14:03:02 <mdgray_> #startmeeting Weekly Open vSwitch for NFV meeting (28th November 2016)
14:03:02 <collabot`> Meeting started Mon Nov 28 14:03:02 2016 UTC.  The chair is mdgray_. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:02 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:03:02 <collabot`> The meeting name has been set to 'weekly_open_vswitch_for_nfv_meeting__28th_november_2016_'
14:03:09 <mdgray_> #topics Roll Call
14:03:15 <mdgray_> #info Mark Gray
14:03:24 <mdgray_> #topic Roll Call
14:03:27 <mdgray_> #info Mark Gray
14:03:28 <lch> #info Chunghan Lee
14:03:41 <tfherbert> #info Tom Herbert
14:03:55 <lch> #info Chunghan Lee
14:03:56 <billyo_> #info Billy O'Mahony
14:04:54 <mdgray_> #topic Readout on the priority path development work (Billy)
14:06:52 <mdgray_> #info Priority path use case is that we need to enable some traffic class with better latency characteristics that others.
14:07:16 <mdgray_> #info Billy presents tests done to measure the current behaviour in OVS.
14:08:35 <mdgray_> #info Billy used three traffic classes (Bulk data transfer, Voice, PTP)
14:11:38 <mdgray_> #info Billy presented latency data for different traffic classes
14:15:10 <mdgray_> #info As the PTP rate is so low it causes a sampling error. There are not enough PTP frames to show the long tail in latency. This is currently speculation. Billy will confirm this.
14:15:45 <mdgray_> #info These tests were a PVP test.
14:23:12 <mdgray_> #info Billy outlined what the graphs should look like. i.e. Latency spikes are only seen in the lower priority traffic
14:23:27 <mdgray_> #info Billy also opened a discussion on the ovs-discuss mailing list
14:23:38 <mdgray_> #link https://mail.openvswitch.org/pipermail/ovs-discuss/2016-November/043081.html
14:23:55 <mdgray_> #topic Colorado Fuel Performance Results (Billy)
14:24:03 <mdgray_> #undo
14:24:03 <collabot`> Removing item from minutes: <MeetBot.ircmeeting.items.Topic object at 0x216d090>
14:24:12 <mdgray_> #info Billy will share excel spread sheet with results
14:24:21 <mdgray_> #topic Colorado Fuel Performance Results (Billy)
14:24:49 <mdgray_> #info Billy has seen a performance result in yardstick tests between Brahmaputra and Colorado
14:30:22 <mdgray_> #info Yardstick pktgen test shows that although ovs-dpdk performs better than vanilla ovs across Brahmaputra and Colorado. However, ovs-dpdk performance less better in colorado than brahmaputra.
14:30:52 <mdgray_> #info Ping latency test looks about the same for colorado and brahmaputra.
14:32:17 <mdgray_> #info netperf tests (which are not part of yardstick) also showed decrease performance.
14:33:00 <mdgray_> #info Billy has not identified the root cause.
14:33:24 <mdgray_> #info Current assumption is the Qemu version is causing problems with virtio.
14:34:16 <mdgray_> #info Brahmaputra was using a different version of Qemu.
14:35:42 <mdgray_> #info Currently, we do not have results for Apex in order to compare.
14:52:56 <mdgray_> #info This is because the Apex infrastructure in that POD is using Cisco UCS which is adding a vlan tag which is not needed. Tom has worked around this by adding a rule. However, we need a more permanent solution. Some solutions were discussed.
14:53:11 <mdgray_> #info There is a similar issue in the FDS project.
14:53:42 <mdgray_> #info Some consensus to fix this in OVS.
14:54:25 <mdgray_> #info When Tom submits the patch, he will cc Billy in order to help move this through the OVS patch process.
14:55:21 <mdgray_> #topic Other?
14:55:34 <mdgray_> #endmeeting