aboutsummaryrefslogtreecommitdiffstats
path: root/docs/report/introduction/methodology_packet_latency.rst
diff options
context:
space:
mode:
Diffstat (limited to 'docs/report/introduction/methodology_packet_latency.rst')
-rw-r--r--docs/report/introduction/methodology_packet_latency.rst23
1 files changed, 23 insertions, 0 deletions
diff --git a/docs/report/introduction/methodology_packet_latency.rst b/docs/report/introduction/methodology_packet_latency.rst
new file mode 100644
index 0000000000..550d12f688
--- /dev/null
+++ b/docs/report/introduction/methodology_packet_latency.rst
@@ -0,0 +1,23 @@
+Packet Latency
+--------------
+
+TRex Traffic Generator (TG) is used for measuring latency of VPP DUTs.
+Reported latency values are measured using following methodology:
+
+- Latency tests are performed at 100% of discovered NDR and PDR rates
+ for each throughput test and packet size (except IMIX).
+- TG sends dedicated latency streams, one per direction, each at the
+ rate of 9 kpps at the prescribed packet size; these are sent in
+ addition to the main load streams.
+- TG reports min/avg/max latency values per stream direction, hence two
+ sets of latency values are reported per test case; future release of
+ TRex is expected to report latency percentiles.
+- Reported latency values are aggregate across two SUTs due to three
+ node topology used for all performance tests; for per SUT latency,
+ reported value should be divided by two.
+- 1usec is the measurement accuracy advertised by TRex TG for the setup
+ used in FD.io labs used by CSIT project.
+- TRex setup introduces an always-on error of about 2*2usec per latency
+ flow additonal Tx/Rx interface latency induced by TRex SW writing and
+ reading packet timestamps on CPU cores without HW acceleration on NICs
+ closer to the interface line.