Release Notes ============= Changes in CSIT |release| ------------------------- #. **VPP performance tests** - *MRR tests* - New Maximum Receive Rate tests measure the packet forwarding rate under the maximum load offered by traffic generator over a set trial duration, regardless of packet loss. MRR tests are used for continuous performance trending and for comparison between releases. - *Service Chaining with SRv6* - New SRv6 (Segment Routing IPv6) proxy tests measure performance of SRv6 Endpoint fronting SR-unaware appliance via masquerading (End.AM), dynamic proxy (End.AD) or static proxy (End.AS) SR functions. #. **Presentation and Analytics Layer** - *Performance trending* - Added continuous performance trending and analysis. New Performance Trending and Performance Analysis jobs executed regular throughput tests, with results being subsequently analysed and trend and anomalies summarized and presented in VPP Performance Dashboard and trendline graphs. #. **Test Framework Optimizations** - *Performance tests efficiency* - Qemu build/install optimizations, warmup phase handling, vpp restart handling. Resulted in improved stability and reduced total execution time by 30% for single pkt size e.g. 64B/78B. - *General code housekeeping* - ongoing RF keywords optimizations, removal of redundant RF keywords. Performance Changes ------------------- Relative performance changes in measured packet throughput in CSIT |release| are calculated against the results from CSIT |release-1| report. Listed mean and standard deviation values are computed based on a series of the same tests executed against respective VPP releases to verify test results repeatibility, with percentage change calculated for mean values. Note that the standard deviation is quite high for a small number of packet throughput tests, what indicates poor test results repeatability and makes the relative change of mean throughput value not fully representative for these tests. The root causes behind poor results repeatibility vary between the test cases. NDR Changes ~~~~~~~~~~~ NDR small packet throughput changes between releases are available in a CSV and pretty ASCII formats: - `csv format for 1t1c <../_static/vpp/performance-changes-1t1c-ndr.csv>`_, - `csv format for 2t2c <../_static/vpp/performance-changes-2t2c-ndr.csv>`_, - `pretty ASCII format for 1t1c <../_static/vpp/performance-changes-1t1c-ndr.txt>`_, - `pretty ASCII format for 2t2c <../_static/vpp/performance-changes-2t2c-ndr.txt>`_. .. note:: Test results have been generated by `FD.io test executor vpp performance jobs`_ with Robot Framework result files csit-vpp-perf-|srelease|-\*.zip `archived here <../_static/archive/>`_. PDR Changes ~~~~~~~~~~~ NDR small packet throughput changes between releases are available in a CSV and pretty ASCII formats: - `csv format for 1t1c <../_static/vpp/performance-changes-1t1c-pdr.csv>`_, - `csv format for 2t2c <../_static/vpp/performance-changes-2t2c-pdr.csv>`_, - `pretty ASCII format for 1t1c <../_static/vpp/performance-changes-1t1c-pdr.txt>`_, - `pretty ASCII format for 2t2c <../_static/vpp/performance-changes-2t2c-pdr.txt>`_. .. note:: Test results have been generated by `FD.io test executor vpp performance jobs`_ with Robot Framework result files csit-vpp-perf-|srelease|-\*.zip `archived here <../_static/archive/>`_. MRR Changes ~~~~~~~~~~~ MRR small packet throughput changes between releases are available in a CSV and pretty ASCII formats: - `csv format for 1t1c <../_static/vpp/performance-changes-1t1c-mrr.csv>`_, - `csv format for 2t2c <../_static/vpp/performance-changes-2t2c-mrr.csv>`_, - `csv format for 4t4c <../_static/vpp/performance-changes-4t4c-mrr.csv>`_, - `pretty ASCII format for 1t1c <../_static/vpp/performance-changes-1t1c-mrr.txt>`_, - `pretty ASCII format for 2t2c <../_static/vpp/performance-changes-2t2c-mrr.txt>`_, - `pretty ASCII format for 4t4c <../_static/vpp/performance-changes-4t4c-mrr.txt>`_. .. note:: Test results have been generated by `FD.io test executor vpp mrr jobs `_ with Robot Framework result files csit-vpp-perf-mrr-daily-master__*__output.xml.gz `archived here `_. Comparison Across Testbeds -------------------------- .. warning:: TODO: Add: Table 1. Test Case 3-Node Hsw 3-Node Skx Skx vs. Hsw Delta [%] Table 2. Test Case 3-Node Skx 2-Node Skx 2-Node vs. 3-Node Delta [%] Throughput Trending ------------------- In addition to reporting throughput changes between VPP releases, CSIT provides continuous performance trending for VPP master branch: #. `VPP Performance Dashboard `_ - per VPP test case throughput trend, trend compliance and summary of detected anomalies. #. `Trending Methodology `_ - throughput test metrics, trend calculations and anomaly classification (progression, regression, outlier). #. `Trendline Graphs `_ - per VPP build MRR throughput measurements against the trendline with anomaly highlights, with associated CSIT test jobs. Known Issues ------------ List of known issues in CSIT |release| for VPP performance tests: +---+-------------------------------------------------+------------+-----------------------------------------------------------------+ | # | Issue | Jira ID | Description | +===+=================================================+============+=================================================================+ | 1 | Sporadic (1 in 200) NDR discovery test failures | CSIT-570 | DPDK reporting rx-errors, indicating L1 issue. Suspected issue | | | on x520. | | with HW combination of X710-X520 in LF testbeds. Not observed | | | | | outside of LF testbeds. | +---+-------------------------------------------------+------------+-----------------------------------------------------------------+ | 2 | Lower than expected NDR throughput of DPDK | CSIT-571 | Suspected NIC firmware or DPDK driver issue affecting NDR and | | | testpmd and VPP L2 path NDR throughput with | | PDR throughput on XL710 and X710 NICs. | | | xl710 and x710 NICs, compared to x520 NICs. | | | +---+-------------------------------------------------+------------+-----------------------------------------------------------------+ | 3 | Tagged Ethernet dot1q and dot1ad L2 path | CSIT-1066 | Tagged Ethernet dot1q and dot1ad L2 path throughput regression: | | | throughput regression. | | NDR -2%..-5%, PDR -2%..-6%, MRR. Affects l2xc and l2bd | | | | | performance tests. | +---+-------------------------------------------------+------------+-----------------------------------------------------------------+ | 4 | IPSec (software, no QAT HW) throughput | CSIT-1064 | IPSec throughput regression: NDR -3%..-8%, PDR -2%..-8%, MRR | | | regression. | | -3%..-7%. Affects IPSec SW tests, QAT HW tests not affected. | +---+-------------------------------------------------+------------+-----------------------------------------------------------------+ | 5 | High failure rate of creating working container | CSIT-1065 | About 20% of orchestrated container topology tests failing data | | | topologies with K8s/Ligato orchestration. | | plane verification indicating configuration issue. Suspected | | | | | issue with Ligato vpp-agent. | +---+-------------------------------------------------+------------+-----------------------------------------------------------------+