	 * insufficient information to to convert
 	 */
@@ -319,6 +322,8 @@ fib_forw_chain_type_to_dpo_proto (fib_forward_chain_type_t fct)
 	return (DPO_PROTO_ETHERNET);
     case FIB_FORW_CHAIN_TYPE_NSH:
         return (DPO_PROTO_NSH);
+    case FIB_FORW_CHAIN_TYPE_BIER:
+	return (DPO_PROTO_BIER);
     case FIB_FORW_CHAIN_TYPE_MPLS_EOS:
     case FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS:
 	return (DPO_PROTO_MPLS);
diff --git a/src/vnet/fib/fib_types.h b/src/vnet/fib/fib_types.h
index ec7f1b374dd..0a4b1699609 100644
--- a/src/vnet/fib/fib_types.h
+++ b/src/vnet/fib/fib_types.h
@@ -20,6 +20,7 @@
 #include <vnet/ip/ip6_packet.h>
 #include <vnet/mpls/packet.h>
 #include <vnet/dpo/dpo.h>
+#include <vnet/bier/bier_types.h>
 
 /**
  * A typedef of a node index.
@@ -50,6 +51,12 @@ typedef enum fib_protocol_t_ {
 #define FIB_PROTOCOL_MAX (FIB_PROTOCOL_MPLS + 1)
 
 /**
+ * Definition outside of enum so it does not need to be included in non-defaulted
+ * switch statements
+ */
+#define FIB_PROTOCOL_IP_MAX (FIB_PROTOCOL_IP6 + 1)
+
+/**
  * Not part of the enum so it does not have to be handled in switch statements
  */
 #define FIB_PROTOCOL_NONE (FIB_PROTOCOL_MAX+1)
@@ -89,6 +96,10 @@ typedef enum fib_forward_chain_type_t_ {
      */
     FIB_FORW_CHAIN_TYPE_MPLS_NON_EOS,
     /**
+     * Contribute an object that is to be used to forward BIER packets.
+     */
+    FIB_FORW_CHAIN_TYPE_BIER,
+    /**
      * Contribute an object that is to be used to forward end-of-stack
      * MPLS packets. This is a convenient ID for clients. A real EOS chain
      * must be pay-load protocol specific. This
@@ -117,6 +128,7 @@ typedef enum fib_forward_chain_type_t_ {
 
 #define FIB_FORW_CHAINS {					\
     [FIB_FORW_CHAIN_TYPE_ETHERNET]      = "ethernet",     	\
+    [FIB_FORW_CHAIN_TYPE_BIER]          = "bier",     	        \
     [FIB_FORW_CHAIN_TYPE_UNICAST_IP4]   = "unicast-ip4",	\
     [FIB_FORW_CHAIN_TYPE_UNICAST_IP6]   = "unicast-ip6",	\
     [FIB_FORW_CHAIN_TYPE_MCAST_IP4]     = "multicast-ip4",	\
@@ -231,6 +243,11 @@ extern dpo_proto_t fib_proto_to_dpo(fib_protocol_t fib_proto);
 extern fib_protocol_t dpo_proto_to_fib(dpo_proto_t dpo_proto);
 
 /**
+ * Convert from BIER next-hop proto to FIB proto
+ */
+extern fib_protocol_t bier_hdr_proto_to_fib(bier_hdr_proto_id_t bproto);
+
+/**
  * Enurmeration of special path/entry types
  */
 typedef enum fib_special_type_t_ {
@@ -311,6 +328,18 @@ typedef enum fib_route_path_flags_t_
      * A path via a UDP encap object.
      */
     FIB_ROUTE_PATH_UDP_ENCAP = (1 << 9),
+    /**
+     * A path that resolves via a BIER F-Mask
+     */
+    FIB_ROUTE_PATH_BIER_FMASK = (1 << 10),
+    /**
+     * A path that resolves via a BIER [ECMP] Table
+     */
+    FIB_ROUTE_PATH_BIER_TABLE = (1 << 11),
+    /**
+     * A path that resolves via a BIER impostion object
+     */
+    FIB_ROUTE_PATH_BIER_IMP = (1 << 12),
 } fib_route_path_flags_t;
 
 /**
@@ -349,47 +378,73 @@ typedef struct fib_route_path_t_ {
     dpo_proto_t frp_proto;
 
     union {
-	/**
-	 * The next-hop address.
-	 * Will be NULL for attached paths.
-	 * Will be all zeros for attached-next-hop paths on a p2p interface
-	 * Will be all zeros for a deag path.
-	 */
-	ip46_address_t frp_addr;
-
         struct {
+            union {
+                /**
+                 * The next-hop address.
+                 * Will be NULL for attached paths.
+                 * Will be all zeros for attached-next-hop paths on a p2p interface
+                 * Will be all zeros for a deag path.
+                 */
+                ip46_address_t frp_addr;
+
+                struct {
+                    /**
+                     * The MPLS local Label to reursively resolve through.
+                     * This is valid when the path type is MPLS.
+                     */
+                    mpls_label_t frp_local_label;
+                    /**
+                     * EOS bit for the resolving label
+                     */
+                    mpls_eos_bit_t frp_eos;
+                };
+            };
+            union {
+                /**
+                 * The interface.
+                 * Will be invalid for recursive paths.
+                 */
+                u32 frp_sw_if_index;
+                /**
+                 * The RPF-ID
+                 */
+                fib_rpf_id_t frp_rpf_id;
+            };
+            union {
+                /**
+                 * The FIB index to lookup the nexthop
+                 * Only valid for recursive paths.
+                 */
+                u32 frp_fib_index;
+                /**
+                 * The BIER table to resolve the fmask in
+                 */
+                u32 frp_bier_fib_index;
+            };
             /**
-             * The MPLS local Label to reursively resolve through.
-             * This is valid when the path type is MPLS.
+             * The outgoing MPLS label Stack. NULL implies no label.
              */
-            mpls_label_t frp_local_label;
-            /**
-             * EOS bit for the resolving label
-             */
-            mpls_eos_bit_t frp_eos;
+            mpls_label_t *frp_label_stack;
         };
-    };
-    union {
         /**
-         * The interface.
-         * Will be invalid for recursive paths.
+         * A path that resolves via a BIER Table.
+         * This would be for a MPLS label at a BIER midpoint or tail
          */
-        u32 frp_sw_if_index;
+        bier_table_id_t frp_bier_tbl;
+
         /**
-         * The RPF-ID
+         * A path via a BIER imposition object.
+         * Present in an mfib path list
          */
-        fib_rpf_id_t frp_rpf_id;
+        index_t frp_bier_imp;
+
         /**
          * UDP encap ID
          */
         u32 frp_udp_encap_id;
     };
     /**
-     * The FIB index to lookup the nexthop
-     * Only valid for recursive paths.
-     */
-    u32 frp_fib_index;
-    /**
      * [un]equal cost path weight
      */
     u8 frp_weight;
@@ -403,10 +458,6 @@ typedef struct fib_route_path_t_ {
      * flags on the path
      */
     fib_route_path_flags_t frp_flags;
-    /**
-     * The outgoing MPLS label Stack. NULL implies no label.
-     */
-    mpls_label_t *frp_label_stack;
 } fib_route_path_t;
 
 /**
diff --git a/src/vnet/ip/ip.api b/src/vnet/ip/ip.api
index 85e4b8e04c3..e9c9f55a6f6 100644
--- a/src/vnet/ip/ip.api
+++ b/src/vnet/ip/ip.api
@@ -19,7 +19,7 @@
     called through a shared memory interface. 
 */
 
-vl_api_version 1.0.0
+vl_api_version 1.0.1
 
 /** \brief Add / del table request
            A table can be added multiple times, but need be deleted only once.
@@ -68,6 +68,7 @@ define ip_fib_dump
 typeonly manual_print manual_endian define fib_path
 {
   u32 sw_if_index;
+  u32 table_id;
   u8 weight;
   u8 preference;
   u8 is_local;
@@ -426,7 +427,7 @@ autoreply define ip_add_del_route
     @param context - sender context, to match reply w/ request
     @param sw_if_index - software index of the new vlan's parent interface
     @param vrf_id - fib table /vrf associated with the route
-
+    @param next_hop_afi - Use dpo_proto_t
     FIXME
 */
 autoreply define ip_mroute_add_del
@@ -438,7 +439,9 @@ autoreply define ip_mroute_add_del
   u32 entry_flags;
   u32 itf_flags;
   u32 rpf_id;
+  u32 bier_imp;
   u16 grp_address_length;
+  u8 next_hop_afi;
   u8 create_vrf_if_needed;
   u8 is_add;
   u8 is_ipv6;
diff --git a/src/vnet/ip/ip_api.c b/src/vnet/ip/ip_api.c
index 20e19205756..0fdb167312e 100644
--- a/src/vnet/ip/ip_api.c
+++ b/src/vnet/ip/ip_api.c
@@ -464,7 +464,7 @@ send_ip_mfib_details (unix_shared_memory_queue_t * q,
   if (!mp)
     return;
   memset (mp, 0, sizeof (*mp));
-  mp->_vl_msg_id = ntohs (VL_API_IP_FIB_DETAILS);
+  mp->_vl_msg_id = ntohs (VL_API_IP_MFIB_DETAILS);
   mp->context = context;
 
   mp->rpf_id = mfib_entry->mfe_rpf_id;
@@ -563,7 +563,7 @@ send_ip6_mfib_details (vpe_api_main_t * am,
   if (!mp)
     return;
   memset (mp, 0, sizeof (*mp));
-  mp->_vl_msg_id = ntohs (VL_API_IP6_FIB_DETAILS);
+  mp->_vl_msg_id = ntohs (VL_API_IP6_MFIB_DETAILS);
   mp->context = context;
 
   mp->table_id = htonl (table_id);
@@ -1283,40 +1283,45 @@ mroute_add_del_handler (u8 is_add,
 			u8 is_local,
 			u32 fib_index,
 			const mfib_prefix_t * prefix,
+			dpo_proto_t nh_proto,
 			u32 entry_flags,
 			fib_rpf_id_t rpf_id,
-			u32 next_hop_sw_if_index, u32 itf_flags)
+			u32 next_hop_sw_if_index, u32 itf_flags, u32 bier_imp)
 {
   stats_dslock_with_hint (1 /* release hint */ , 2 /* tag */ );
 
   fib_route_path_t path = {
     .frp_sw_if_index = next_hop_sw_if_index,
-    .frp_proto = fib_proto_to_dpo (prefix->fp_proto),
+    .frp_proto = nh_proto,
   };
 
   if (is_local)
     path.frp_flags |= FIB_ROUTE_PATH_LOCAL;
 
-
-  if (!is_local && ~0 == next_hop_sw_if_index)
+  if (DPO_PROTO_BIER == nh_proto)
+    {
+      path.frp_bier_imp = bier_imp;
+      path.frp_flags = FIB_ROUTE_PATH_BIER_IMP;
+    }
+  else if (!is_local && ~0 == next_hop_sw_if_index)
     {
       mfib_table_entry_update (fib_index, prefix,
 			       MFIB_SOURCE_API, rpf_id, entry_flags);
+      goto done;
+    }
+
+  if (is_add)
+    {
+      mfib_table_entry_path_update (fib_index, prefix,
+				    MFIB_SOURCE_API, &path, itf_flags);
     }
   else
     {
-      if (is_add)
-	{
-	  mfib_table_entry_path_update (fib_index, prefix,
-					MFIB_SOURCE_API, &path, itf_flags);
-	}
-      else
-	{
-	  mfib_table_entry_path_remove (fib_index, prefix,
-					MFIB_SOURCE_API, &path);
-	}
+      mfib_table_entry_path_remove (fib_index, prefix,
+				    MFIB_SOURCE_API, &path);
     }
 
+done:
   stats_dsunlock ();
   return (0);
 }
@@ -1325,9 +1330,11 @@ static int
 api_mroute_add_del_t_handler (vl_api_ip_mroute_add_del_t * mp)
 {
   fib_protocol_t fproto;
+  dpo_proto_t nh_proto;
   u32 fib_index;
   int rv;
 
+  nh_proto = mp->next_hop_afi;
   fproto = (mp->is_ipv6 ? FIB_PROTOCOL_IP6 : FIB_PROTOCOL_IP4);
   rv = add_del_mroute_check (fproto,
 			     mp->table_id,
@@ -1360,10 +1367,12 @@ api_mroute_add_del_t_handler (vl_api_ip_mroute_add_del_t * mp)
   return (mroute_add_del_handler (mp->is_add,
 				  mp->is_local,
 				  fib_index, &pfx,
+				  nh_proto,
 				  ntohl (mp->entry_flags),
 				  ntohl (mp->rpf_id),
 				  ntohl (mp->next_hop_sw_if_index),
-				  ntohl (mp->itf_flags)));
+				  ntohl (mp->itf_flags),
+				  ntohl (mp->bier_imp)));
 }
 
 void
diff --git a/src/vnet/mfib/mfib_entry.c b/src/vnet/mfib/mfib_entry.c
index 2302b9a1707..aaecf0f6eb6 100644
--- a/src/vnet/mfib/mfib_entry.c
+++ b/src/vnet/mfib/mfib_entry.c
@@ -300,7 +300,7 @@ mfib_entry_src_find (const mfib_entry_t *mfib_entry,
 
 static mfib_entry_src_t *
 mfib_entry_src_find_or_create (mfib_entry_t *mfib_entry,
-                              mfib_source_t source)
+                               mfib_source_t source)
 {
     mfib_entry_src_t *esrc;
 
@@ -535,6 +535,7 @@ mfib_entry_src_collect_forwarding (fib_node_index_t pl_index,
     case FIB_FORW_CHAIN_TYPE_MPLS_EOS:
     case FIB_FORW_CHAIN_TYPE_ETHERNET:
     case FIB_FORW_CHAIN_TYPE_NSH:
+    case FIB_FORW_CHAIN_TYPE_BIER:
         ASSERT(0);
         break;
     }
diff --git a/src/vnet/mpls/mpls.api b/src/vnet/mpls/mpls.api
index 3c817db1306..f9171bfe22e 100644
--- a/src/vnet/mpls/mpls.api
+++ b/src/vnet/mpls/mpls.api
@@ -13,7 +13,7 @@
  * limitations under the License.
  */
 
-vl_api_version 1.0.0
+vl_api_version 1.0.1
 
 /** \brief Bind/Unbind an MPLS local label to an IP prefix. i.e. create
            a per-prefix label entry.
@@ -114,6 +114,7 @@ define mpls_tunnel_dump
 typeonly manual_print manual_endian define fib_path2
 {
   u32 sw_if_index;
+  u32 table_id;
   u8 weight;
   u8 preference;
   u8 is_local;
diff --git a/src/vnet/mpls/mpls_api.c b/src/vnet/mpls/mpls_api.c
index a55daa2b0c4..f345ca6aeea 100644
--- a/src/vnet/mpls/mpls_api.c
+++ b/src/vnet/mpls/mpls_api.c
@@ -366,7 +366,7 @@ typedef struct mpls_tunnel_send_walk_ctx_t_
 static void
 send_mpls_tunnel_entry (u32 mti, void *arg)
 {
-  fib_route_path_encode_t *api_rpaths, *api_rpath;
+  fib_route_path_encode_t *api_rpaths = NULL, *api_rpath;
   mpls_tunnel_send_walk_ctx_t *ctx;
   vl_api_mpls_tunnel_details_t *mp;
   const mpls_tunnel_t *mt;
diff --git a/src/vnet/srv6/sr_policy_rewrite.c b/src/vnet/srv6/sr_policy_rewrite.c
index 2f90993a4be..514dd6584d0 100755
--- a/src/vnet/srv6/sr_policy_rewrite.c
+++ b/src/vnet/srv6/sr_policy_rewrite.c
@@ -367,7 +367,7 @@ update_lb (ip6_sr_policy_t * sr_policy)
 
       /* Add FIB entry for BSID */
       fhc = fib_table_get_flow_hash_config (sr_policy->fib_table,
-					    dpo_proto_to_fib (DPO_PROTO_IP6));
+					    FIB_PROTOCOL_IP6);
 
       dpo_set (&sr_policy->bsid_dpo, DPO_LOAD_BALANCE, DPO_PROTO_IP6,
 	       load_balance_create (0, DPO_PROTO_IP6, fhc));
diff --git a/src/vnet/vnet_all_api_h.h b/src/vnet/vnet_all_api_h.h
index c71af079e11..e5f580496b1 100644
--- a/src/vnet/vnet_all_api_h.h
+++ b/src/vnet/vnet_all_api_h.h
@@ -62,6 +62,7 @@
 #include <vnet/tcp/tcp.api.h>
 #include <vnet/dns/dns.api.h>
 #include <vnet/udp/udp.api.h>
+#include <vnet/bier/bier.api.h>
 
 /*
  * fd.io coding-style-patch-verification: ON
diff --git a/src/vpp/api/api.c b/src/vpp/api/api.c
index 36953aec623..c0c6f61f95b 100644
--- a/src/vpp/api/api.c
+++ b/src/vpp/api/api.c
@@ -87,6 +87,7 @@
 #include <vnet/dpo/lookup_dpo.h>
 #include <vnet/dpo/classify_dpo.h>
 #include <vnet/dpo/ip_null_dpo.h>
+
 #define vl_typedefs		/* define message structures */
 #include <vpp/api/vpe_all_api_h.h>
 #undef vl_typedefs
diff --git a/src/vpp/api/vpe.api b/src/vpp/api/vpe.api
index f0f6f41f7eb..0807e4b3985 100644
--- a/src/vpp/api/vpe.api
+++ b/src/vpp/api/vpe.api
@@ -51,6 +51,7 @@ vl_api_version 1.0.0
  * COP APIs: see ... /src/vnet/cop/{cop.api, cop_api.c}
  * POLICER APIs: see ... /src/vnet/policer/{policer.api, policer_api.c}
  * STATS APIs: see .../src/vpp/stats/{stats.api, stats.c}
+ * BIER APIs: see ... /src/vnet/policer/{bier.api, bier_api.c}
  */
 
 /** \brief Create a new subinterface with the given vlan id
diff --git a/test/patches/scapy-2.3.3/bier.patch b/test/patches/scapy-2.3.3/bier.patch
new file mode 100644
index 00000000000..024805d0501
--- /dev/null
+++ b/test/patches/scapy-2.3.3/bier.patch
@@ -0,0 +1,45 @@
+diff --git a/scapy/contrib/bier.py b/scapy/contrib/bier.py
+new file mode 100644
+index 0000000..e173cdb
+--- /dev/null
++++ b/scapy/contrib/bier.py
+@@ -0,0 +1,39 @@
++# http://trac.secdev.org/scapy/ticket/31
++
++# scapy.contrib.description = MPLS
++# scapy.contrib.status = loads
++
++from scapy.packet import *
++from scapy.fields import *
++from scapy.layers.inet import IP
++from scapy.layers.inet6 import IPv6
++
++class BIERLength:
++   BIER_LEN_64 = 0
++   BIER_LEN_128 = 1
++   BIER_LEN_256 = 2
++
++
++
++BIERnhcls = {  1: "MPLS",
++               2: "MPLS",
++               4: "IPv4",
++               5: "IPv6" }
++
++class BIER(Packet):
++   name = "BIER"
++   fields_desc =  [ BitField("id", 5, 4),
++                    BitField("version", 0, 4),
++                    BitField("length", 0, 4),
++                    BitField("entropy", 0, 20),
++                    BitField("OAM", 0, 2),
++                    BitField("RSV", 0, 2),
++                    BitField("DSCP", 0, 6),
++                    BitEnumField("Proto", 2, 6, BIERnhcls),
++                    ShortField("BFRID", 0),
++                    StrFixedLenField("BitString",
++                                     chr(255)*32, 32) ]
++
++
++bind_layers(BIER, IP,   Proto=4)
++bind_layers(BIER, IPv6, Proto=5)
diff --git a/test/patches/scapy-2.3.3/mpls.py.patch b/test/patches/scapy-2.3.3/mpls.py.patch
index f63a70a3cd7..f3293e8eee2 100644
--- a/test/patches/scapy-2.3.3/mpls.py.patch
+++ b/test/patches/scapy-2.3.3/mpls.py.patch
@@ -1,16 +1,35 @@
 diff --git a/scapy/contrib/mpls.py b/scapy/contrib/mpls.py
-index 640a0c5..6af1d4a 100644
+index 640a0c5..944723a 100644
 --- a/scapy/contrib/mpls.py
 +++ b/scapy/contrib/mpls.py
-@@ -18,6 +18,8 @@ class MPLS(Packet):
+@@ -6,6 +6,7 @@
+ from scapy.packet import Packet, bind_layers, Padding
+ from scapy.fields import BitField,ByteField
+ from scapy.layers.inet import IP
++from scapy.contrib.bier import BIER
+ from scapy.layers.inet6 import IPv6
+ from scapy.layers.l2 import Ether, GRE
+ 
+@@ -17,9 +18,12 @@ class MPLS(Packet):
  
     def guess_payload_class(self, payload):
-        if len(payload) >= 1:
-+           if not self.s:
-+              return MPLS
-            ip_version = (ord(payload[0]) >> 4) & 0xF
-            if ip_version == 4:
-                return IP
+-       if len(payload) >= 1:
+-           ip_version = (ord(payload[0]) >> 4) & 0xF
+-           if ip_version == 4:
+-               return IP
+-           elif ip_version == 6:
+-               return IPv6
+-       return Padding
++         if not self.s:
++             return MPLS
++         ip_version = (ord(payload[0]) >> 4) & 0xF
++         if ip_version == 4:
++            return IP
++         elif ip_version == 5:
++            return BIER
++         elif ip_version == 6:
++            return IPv6
++         return Padding
 @@ -27,3 +29,4 @@ class MPLS(Packet):
  
  bind_layers(Ether, MPLS, type=0x8847)
diff --git a/test/test_bier.py b/test/test_bier.py
new file mode 100644
index 00000000000..1a4567bd656
--- /dev/null
+++ b/test/test_bier.py
@@ -0,0 +1,390 @@
+#!/usr/bin/env python
+
+import unittest
+import socket
+
+from framework import VppTestCase, VppTestRunner
+from vpp_ip_route import VppIpRoute, VppRoutePath, VppMplsRoute, \
+    VppMplsTable, VppIpMRoute, VppMRoutePath, VppIpTable, \
+    MRouteEntryFlags, MRouteItfFlags, MPLS_LABEL_INVALID, DpoProto
+from vpp_bier import *
+
+from scapy.packet import Raw
+from scapy.layers.l2 import Ether
+from scapy.layers.inet import IP, UDP, ICMP
+from scapy.layers.inet6 import IPv6
+from scapy.contrib.mpls import MPLS
+from scapy.contrib.bier import *
+
+
+class TestBFIB(VppTestCase):
+    """ BIER FIB Test Case """
+
+    def test_bfib(self):
+        """ BFIB Unit Tests """
+        error = self.vapi.cli("test bier")
+
+        if error:
+            self.logger.critical(error)
+        self.assertEqual(error.find("Failed"), -1)
+
+
+class TestBier(VppTestCase):
+    """ BIER Test Case """
+
+    def setUp(self):
+        super(TestBier, self).setUp()
+
+        # create 2 pg interfaces
+        self.create_pg_interfaces(range(3))
+
+        # create the default MPLS table
+        self.tables = []
+        tbl = VppMplsTable(self, 0)
+        tbl.add_vpp_config()
+        self.tables.append(tbl)
+
+        tbl = VppIpTable(self, 10)
+        tbl.add_vpp_config()
+        self.tables.append(tbl)
+
+        # setup both interfaces
+        for i in self.pg_interfaces:
+            if i == self.pg2:
+                i.set_table_ip4(10)
+            i.admin_up()
+            i.config_ip4()
+            i.resolve_arp()
+            i.enable_mpls()
+
+    def tearDown(self):
+        for i in self.pg_interfaces:
+            i.disable_mpls()
+            i.unconfig_ip4()
+            i.set_table_ip4(0)
+            i.admin_down()
+        super(TestBier, self).tearDown()
+
+    def send_and_assert_no_replies(self, intf, pkts, remark):
+        intf.add_stream(pkts)
+        self.pg_enable_capture(self.pg_interfaces)
+        self.pg_start()
+        for i in self.pg_interfaces:
+            i.assert_nothing_captured(remark=remark)
+
+    def send_and_expect(self, input, pkts, output):
+        self.vapi.cli("trace add bier-mpls-lookup 10")
+        input.add_stream(pkts)
+        self.pg_enable_capture(self.pg_interfaces)
+        self.pg_start()
+        rx = output.get_capture(len(pkts))
+
+    def test_bier_midpoint(self):
+        """BIER midpoint"""
+
+        #
+        # Add a BIER table for sub-domain 0, set 0, and BSL 256
+        #
+        bti = VppBierTableID(0, 0, BIERLength.BIER_LEN_256)
+        bt = VppBierTable(self, bti, 77)
+        bt.add_vpp_config()
+
+        #
+        # A packet with no bits set gets dropped
+        #
+        p = (Ether(dst=self.pg0.local_mac, src=self.pg0.remote_mac) /
+             MPLS(label=77, ttl=255) /
+             BIER(length=BIERLength.BIER_LEN_256,
+                  BitString=chr(0)*64) /
+             IPv6(src=self.pg0.remote_ip6, dst=self.pg0.remote_ip6) /
+             UDP(sport=1234, dport=1234) /
+             Raw())
+        pkts = [p]
+
+        self.send_and_assert_no_replies(self.pg0, pkts,
+                                        "Empty Bit-String")
+
+        #
+        # Add a BIER route for each bit-position in the table via a different
+        # next-hop. Testing whether the BIER walk and replicate forwarding
+        # function works for all bit posisitons.
+        #
+        nh_routes = []
+        bier_routes = []
+        for i in range(1, 256):
+            nh = "10.0.%d.%d" % (i / 255, i % 255)
+            nh_routes.append(VppIpRoute(self, nh, 32,
+                                        [VppRoutePath(self.pg1.remote_ip4,
+                                                      self.pg1.sw_if_index,
+                                                      labels=[2000+i])]))
+            nh_routes[-1].add_vpp_config()
+
+            bier_routes.append(VppBierRoute(self, bti, i, nh, 100+i))
+            bier_routes[-1].add_vpp_config()
+
+        #
+        # A packet with all bits set gets spat out to BP:1
+        #
+        p = (Ether(dst=self.pg0.local_mac, src=self.pg0.remote_mac) /
+             MPLS(label=77, ttl=255) /
+             BIER(length=BIERLength.BIER_LEN_256) /
+             IPv6(src=self.pg0.remote_ip6, dst=self.pg0.remote_ip6) /
+             UDP(sport=1234, dport=1234) /
+             Raw())
+        pkts = [p]
+
+        self.pg0.add_stream(pkts)
+        self.pg_enable_capture(self.pg_interfaces)
+        self.pg_start()
+
+        rx = self.pg1.get_capture(255)
+
+        for rxp in rx:
+            #
+            # The packets are not required to be sent in bit-position order
+            # when we setup the routes above we used the bit-position to
+            # construct the out-label. so use that here to determine the BP
+            #
+            olabel = rxp[MPLS]
+            bp = olabel.label - 2000
+
+            blabel = olabel[MPLS].payload
+            self.assertEqual(blabel.label, 100+bp)
+
+            bier_hdr = blabel[MPLS].payload
+
+            self.assertEqual(bier_hdr.id, 5)
+            self.assertEqual(bier_hdr.version, 0)
+            self.assertEqual(bier_hdr.length, BIERLength.BIER_LEN_256)
+            self.assertEqual(bier_hdr.entropy, 0)
+            self.assertEqual(bier_hdr.OAM, 0)
+            self.assertEqual(bier_hdr.RSV, 0)
+            self.assertEqual(bier_hdr.DSCP, 0)
+            self.assertEqual(bier_hdr.Proto, 5)
+
+            # The bit-string should consist only of the BP given by i.
+            i = 0
+            bitstring = ""
+            bpi = bp - 1
+            while (i < bpi/8):
+                bitstring = chr(0) + bitstring
+                i += 1
+            bitstring = chr(1 << bpi % 8) + bitstring
+
+            while len(bitstring) < 32:
+                bitstring = chr(0) + bitstring
+
+            self.assertEqual(len(bitstring), len(bier_hdr.BitString))
+            self.assertEqual(bitstring, bier_hdr.BitString)
+
+    def test_bier_head(self):
+        """BIER head"""
+
+        #
+        # Add a BIER table for sub-domain 0, set 0, and BSL 256
+        #
+        bti = VppBierTableID(0, 0, BIERLength.BIER_LEN_256)
+        bt = VppBierTable(self, bti, 77)
+        bt.add_vpp_config()
+
+        #
+        # 2 bit positions via two next hops
+        #
+        nh1 = "10.0.0.1"
+        nh2 = "10.0.0.2"
+        ip_route_1 = VppIpRoute(self, nh1, 32,
+                                [VppRoutePath(self.pg1.remote_ip4,
+                                              self.pg1.sw_if_index,
+                                              labels=[2001])])
+        ip_route_2 = VppIpRoute(self, nh2, 32,
+                                [VppRoutePath(self.pg1.remote_ip4,
+                                              self.pg1.sw_if_index,
+                                              labels=[2002])])
+        ip_route_1.add_vpp_config()
+        ip_route_2.add_vpp_config()
+
+        bier_route_1 = VppBierRoute(self, bti, 1, nh1, 101)
+        bier_route_2 = VppBierRoute(self, bti, 2, nh2, 102)
+        bier_route_1.add_vpp_config()
+        bier_route_2.add_vpp_config()
+
+        #
+        # An imposition object with both bit-positions set
+        #
+        bi = VppBierImp(self, bti, 333, chr(0x3) * 32)
+        bi.add_vpp_config()
+
+        #
+        # Add a multicast route that will forward into the BIER doamin
+        #
+        route_ing_232_1_1_1 = VppIpMRoute(
+            self,
+            "0.0.0.0",
+            "232.1.1.1", 32,
+            MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
+            paths=[VppMRoutePath(self.pg0.sw_if_index,
+                                 MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT),
+                   VppMRoutePath(0xffffffff,
+                                 MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                                 proto=DpoProto.DPO_PROTO_BIER,
+                                 bier_imp=bi.bi_index)])
+        route_ing_232_1_1_1.add_vpp_config()
+
+        #
+        # inject a packet an IP. We expect it to be BIER encapped,
+        # replicated.
+        #
+        p = (Ether(dst=self.pg0.local_mac,
+                   src=self.pg0.remote_mac) /
+             IP(src="1.1.1.1", dst="232.1.1.1") /
+             UDP(sport=1234, dport=1234))
+
+        self.pg0.add_stream([p])
+        self.pg_enable_capture(self.pg_interfaces)
+        self.pg_start()
+
+        rx = self.pg1.get_capture(2)
+
+    def test_bier_tail(self):
+        """BIER Tail"""
+
+        #
+        # Add a BIER table for sub-domain 0, set 0, and BSL 256
+        #
+        bti = VppBierTableID(0, 0, BIERLength.BIER_LEN_256)
+        bt = VppBierTable(self, bti, 77)
+        bt.add_vpp_config()
+
+        #
+        # disposition table
+        #
+        bdt = VppBierDispTable(self, 8)
+        bdt.add_vpp_config()
+
+        #
+        # BIER route in table that's for-us
+        #
+        bier_route_1 = VppBierRoute(self, bti, 1, "0.0.0.0", 0,
+                                    disp_table=8)
+        bier_route_1.add_vpp_config()
+
+        #
+        # An entry in the disposition table
+        #
+        bier_de_1 = VppBierDispEntry(self, bdt.id, 99,
+                                     BIER_HDR_PAYLOAD.BIER_HDR_PROTO_IPV4,
+                                     "0.0.0.0", 0, rpf_id=8192)
+        bier_de_1.add_vpp_config()
+
+        #
+        # A multicast route to forward post BIER disposition
+        #
+        route_eg_232_1_1_1 = VppIpMRoute(
+            self,
+            "0.0.0.0",
+            "232.1.1.1", 32,
+            MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
+            paths=[VppMRoutePath(self.pg1.sw_if_index,
+                                 MRouteItfFlags.MFIB_ITF_FLAG_FORWARD)])
+        route_eg_232_1_1_1.add_vpp_config()
+        route_eg_232_1_1_1.update_rpf_id(8192)
+
+        #
+        # A packet with all bits set gets spat out to BP:1
+        #
+        p = (Ether(dst=self.pg0.local_mac, src=self.pg0.remote_mac) /
+             MPLS(label=77, ttl=255) /
+             BIER(length=BIERLength.BIER_LEN_256, BFRID=99) /
+             IP(src="1.1.1.1", dst="232.1.1.1") /
+             UDP(sport=1234, dport=1234) /
+             Raw())
+
+        self.send_and_expect(self.pg0, [p], self.pg1)
+
+    def test_bier_e2e(self):
+        """ BIER end-to-end """
+
+        #
+        # Add a BIER table for sub-domain 0, set 0, and BSL 256
+        #
+        bti = VppBierTableID(0, 0, BIERLength.BIER_LEN_256)
+        bt = VppBierTable(self, bti, 77)
+        bt.add_vpp_config()
+
+        #
+        # Impostion Sets bit string 101010101....
+        #  sender 333
+        #
+        bi = VppBierImp(self, bti, 333, chr(0x5) * 32)
+        bi.add_vpp_config()
+
+        #
+        # Add a multicast route that will forward into the BIER doamin
+        #
+        route_ing_232_1_1_1 = VppIpMRoute(
+            self,
+            "0.0.0.0",
+            "232.1.1.1", 32,
+            MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
+            paths=[VppMRoutePath(self.pg0.sw_if_index,
+                                 MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT),
+                   VppMRoutePath(0xffffffff,
+                                 MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                                 proto=DpoProto.DPO_PROTO_BIER,
+                                 bier_imp=bi.bi_index)])
+        route_ing_232_1_1_1.add_vpp_config()
+
+        #
+        # disposition table 8
+        #
+        bdt = VppBierDispTable(self, 8)
+        bdt.add_vpp_config()
+
+        #
+        # BIER route in table that's for-us, resolving through
+        # disp table 8.
+        #
+        bier_route_1 = VppBierRoute(self, bti, 1, "0.0.0.0",
+                                    MPLS_LABEL_INVALID,
+                                    disp_table=8)
+        bier_route_1.add_vpp_config()
+
+        #
+        # An entry in the disposition table for sender 333
+        #  lookup in VRF 10
+        #
+        bier_de_1 = VppBierDispEntry(self, bdt.id, 333,
+                                     BIER_HDR_PAYLOAD.BIER_HDR_PROTO_IPV4,
+                                     "0.0.0.0", 10, rpf_id=8192)
+        bier_de_1.add_vpp_config()
+
+        #
+        # Add a multicast route that will forward the traffic
+        # post-disposition
+        #
+        route_eg_232_1_1_1 = VppIpMRoute(
+            self,
+            "0.0.0.0",
+            "232.1.1.1", 32,
+            MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
+            table_id=10,
+            paths=[VppMRoutePath(self.pg1.sw_if_index,
+                                 MRouteItfFlags.MFIB_ITF_FLAG_FORWARD)])
+        route_eg_232_1_1_1.add_vpp_config()
+        route_eg_232_1_1_1.update_rpf_id(8192)
+
+        #
+        # inject a packet in VRF-0. We expect it to be BIER encapped,
+        # replicated, then hit the disposition and be forwarded
+        # out of VRF 10, i.e. on pg1
+        #
+        p = (Ether(dst=self.pg0.local_mac,
+                   src=self.pg0.remote_mac) /
+             IP(src="1.1.1.1", dst="232.1.1.1") /
+             UDP(sport=1234, dport=1234))
+
+        self.send_and_expect(self.pg0, p*65, self.pg1)
+
+
+if __name__ == '__main__':
+    unittest.main(testRunner=VppTestRunner)
diff --git a/test/test_ip_mcast.py b/test/test_ip_mcast.py
index 7cad683cac5..8ca92df26b2 100644
--- a/test/test_ip_mcast.py
+++ b/test/test_ip_mcast.py
@@ -5,7 +5,7 @@ import unittest
 from framework import VppTestCase, VppTestRunner
 from vpp_sub_interface import VppSubInterface, VppDot1QSubint, VppDot1ADSubint
 from vpp_ip_route import VppIpMRoute, VppMRoutePath, VppMFibSignal, \
-    MRouteItfFlags, MRouteEntryFlags, VppIpTable
+    MRouteItfFlags, MRouteEntryFlags, VppIpTable, DpoProto
 
 from scapy.packet import Raw
 from scapy.layers.l2 import Ether
@@ -324,10 +324,6 @@ class TestIPMcast(VppTestCase):
         self.verify_capture_ip4(self.pg6, tx)
         self.verify_capture_ip4(self.pg7, tx)
 
-        route_232_1_1_1.remove_vpp_config()
-        route_1_1_1_1_232_1_1_1.remove_vpp_config()
-        route_232.remove_vpp_config()
-
     def test_ip6_mcast(self):
         """ IPv6 Multicast Replication """
 
@@ -354,13 +350,17 @@ class TestIPMcast(VppTestCase):
             "ff01::1", 128,
             MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
             [VppMRoutePath(self.pg0.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT),
+                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg1.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD),
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg2.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD),
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg3.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD)],
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6)],
             is_ip6=1)
         route_ff01_1.add_vpp_config()
 
@@ -374,11 +374,14 @@ class TestIPMcast(VppTestCase):
             "ff01::1", 256,
             MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
             [VppMRoutePath(self.pg0.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT),
+                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg1.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD),
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg2.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD)],
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6)],
             is_ip6=1)
         route_2001_ff01_1.add_vpp_config()
 
@@ -392,9 +395,11 @@ class TestIPMcast(VppTestCase):
             "ff01::", 16,
             MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
             [VppMRoutePath(self.pg0.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT),
+                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg1.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD)],
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6)],
             is_ip6=1)
         route_ff01.add_vpp_config()
 
@@ -474,10 +479,6 @@ class TestIPMcast(VppTestCase):
         self.pg3.assert_nothing_captured(
             remark="IP multicast packets forwarded on PG3")
 
-        route_ff01.remove_vpp_config()
-        route_ff01_1.remove_vpp_config()
-        route_2001_ff01_1.remove_vpp_config()
-
     def _mcast_connected_send_stream(self, dst_ip):
         self.vapi.cli("clear trace")
         tx = self.create_stream_ip4(self.pg0,
@@ -584,8 +585,10 @@ class TestIPMcast(VppTestCase):
         signal_232_1_1_1_itf_0.compare(signals[1])
         signal_232_1_1_2_itf_0.compare(signals[0])
 
-        route_232_1_1_1.remove_vpp_config()
-        route_232_1_1_2.remove_vpp_config()
+        route_232_1_1_1.update_entry_flags(
+            MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE)
+        route_232_1_1_2.update_entry_flags(
+            MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE)
 
     def test_ip_mcast_signal(self):
         """ IP Multicast Signal """
@@ -679,11 +682,6 @@ class TestIPMcast(VppTestCase):
         signals = self.vapi.mfib_signal_dump()
         self.assertEqual(0, len(signals))
 
-        #
-        # Cleanup
-        #
-        route_232_1_1_1.remove_vpp_config()
-
     def test_ip_mcast_vrf(self):
         """ IP Multicast Replication in non-default table"""
 
@@ -733,11 +731,14 @@ class TestIPMcast(VppTestCase):
             "ff01::1", 256,
             MRouteEntryFlags.MFIB_ENTRY_FLAG_NONE,
             [VppMRoutePath(self.pg8.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT),
+                           MRouteItfFlags.MFIB_ITF_FLAG_ACCEPT,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg1.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD),
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6),
              VppMRoutePath(self.pg2.sw_if_index,
-                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD)],
+                           MRouteItfFlags.MFIB_ITF_FLAG_FORWARD,
+                           proto=DpoProto.DPO_PROTO_IP6)],
             table_id=10,
             is_ip6=1)
         route_2001_ff01_1.add_vpp_config()
diff --git a/test/vpp_bier.py b/test/vpp_bier.py
new file mode 100644
index 00000000000..58c4f7248da
--- /dev/null
+++ b/test/vpp_bier.py
@@ -0,0 +1,267 @@
+"""
+  BIER Tables and Routes
+"""
+
+import socket
+from vpp_object import VppObject
+
+
+class BIER_HDR_PAYLOAD:
+    BIER_HDR_PROTO_MPLS_DOWN_STREAM = 1
+    BIER_HDR_PROTO_MPLS_UP_STREAM = 2
+    BIER_HDR_PROTO_ETHERNET = 3
+    BIER_HDR_PROTO_IPV4 = 4
+    BIER_HDR_PROTO_IPV6 = 5
+    BIER_HDR_PROTO_VXLAN = 6
+    BIER_HDR_PROTO_CTRL = 7
+    BIER_HDR_PROTO_OAM = 8
+
+
+class VppBierTableID():
+    def __init__(self, set_id, sub_domain_id, hdr_len_id):
+        self.set_id = set_id
+        self.sub_domain_id = sub_domain_id
+        self.hdr_len_id = hdr_len_id
+
+
+def find_bier_table(test, bti):
+    tables = test.vapi.bier_table_dump()
+    for t in tables:
+        if bti.set_id == t.bt_tbl_id.bt_set \
+           and bti.sub_domain_id == t.bt_tbl_id.bt_sub_domain \
+           and bti.hdr_len_id == t.bt_tbl_id.bt_hdr_len_id:
+            return True
+    return False
+
+
+def find_bier_route(test, bti, bp):
+    routes = test.vapi.bier_route_dump(bti)
+    for r in routes:
+        if bti.set_id == r.br_tbl_id.bt_set \
+           and bti.sub_domain_id == r.br_tbl_id.bt_sub_domain \
+           and bti.hdr_len_id == r.br_tbl_id.bt_hdr_len_id \
+           and bp == r.br_bp:
+            return True
+    return False
+
+
+def find_bier_disp_table(test, bdti):
+    tables = test.vapi.bier_disp_table_dump()
+    for t in tables:
+        if bdti == t.bdt_tbl_id:
+            return True
+    return False
+
+
+def find_bier_disp_entry(test, bdti, bp):
+    entries = test.vapi.bier_disp_entry_dump(bdti)
+    for e in entries:
+        if bp == e.bde_bp \
+           and bdti == e.bde_tbl_id:
+            return True
+    return False
+
+
+def find_bier_imp(test, bti, bp):
+    imps = test.vapi.bier_imp_dump()
+    for i in imps:
+        if bti.set_id == i.bi_tbl_id.bt_set \
+           and bti.sub_domain_id == i.bi_tbl_id.bt_sub_domain \
+           and bti.hdr_len_id == i.bi_tbl_id.bt_hdr_len_id \
+           and bp == i.bi_src:
+            return True
+    return False
+
+
+class VppBierTable(VppObject):
+    """
+    BIER Table
+    """
+
+    def __init__(self, test, id, mpls_label):
+        self._test = test
+        self.id = id
+        self.mpls_label = mpls_label
+
+    def add_vpp_config(self):
+        self._test.vapi.bier_table_add_del(
+            self.id,
+            self.mpls_label,
+            is_add=1)
+        self._test.registry.register(self, self._test.logger)
+
+    def remove_vpp_config(self):
+        self._test.vapi.bier_table_add_del(
+            self.id,
+            self.mpls_label,
+            is_add=0)
+
+    def __str__(self):
+        return self.object_id()
+
+    def object_id(self):
+        return "bier-table;[%d:%d:%d]" % (self.id.set_id,
+                                          self.id.sub_domain_id,
+                                          self.id.hdr_len_id)
+
+    def query_vpp_config(self):
+        return find_bier_table(self._test, self.id)
+
+
+class VppBierRoute(VppObject):
+    """
+    BIER route
+    """
+
+    def __init__(self, test, tbl_id, bp, nh, out_label,
+                 disp_table=0):
+        self._test = test
+        self.tbl_id = tbl_id
+        self.out_label = out_label
+        self.bp = bp
+        self.disp_table = disp_table
+        self.nh = socket.inet_pton(socket.AF_INET, nh)
+
+    def add_vpp_config(self):
+        self._test.vapi.bier_route_add_del(
+            self.tbl_id,
+            self.bp,
+            self.nh,
+            self.out_label,
+            self.disp_table,
+            is_add=1)
+        self._test.registry.register(self, self._test.logger)
+
+    def remove_vpp_config(self):
+        self._test.vapi.bier_route_add_del(
+            self.tbl_id,
+            self.bp,
+            self.nh,
+            self.out_label,
+            self.disp_table,
+            is_add=0)
+
+    def __str__(self):
+        return self.object_id()
+
+    def object_id(self):
+        return "bier-route;[%d:%d:%d:%d]" % (self.tbl_id.set_id,
+                                             self.tbl_id.sub_domain_id,
+                                             self.tbl_id.hdr_len_id,
+                                             self.bp)
+
+    def query_vpp_config(self):
+        return find_bier_route(self._test, self.tbl_id, self.bp)
+
+
+class VppBierImp(VppObject):
+    """
+    BIER route
+    """
+
+    def __init__(self, test, tbl_id, src, ibytes):
+        self._test = test
+        self.tbl_id = tbl_id
+        self.ibytes = ibytes
+        self.src = src
+
+    def add_vpp_config(self):
+        res = self._test.vapi.bier_imp_add(
+            self.tbl_id,
+            self.src,
+            self.ibytes)
+        self.bi_index = res.bi_index
+        self._test.registry.register(self, self._test.logger)
+
+    def remove_vpp_config(self):
+        self._test.vapi.bier_imp_del(
+            self.bi_index)
+
+    def __str__(self):
+        return self.object_id()
+
+    def object_id(self):
+        return "bier-imp;[%d:%d:%d:%d]" % (self.tbl_id.set_id,
+                                           self.tbl_id.sub_domain_id,
+                                           self.tbl_id.hdr_len_id,
+                                           self.src)
+
+    def query_vpp_config(self):
+        return find_bier_imp(self._test, self.tbl_id, self.src)
+
+
+class VppBierDispTable(VppObject):
+    """
+    BIER Disposition Table
+    """
+
+    def __init__(self, test, id):
+        self._test = test
+        self.id = id
+
+    def add_vpp_config(self):
+        self._test.vapi.bier_disp_table_add_del(
+            self.id,
+            is_add=1)
+        self._test.registry.register(self, self._test.logger)
+
+    def remove_vpp_config(self):
+        self._test.vapi.bier_disp_table_add_del(
+            self.id,
+            is_add=0)
+
+    def __str__(self):
+        return self.object_id()
+
+    def object_id(self):
+        return "bier-disp-table;[%d]" % (self.id)
+
+    def query_vpp_config(self):
+        return find_bier_disp_table(self._test, self.id)
+
+
+class VppBierDispEntry(VppObject):
+    """
+    BIER Disposition Entry
+    """
+
+    def __init__(self, test, tbl_id, bp, payload_proto, nh, nh_tbl,
+                 rpf_id=~0):
+        self._test = test
+        self.tbl_id = tbl_id
+        self.nh_tbl = nh_tbl
+        self.bp = bp
+        self.payload_proto = payload_proto
+        self.rpf_id = rpf_id
+        self.nh = socket.inet_pton(socket.AF_INET, nh)
+
+    def add_vpp_config(self):
+        self._test.vapi.bier_disp_entry_add_del(
+            self.tbl_id,
+            self.bp,
+            self.payload_proto,
+            self.nh,
+            self.nh_tbl,
+            self.rpf_id,
+            is_add=1)
+        self._test.registry.register(self, self._test.logger)
+
+    def remove_vpp_config(self):
+        self._test.vapi.bier_disp_entry_add_del(
+            self.tbl_id,
+            self.bp,
+            self.payload_proto,
+            self.nh,
+            self.nh_tbl,
+            self.rpf_id,
+            is_add=0)
+
+    def __str__(self):
+        return self.object_id()
+
+    def object_id(self):
+        return "bier-disp-entry;[%d:%d]" % (self.tbl_id,
+                                            self.bp)
+
+    def query_vpp_config(self):
+        return find_bier_disp_entry(self._test, self.tbl_id, self.bp)
diff --git a/test/vpp_ip_route.py b/test/vpp_ip_route.py
index 34aa5428d18..3bc3b65111c 100644
--- a/test/vpp_ip_route.py
+++ b/test/vpp_ip_route.py
@@ -34,7 +34,8 @@ class DpoProto:
     DPO_PROTO_IP6 = 1
     DPO_PROTO_MPLS = 2
     DPO_PROTO_ETHERNET = 3
-    DPO_PROTO_NSH = 4
+    DPO_PROTO_BIER = 4
+    DPO_PROTO_NSH = 5
 
 
 def find_route(test, ip_addr, len, table_id=0, inet=AF_INET):
@@ -138,10 +139,15 @@ class VppRoutePath(object):
 
 class VppMRoutePath(VppRoutePath):
 
-    def __init__(self, nh_sw_if_index, flags):
-        super(VppMRoutePath, self).__init__("0.0.0.0",
-                                            nh_sw_if_index)
+    def __init__(self, nh_sw_if_index, flags,
+                 proto=DpoProto.DPO_PROTO_IP4,
+                 bier_imp=0):
+        super(VppMRoutePath, self).__init__(
+            "::" if proto is DpoProto.DPO_PROTO_IP6 else "0.0.0.0",
+            nh_sw_if_index,
+            proto=proto)
         self.nh_i_flags = flags
+        self.bier_imp = bier_imp
 
 
 class VppIpRoute(VppObject):
@@ -283,8 +289,10 @@ class VppIpMRoute(VppObject):
                                               self.grp_addr,
                                               self.grp_addr_len,
                                               self.e_flags,
+                                              path.proto,
                                               path.nh_itf,
                                               path.nh_i_flags,
+                                              bier_imp=path.bier_imp,
                                               rpf_id=self.rpf_id,
                                               table_id=self.table_id,
                                               is_ipv6=self.is_ip6)
@@ -296,6 +304,7 @@ class VppIpMRoute(VppObject):
                                               self.grp_addr,
                                               self.grp_addr_len,
                                               self.e_flags,
+                                              path.proto,
                                               path.nh_itf,
                                               path.nh_i_flags,
                                               table_id=self.table_id,
@@ -308,6 +317,7 @@ class VppIpMRoute(VppObject):
                                           self.grp_addr,
                                           self.grp_addr_len,
                                           self.e_flags,
+                                          0,
                                           0xffffffff,
                                           0,
                                           table_id=self.table_id,
@@ -319,6 +329,7 @@ class VppIpMRoute(VppObject):
                                           self.grp_addr,
                                           self.grp_addr_len,
                                           self.e_flags,
+                                          0,
                                           0xffffffff,
                                           0,
                                           rpf_id=self.rpf_id,
@@ -334,16 +345,21 @@ class VppIpMRoute(VppObject):
                                           self.grp_addr,
                                           self.grp_addr_len,
                                           self.e_flags,
+                                          path.proto,
                                           path.nh_itf,
                                           path.nh_i_flags,
                                           table_id=self.table_id,
                                           is_ipv6=self.is_ip6)
 
     def query_vpp_config(self):
-        dump = self._test.vapi.ip_fib_dump()
+        if self.is_ip6:
+            dump = self._test.vapi.ip6_mfib_dump()
+        else:
+            dump = self._test.vapi.ip_mfib_dump()
         for e in dump:
-            if self.grp_addr == e.address \
+            if self.grp_addr == e.grp_address \
                and self.grp_addr_len == e.address_length \
+               and self.src_addr == e.src_address \
                and self.table_id == e.table_id:
                 return True
         return False
diff --git a/test/vpp_papi_provider.py b/test/vpp_papi_provider.py
index 5085ff2801e..893ef0d89b9 100644
--- a/test/vpp_papi_provider.py
+++ b/test/vpp_papi_provider.py
@@ -2057,14 +2057,17 @@ class VppPapiProvider(object):
                           grp_address,
                           grp_address_length,
                           e_flags,
+                          next_hop_afi,
                           next_hop_sw_if_index,
                           i_flags,
+                          bier_imp=0,
                           rpf_id=0,
                           table_id=0,
                           is_add=1,
                           is_ipv6=0,
                           is_local=0):
         """
+        IP Multicast Route add/del
         """
         return self.api(
             self.papi.ip_mroute_add_del,
@@ -2076,6 +2079,8 @@ class VppPapiProvider(object):
              'is_add': is_add,
              'is_ipv6': is_ipv6,
              'is_local': is_local,
+             'bier_imp': bier_imp,
+             'next_hop_afi': next_hop_afi,
              'grp_address_length': grp_address_length,
              'grp_address': grp_address,
              'src_address': src_address})
@@ -2086,6 +2091,9 @@ class VppPapiProvider(object):
     def ip_mfib_dump(self):
         return self.api(self.papi.ip_mfib_dump, {})
 
+    def ip6_mfib_dump(self):
+        return self.api(self.papi.ip6_mfib_dump, {})
+
     def lisp_enable_disable(self, is_enabled):
         return self.api(
             self.papi.lisp_enable_disable,
@@ -2634,3 +2642,114 @@ class VppPapiProvider(object):
                          'nh': nh,
                          'is_add': is_add,
                          'is_ip6': is_ip6})
+
+    def bier_table_add_del(self,
+                           bti,
+                           mpls_label,
+                           is_add=1):
+        """ BIER Table add/del """
+        return self.api(
+            self.papi.bier_table_add_del,
+            {'bt_tbl_id': {"bt_set": bti.set_id,
+                           "bt_sub_domain": bti.sub_domain_id,
+                           "bt_hdr_len_id": bti.hdr_len_id},
+             'bt_label': mpls_label,
+             'bt_is_add': is_add})
+
+    def bier_table_dump(self):
+        return self.api(self.papi.bier_table_dump, {})
+
+    def bier_route_add_del(self,
+                           bti,
+                           bp,
+                           next_hop,
+                           next_hop_label,
+                           next_hop_table_id,
+                           next_hop_is_ip4=1,
+                           is_add=1):
+        """ BIER Route add/del """
+        return self.api(
+            self.papi.bier_route_add_del,
+            {'br_tbl_id': {"bt_set": bti.set_id,
+                           "bt_sub_domain": bti.sub_domain_id,
+                           "bt_hdr_len_id": bti.hdr_len_id},
+             'br_bp': bp,
+             'br_n_paths': 1,
+             'br_paths': [{'next_hop': next_hop,
+                           'afi': 0,
+                           'n_labels': 1,
+                           'table_id': next_hop_table_id,
+                           'label_stack': [next_hop_label]}],
+             'br_is_add': is_add})
+
+    def bier_route_dump(self, bti):
+        return self.api(
+            self.papi.bier_route_dump,
+            {'br_tbl_id': {"bt_set": bti.set_id,
+                           "bt_sub_domain": bti.sub_domain_id,
+                           "bt_hdr_len_id": bti.hdr_len_id}})
+
+    def bier_imp_add(self,
+                     bti,
+                     src,
+                     ibytes,
+                     is_add=1):
+        """ BIER Imposition Add """
+        return self.api(
+            self.papi.bier_imp_add,
+            {'bi_tbl_id': {"bt_set": bti.set_id,
+                           "bt_sub_domain": bti.sub_domain_id,
+                           "bt_hdr_len_id": bti.hdr_len_id},
+             'bi_src': src,
+             'bi_n_bytes': len(ibytes),
+             'bi_bytes': ibytes})
+
+    def bier_imp_del(self, bi_index):
+        """ BIER Imposition del """
+        return self.api(
+            self.papi.bier_imp_del,
+            {'bi_index': bi_index})
+
+    def bier_imp_dump(self):
+        return self.api(self.papi.bier_imp_dump, {})
+
+    def bier_disp_table_add_del(self,
+                                bdti,
+                                is_add=1):
+        """ BIER Disposition Table add/del """
+        return self.api(
+            self.papi.bier_disp_table_add_del,
+            {'bdt_tbl_id': bdti,
+             'bdt_is_add': is_add})
+
+    def bier_disp_table_dump(self):
+        return self.api(self.papi.bier_disp_table_dump, {})
+
+    def bier_disp_entry_add_del(self,
+                                bdti,
+                                bp,
+                                payload_proto,
+                                next_hop,
+                                next_hop_tbl_id=0,
+                                next_hop_rpf_id=~0,
+                                next_hop_is_ip4=1,
+                                is_add=1):
+        """ BIER Route add/del """
+        return self.api(
+            self.papi.bier_disp_entry_add_del,
+            {'bde_tbl_id': bdti,
+             'bde_bp': bp,
+             'bde_payload_proto': payload_proto,
+             'bde_n_paths': 1,
+             'bde_paths': [{'next_hop': next_hop,
+                            'table_id': next_hop_tbl_id,
+                            'afi': 0,
+                            'rpf_id': next_hop_rpf_id,
+                            'n_labels': 0,
+                            'label_stack': [0]}],
+             'bde_is_add': is_add})
+
+    def bier_disp_entry_dump(self, bdti):
+        return self.api(
+            self.papi.bier_disp_entry_dump,
+            {'bde_tbl_id': bdti})