13:17:22 <bryan_att> #startmeeting OPNFV Traffic Profiles
13:17:22 <collabot> Meeting started Wed Aug 26 13:17:22 2015 UTC.  The chair is bryan_att. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:17:22 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:17:22 <collabot> The meeting name has been set to 'opnfv_traffic_profiles'
13:17:26 <collabot> morgan_orange: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
13:17:42 <morgan_orange> oups You already created one
13:17:45 <morgan_orange> #info Morgan Richomme
13:17:51 <bryan_att> #info Bryan Sullivan
13:19:02 <bryan_att> #info So I uploaded the mobility profile to gerrit, as far as it has been developed
13:19:15 <morgan_orange> saw that, I made some comments
13:19:23 <morgan_orange> in gerrit
13:19:28 <morgan_orange> more questions than comments
13:19:47 <bryan_att> #info There was a request in Yardstick to add a control plane component which is not complete yet, so there is a placeholder
13:20:21 <bryan_att> let me look at the gerrit comments
13:21:03 <morgan_orange> comments on space (could be a problem when automatically generating the doc)
13:21:35 <bryan_att> not sure what you mean by space - a formatting issue?
13:21:45 <morgan_orange> yes, if you edit the file
13:21:54 <morgan_orange> https://gerrit.opnfv.org/gerrit/#/c/1312/1/commons/mobile/readme.rst
13:22:04 <morgan_orange> you can see lots of space in red
13:22:05 <bryan_att> what exactly is wrong?
13:22:25 <morgan_orange> there are lots of useless space characters
13:22:40 <morgan_orange> that could lead to strange format when producing the html or the pdf from the rst file
13:22:48 <morgan_orange> just to delete the space character
13:23:04 <bryan_att> OK, it renders OK in the tool I use but I wil replace them
13:23:17 <morgan_orange> sometimes the ediotr add spaces
13:23:43 <morgan_orange> in R1 I had similar issues with the functest guides, but gerrit detect them properly
13:24:22 <bryan_att> #info I'll add some more background description to address the questions
13:24:57 <bryan_att> #info The profile is "per average user session"
13:25:51 <bryan_att> #info I don't have busy-hour/curve info but I'm not sure how that would help in how we intend to use the profile
13:25:53 <morgan_orange> so does it make sense to provide several profiles? average (as it is) heavy data consummer?
13:26:30 <morgan_orange> was wondering also if temporal view would make sense average traffic over time (for duration/robustness testcases)
13:27:15 <morgan_orange> hi Malla_
13:27:31 <bryan_att> #info Tweaking a load generator to increase the load based upon this average, or to simulate load variation over time is something I would say is more test design, not profile
13:27:32 <morgan_orange> view from NTT on traffic profile will be interesting :)
13:28:11 <bryan_att> #info The point of this profile is to provide an average traffic mix for an average session
13:28:22 <morgan_orange> ok
13:29:01 <morgan_orange> bryan_att: last question was more due to the influence of CI
13:29:20 <morgan_orange> I think it could make sense to provide a json description in addition of the guide
13:29:28 <morgan_orange> but it is not very important
13:29:34 <Malla_> HI Morgan
13:29:47 <bryan_att> #info For the json description are there any standard formats for that?
13:30:31 <morgan_orange> not for traffic profile as far as I know but it is just a way to represent what you have in the table
13:30:53 <morgan_orange> so it is not a big deal to move from the text representation to a json one
13:31:14 <bryan_att> #info I was thinking it might help drive more consistent traffic generator behavior if we did have some semantic expression of the profile, but it depends what they support, and it seems there is no standard
13:31:49 <morgan_orange> agree, it helps defining a formal view
13:31:55 <bryan_att> #info But I can propose a json representation if one says they can use it
13:32:17 <morgan_orange> could be interested to get trevor_intel view and more generally the ETSI guy view
13:32:45 <morgan_orange> I know that there is a document on test pre-requisites and as far as I remeber there are some traffic (document initiated by Spirent and Ixia)
13:33:26 <bryan_att> OK. are those docs on the wiki or git?
13:33:32 <morgan_orange> coudl be interesting to ask them (Pierre from Ixia was on the bridge some minutes ago) iben shall not be far and trevor_intel, are you there?
13:34:06 <morgan_orange> no they should be on ETSI portal....I think it is not published yet
13:34:53 <morgan_orange> #link http://nfvwiki.etsi.org/index.php?title=HT02_-_Test_Methodology_for_NFV
13:35:12 <morgan_orange> #link http://docbox.etsi.org/ISG/NFV/Open/Drafts/TST001_-_Pre-deployment_Validation/
13:35:45 <bryan_att> #info One more thing; the wiki https://wiki.opnfv.org/opnfv_functional_testing
13:36:22 <bryan_att> #info I added a link to the traffic profile page https://wiki.opnfv.org/traffic_profile_management
13:36:43 <bryan_att> #info How much do we want to maintain on the wiki vs a document in git?
13:36:53 <morgan_orange> the reference is git
13:37:14 <morgan_orange> wiki is used to communicate and maintenance is in best effort
13:37:32 <morgan_orange> but before people can contribute or dive into git , it is good to have something publically available
13:37:50 <bryan_att> #info I'd rather just point the wiki to the doc in git, and migrate the wiki content there
13:38:04 <morgan_orange> ok
13:38:13 <morgan_orange> avoid divergence between the 2 sources
13:38:39 <morgan_orange> do you plan to automatically generate the doc? so the pointer will be on the git or on the artifact?
13:39:09 <bryan_att> #info Yes, I don't want to confuse anyone. I think much of the rationale info as to why there are multiple profiles etc should be in a doc
13:39:25 <morgan_orange> ok
13:39:32 <bryan_att> #info The doc will be auto-generated and linked to the wiki, not scraped
13:40:20 <bryan_att> #info It's easy ala the link to the "Current working draft " on Copper
13:40:36 <bryan_att> #info Linked from the artifacts page
13:42:11 <bryan_att> #info Last thing; I will send out a note to collect the other profiles into the same repo
13:42:24 <morgan_orange> shall we discuss of the list of profiles that could be useful and how to motivate contributors?
13:42:37 <bryan_att> sure
13:42:56 <morgan_orange> I think a communication on the thursday meeting will be needed to share with all the test projects our wish to define reference profiles
13:43:24 <morgan_orange> regarding the list of profiles, I added IMS residential as we plan to add a vIMS testcase in B-Release
13:43:34 <morgan_orange> but then I have a question
13:43:40 <bryan_att> ok
13:43:54 <morgan_orange> we can consider a profile from a pure IP point of view: nb of UDP, TCP packets,
13:44:14 <morgan_orange> but to generate traffic we need also a description at L7 level: Nb of Register, Invite
13:44:17 <morgan_orange> ...
13:44:24 <morgan_orange> not easy to summarize
13:44:49 <bryan_att> #info There is an open section in the profile for a control plane profile part
13:45:31 <morgan_orange> ok
13:45:43 <bryan_att> #info For now through the "session management" overhead will be developed as part of turning the profile into a specific test scenarion for a "DUT/SUT"
13:46:17 <morgan_orange> OK
13:46:22 <bryan_att> #info For example, a SGW to Gi-Lan elements to FW/NAT
13:47:18 <bryan_att> #info We have to craft a session profile in order to generate the traffic - otherwise we are just blasting data at a certain mix
13:48:08 <morgan_orange> :)
13:48:18 <bryan_att> #info So there is a placeholder, and I will work with Ixia and Spirent on how to describe the session (control plane) profiles
13:48:36 <morgan_orange> ok
13:48:37 <bryan_att> #info If anyone has something they want to bring to the table then great
13:49:10 <bryan_att> #info Otherwise it will be like this: A session for a particular application lasts an average time
13:49:24 <morgan_orange> could it be the occasion to try to motivate Telco members? providing realistic real inputs shall be possible even from the Telco that did not contribute so far?
13:49:29 <bryan_att> #info During that session there may be a mix of applications used in varying percentages
13:50:08 <bryan_att> #info The applications have certain control plane protocol implications (number/type of messages) per the standards
13:50:13 <morgan_orange> could be interesting to send a mail to the mailing list for them and start listing other profiles we will need figure (cable, metro, busines,..)
13:50:55 <bryan_att> #info It would be great to get end-user input into the session / control plane profiles
13:51:11 <bryan_att> #info I will send a note for input, and also basic design ideas
13:51:34 <morgan_orange> I will try to find some additional resources in Orange to work on this topic
13:51:38 <bryan_att> #info I am not an expert in this but I developed profiles for WAP testing back in the day...
13:52:14 <bryan_att> #info And most apps including streaming were included in the tests
13:52:24 <morgan_orange> pretty sure we have experts on this kind of topics not far from us (not an expert too)..just have to find them..
13:52:44 <bryan_att> #info But mobility traffic is at the same time more complex today (more apps) and more skewed toward one type of app (video)
13:52:55 <morgan_orange> dunno if mobility traffic is different in the US and in Europe
13:53:27 <bryan_att> #info We are probably good with one global profile
13:53:38 <bryan_att> OK, that's about it from me for now
13:53:47 <bryan_att> Any other points?
13:53:50 <morgan_orange> same for me, thanks for leading that task
13:54:24 <bryan_att> sure, my pleasure
13:54:30 <bryan_att> ok I will end the meeting
13:54:38 <bryan_att> #endmeeting