summaryrefslogtreecommitdiffstats
path: root/docs/gettingstarted/developers/fib20/mplsfib.rst
diff options
context:
space:
mode:
authorjdenisco <jdenisco@cisco.com>2018-08-29 13:19:43 -0400
committerNeale Ranns <nranns@cisco.com>2018-08-30 15:14:49 +0000
commit0923a2376fcdc07fc298b9422233f4a5c1777e7d (patch)
tree7164f6f7083d48c7cd47abdbea7377d4f323dc26 /docs/gettingstarted/developers/fib20/mplsfib.rst
parent5e2c54d02906a7b4787a1f41b59baf97c3a4a840 (diff)
docs: FIB 2.0 start
Change-Id: I87cd2eae133c9f5b9f7764a0f7a41bcc28523e4c Signed-off-by: jdenisco <jdenisco@cisco.com>
Diffstat (limited to 'docs/gettingstarted/developers/fib20/mplsfib.rst')
-rw-r--r--docs/gettingstarted/developers/fib20/mplsfib.rst90
1 files changed, 90 insertions, 0 deletions
diff --git a/docs/gettingstarted/developers/fib20/mplsfib.rst b/docs/gettingstarted/developers/fib20/mplsfib.rst
new file mode 100644
index 00000000000..5fee7d0cd7f
--- /dev/null
+++ b/docs/gettingstarted/developers/fib20/mplsfib.rst
@@ -0,0 +1,90 @@
+.. _mplsfib:
+
+MPLS FIB
+----------
+
+There is a tight coupling between IP and MPLS forwarding. MPLS forwarding
+equivalence classes (FECs) are often an IP prefix Рthat is to say that traffic
+matching a given IP prefix is routed into a MPLS label switch path (LSP). It is
+thus necessary to be able to associated a given prefix/route with an [out-going]
+MPLS label that will be imposed when the packet is forwarded. This is configured
+as:
+
+.. code-block:: console
+
+ $ ip route add 1.1.1.1/32 via 10.10.10.10 GigabitEthernet0/8/0 out-label 33
+
+packets matching 1.1.1.1/32 will be forwarded out GigabitEthernet0/8/0 and have MPLS label
+33 imposed. More than one out-going label can be specified. Out-going MPLS labels
+can be applied to recursive and non-recursive routes, e.g;
+
+.. code-block:: console
+
+ $ ip route add 2.2.2.0/24 via 1.1.1.1 out-label 34
+
+packets matching 2.2.2.0/24 will thus have two MPLS labels imposed; 34 and 33.
+This is the realisation of, e,g, an MPLS BGP VPNv4.
+
+To associate/allocate a local-label for a prefix, and thus have packets to that
+local-label forwarded equivalently to the prefix do;
+
+.. code-block:: console
+
+ $ mpls local-label 99 2.2.2.0/24
+
+In the API this action is called a *bind*.
+
+The router receiving the MPLS encapsulated packets needs to be programmed with
+actions associated which each label value Рthis is the role of the MPLS FIB.
+The MPLS FIB Is a table, whose key is the MPLS label value and end-of-stack (EOS)
+bit, which stores the action to perform on packets with matching encapsulation.
+
+Currently supported actions are:
+
+#. Pop the label and perform an IPv[46] lookup in a specified table
+#. Pop the label and forward via a specified next-hop (this is penultimate-hop-pop, PHP)
+#. Swap the label and forward via a specified next-hop.
+
+These can be programmed respectively by:
+
+#. mpls local-label 33 ip4-lookup-in-table X
+#. mpls local-label 33 via 10.10.10.10 GigabitEthernet0/8/0
+#. mpls local-label 33 via 10.10.10.10 GigabitEthernet0/8/0 out-label 66
+
+the latter is an example of an MPLS cross connect. Any description of a next-hop,
+recursive, non-recursive, labelled, non-labelled, etc, that is valid for an IP
+prefix, is also valid for an MPLS local-label.
+
+Implementation
+^^^^^^^^^^^^^^^
+
+The MPLS FIB is implemented using exactly the same data structures as the IP FIB.
+The only difference is the implementation of the table. Whereas for IPv4 this is
+an mtrie and for IPv6 a hash table, for MPLS it is a flat array indexed by a 21
+bit key (label & EOS bit). This implementation is chosen to favour packet
+forwarding speed.
+
+MPLS Tunnels
+^^^^^^^^^^^^^
+
+VPP no longer supports MPLS tunnels that are coupled to a particular transport,
+
+i.e. MPLSoGRE or MPLSoEth. Such tight coupling is not beneficial. Instead VPP supports;
+
+#. MPLS LSPs associated with IP prefixes and MPLS local-labels (as described above) which are transport independent (i.e. the IP route could be reachable over a GRE tunnel, or any other interface type).
+#. A generic uni-directional MPLS tunnel interface that is transport independent.
+
+An MPLS tunnel is effectively an LSP with an associated interface. The LSP can be
+described by any next-hop type (recursive, non-recursive etc), e.g.:
+
+mpls tunnel add via 10.10.10.10 GigabitEthernet0/8/0 out-label 66
+IP routes and/or MPLS x-connects can be routed via the interface, e.g.
+
+.. code-block:: console
+
+ $ ip route add 2.2.2.0/24 via mpls-tunnel0
+
+packets matching the route for 2.2.2.0/24 would thus have label 66 imposed since
+it is transmitted via the tunnel.
+
+These MPLS tunnels can be used to realise MPLS RSVP-TE tunnels.