14:01:19 #startmeeting telcowg 14:01:20 Meeting started Wed Mar 11 14:01:19 2015 UTC and is due to finish in 60 minutes. The chair is sgordon. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:21 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:24 The meeting name has been set to 'telcowg' 14:01:28 #link https://etherpad.openstack.org/p/nfv-meeting-agenda 14:01:30 hi 14:01:34 #topic roll call 14:01:45 hi 14:01:46 hi 14:01:49 hi 14:01:53 * beagles waves, panting from running from other room 14:02:19 #topic action items from last week 14:02:41 #info aveiga's basic use case template was merged 14:02:59 #link http://git.openstack.org/cgit/stackforge/telcowg-usecases/tree/template.rst 14:03:18 #info sgordon submitted cloudon's vIMS use case for review - needs some further updates to meet template 14:03:24 sgordon: the basic template seems not to work for the doc build :( 14:03:34 #link https://review.openstack.org/158997 14:03:43 #info daschab submitted security segregation use case for review 14:03:48 #link https://review.openstack.org/163399 14:04:01 #info mkoderer notes that the basic template seems not to work for the doc build 14:04:07 mkoderer, do you have a link to the fail 14:04:12 sgordon: https://review.openstack.org/#/c/163399/ 14:04:46 well that is a friendly error isnt it 14:04:50 sphinx is failing with "WARNING: Literal block expected; none found" 14:05:05 sgordon: I need to check what goes wrong there 14:05:35 so in the original review where we merged the template it did build: https://review.openstack.org/#/c/158028/ 14:05:51 so it must be something specific to the changes in that submission - i guess? 14:06:06 sgordon: I guess the template won't be used for the spinx build 14:06:48 sgordon: and the IMS one run through because you used the wrong directory ;) 14:06:48 mkoderer, let's try work on that after the meeting 14:06:49 https://review.openstack.org/#/c/158997/ 14:06:51 hahaha 14:06:51 #help to corelate mNO/MVNO use case with IPv6 OPNFV when time permits - Thanks 14:06:56 sgordon: ok 14:06:57 yeah i havent rebased it yet 14:07:25 sooo 14:07:30 #topic operators mid-cycle 14:07:45 we had a telco working session at the operators mid-cycle yesterday 14:07:51 #link https://etherpad.openstack.org/p/PHL-ops-telco 14:08:03 minutes are recorded in the above etherpad 14:08:24 we had a mix of people who have been in these meetings and some who were coming at it for the first time 14:08:52 and i think it's fair to say that the main topic of conversation was interaction with OPNFV and how the two groups fit together 14:09:18 there was also an OPNFV call on the topic of working with openstack earlier in the day 14:09:20 what are 'the two groups'? 14:09:37 gongysh, OPNFV and OpenStack Telco Working group 14:09:47 thanks 14:10:20 we are in a bit of an odd place because we created the latter as a place to collaborate on NFV and telco use cases and their openstack requirements before OPNFV existed 14:10:21 #info I am working in OPNFV and as well in teclowg since last year but missed lot here 14:11:13 the main action item that came out of that was for me to re-connect with chris price on the OPNFV side and start some more cross-threads to discuss further 14:11:33 we did also walk through the way we are handling use case submission and review at the moment 14:11:53 Lots of OpenStack requirements going to be developed in both OPNFV and ETSI-NFV phase 2 14:12:08 #info The Community page ion OPNFV has Openstack listed and is working and we can co-ordiate with it 14:12:21 well this is really the issue 14:12:30 how? 14:12:43 - ETSI, OPNFV, and this group, all currently define/update use cases 14:12:56 - some are members of one but not the others 14:13:11 - how to avoid duplication 14:13:26 #link https://wiki.opnfv.org/community/openstack 14:13:29 if we share common discussion 14:13:52 at least for defining use cases 14:14:18 Re ETSI-NFV Phase 2, I suspect most of the OpenStack related item will come out of the IFA stream. 14:14:19 #link https://wiki.opnfv.org/community/openstack 14:14:51 adrian-hoban, IFA? 14:15:09 We need to define the scope for all these groups... 14:15:10 if i am being brutally honest i am more concerned about the OPNFV interaction versus ETSI, unless we are expecting significantly more timely publication of the next round of ETSI outputs 14:15:53 OPNFV are defining and pushing out use cases, gaps, requirements for us to interact with now 14:15:55 sgordon, do u mean to collabrate with any one? 14:16:14 vks: IFA = Interfaces and Architecture Working Group 14:16:20 #info Sean Rob from EMC/Vmware and visited OPNFV and had indicated the use case can be co-ordinaged by product team too 14:16:33 vks, the ETSI-NFV collaboration with this group thus far has been "here is a PDF of things and you already did half of them because we took so long" 14:16:41 ;) 14:16:44 :-) 14:16:56 Can I #link? 14:16:57 ;) 14:17:11 #info we can do agap analysis between projects listed in link I showed earlier (requirement) projects with openstack modules 14:17:16 #link https://wiki.opnfv.org/community/openstack 14:17:41 dneary, http://lists.openstack.org/pipermail/openstack-dev/2014-October/048021.html 14:17:48 dneary, we already linked that page 14:18:02 sgordon, Great, thanks 14:18:23 so what i am getting at is i see more immediate overlap with OPNFV activities than ETSI-NFV ones 14:19:11 Yes. Although the ETSI NFV Open Source WI is also focussing on OpenStack interactions 14:19:17 #info But this is as of October 2014 , is it current? 14:19:31 sgordon, i think we are in better position now 14:19:40 we have few use cases with us 14:19:46 I don't know if there is a specific mailing list, or SPOC, for that work item though 14:19:48 rprakash, this is the thing 14:19:52 #info I though we are talking of gap between OPNFV requirments project and Openstack Modules 14:19:57 as far as i know no newer version has been published 14:20:26 rprakash, can you stop trying to info things that are just your commentary? 14:20:32 thanks 14:20:32 sgordon: I think that's a fair assessment. There's a but... If we collectively don't influence the ETSI-NFV specs, there are requirements coming down the pipe that will be difficult to implement... 14:20:55 sure, but how much of that influence is co-ordinated via this effort today? 14:21:03 i think a lot of participants dont have the visibility 14:21:54 IMHO we are currently only keeping our self busy :) 14:22:26 #info I will try get that for IFA 005-006 for NB Interfaces and that may be say next week 14:22:26 #info VIM NB Interfaces 14:22:27 OK I need some help in irc practices will do 14:22:31 rprakash, Even in OPNFV the discussion has moved past that, I think - we're breaking down requirements projects into individual feature requests - and in the future that will be Gerrit patch submissions and bug reports 14:22:43 at the end nobody will get the needed changes implemented 14:22:58 mkoderer, indeed and that is really the key 14:23:13 identifying where the body of the participants actually willing to do the work are connected to each other 14:23:59 sgordon: Working with the OPNFV community makes sense and will have more developer representation 14:24:20 adrian-hoban_, +1 14:24:39 adrian-hoban_: but how... 14:24:42 that is my suspicion, also it looks like dneary fearlessly volunteered to maintain a wiki of the changes they are submitting 14:24:43 HOWEVER 14:25:04 sgordon, Yes, I am that sucker 14:25:07 the concern raised with this was it excludes NFV and telco users, including some who are members of this group, that are not members of OPNFV 14:25:22 ...concern raised at the mid-cycle session yesterday that is 14:25:42 sgordon, I hope that the only metadata we'll have there which won't be upstream is the OPNFV project name it came from 14:25:53 this comes back to how we create some loose framework for collaboration here 14:26:10 versus the current situation where OPNFV members submit their use cases to both groups, for example 14:26:35 sgordon, rprakash was asking me what belongs in OPNFV and what belongs in OpenStack - would it be useful to clarify that (from the OPNFV point of view)? 14:26:46 this is not a problem i expect to solve today but what i took the AI to try and get at 14:27:06 dneary, sure - but is there a common OPNFV POV on this? 14:27:15 sgordon, Getting towards it... 14:28:25 OPNFV - take telco use-cases defined in telco language, and convert them into requirements expressed in non-telco-readable language; from that break high level requirements into actionable feature requests for upstream projects (including OpenStack) 14:28:44 #info Clarification needed to establish interaction/overlap of ETSI-NFV, OPNFV, and OpenStack Telco Working group 14:28:49 Is it recommended to use new template listed to be used for use cases rather than one we used earlier as nfv group?4 14:28:59 OpenStack: OPNFV members submit blueprints, and track those blueprints through review, code creation, code review and merge 14:29:37 For the moment, OPNFV is essentially taking the product of ETSI NFV ISG as the higher level telco use-case inputs 14:29:39 dneary: seems like an overlap to this group 14:29:48 dneary, really this asks more questions than it answers 14:29:50 mkoderer, +1 14:29:59 So, for ETSI members, there are 2 opportunities to collaborate: 14:30:03 "The working group aims to define the use cases and identify and prioritise the requirements which are needed to deploy, manage, and run telecommunication services on top of OpenStack. This work includes identifying functional gaps, creating blueprints, submitting and reviewing patches to the relevant OpenStack projects and tracking their completion in support of telecommunication services." 14:30:24 1. Breaking down the high level requests inside OPNFV (you don't need to be a member to participate in a specific project) 14:30:25 #ilink http://git.openstack.org/cgit/stackforge/telcowg-usecases/tree/template.rst 14:30:37 2. Participating in the review & refinement of blueprints 14:30:39 I mean this template 14:31:09 rprakash, im not really clear on the question - at this time if you were to submit to our repo we would expect you to use our template, otherwise we wouldnt merge it 14:31:43 what we're talking about though is whether we're just spinning our wheels here and that use case definition belongs in opnfv 14:31:45 sgordon, Yes, IMHO, there is overlap with this group. Ensuring that people in this group are ensuring that OPNFV's product is useful to OpenStack, and ensuring that OPNFV people are participating here, is a key goal. 14:31:54 OK thta was my question , is this new template and are we doing away with old one in telcowg 14:31:57 what we need to see though is that the use case definitions are in fact in non-telco speak 14:32:01 The delta is where OPNFV includes in scope projects other than OpenStack 14:32:04 and are coming through with the blueprints and specs 14:32:18 which at the moment based on the proposals i am seeing in the opnfv list is not necessarily the case 14:32:20 sgordon, Agreed 14:32:29 sgordon, That's been a challenge up to this point 14:32:54 sgordon, maybe adding that ETSI NFV has activity to support with creating use cases in non-telco language along the lines of their requirements 14:33:06 Project reviews and blueprints have not been getting enough eyeballs. We're making progress in some cases, but there are a lot of OPNFV projects now 14:34:23 fzdarsky: I don't think you'll get that from ETSI-NFV in a form that will be usable here. 14:34:30 fzdarsky, honestly i would again prefer to focus on OPNFV 14:34:52 if something comes out of ETSI-NFV that is directly understandable by someone coming at it from a non-telco case then i will be the first to take notice 14:34:59 let me ask you I have specific MNO-MVNO use case in telco-wg under prep work how do I deal this wrt telcowg and opnfv, help me build this 14:35:01 but need to see evidence of that actually happening before i will believe it 14:35:06 adrian-hoban_, I think doing that in ETSI NFV is the only way to ensure full coverage 14:35:22 rprakash, What is MNO-MVNO? 14:35:28 (actually, that's kind of the point) 14:35:29 The ETSI-NFV normative work has to stay technology neutron, so there is a level of abstraction between what is developed there, and what shows up here in blueprints 14:35:42 dneary, yes we covered this in a previous meeting also 14:35:52 Mobile Network Operator & Mobile Virtual Network Operator 14:36:03 the use case proposal was very acronym heavy without necessarily explaining itself 14:36:13 adrian-hoban_, that is exactly why ETSI NFV created the work item for interfacing with open-source 14:36:39 adrian-hoban_, so that IFA WG et al can stay technology neutral 14:36:50 #info ETSI NIFV IFA and other groups publish and only published reports we can use 14:37:06 #info OPNFV aims overlap significantly with telco working group aims 14:37:25 fzdarsky: I suspect that for this context, it will end up being a close collaboration with OPNFV, which brings us back to sgordon's initial point 14:37:33 #info ETSI-NFV Open Source work item may overlap but not as clear at this point 14:37:51 I thought that helps co-ordiate activities as we can use synergy in both forums 14:38:13 in my opinion working close with OPNFV now will be gud for now 14:38:25 #info sgordon and dneary agree OPNFV project reviews and blueprints need more eyeballs - aim being combination of both understandable/clear use case and implementation proposal make it through to OpenStack specification 14:38:47 vks, +1 14:39:01 i think this is where there is the most obvious overlap *today* 14:39:22 So - how can I help enable that work together? 14:39:24 can we then agree that let requirements come form OPNFV but use case be built here in telcowg? 14:39:38 dneary, so what is the process in opnfv land for submitting/reviewing a use case 14:39:45 I can identify the 3-4 most important meetings for OpenStack people to be present? 14:39:46 rprakash: I don't agree 14:39:48 rprakash, i dont think we can 14:39:57 sgordon, first by segregating use cases at single repo 14:39:58 for starters opnfv would have to agree 14:40:05 since that appears to be part of their mandate too 14:40:07 (use cases) 14:40:22 and as dneary said they are tracking at a slightly broader level 14:40:24 don't forget that there are policical things ongonig :) 14:40:37 in that the use case may cover multiple projects that include openstack 14:40:42 We won't join OPNFV in it's current state... 14:40:42 mkoderer, lol - oh yes 14:40:42 sgordon, OPNFV has requirements projects, the idea is to express the high level requirements that a use-case places on the platform (some examples: multi-site, IPv6, fault management, resource reservation) 14:40:46 sgordon, i meant to say blueprints whatever we have now 14:40:56 mkoderer, right and this comes back to my biggest concern 14:41:07 what is the process for people who arent OPNFV members and dont want to be 14:41:19 because aveiga and amitry are also not currently 14:41:20 Each requirements project comes up with a project definition that should say what the work result of the project will be, the upstream projects affected, and the people committed to working on the project 14:41:42 so that's basically 4-5 of our regular contributors who arent in OPNFV at all 14:41:56 Being a OpenStack foundation memeber should be sufficent to file use cases :) 14:42:00 vks, ack 14:42:01 Once the project is approved by the OPNFV TSC (approval process mostly rotates around clarifying project definition and deliverables, no attention is given to implementation details in my experience) 14:42:28 #info mkoderer notes not everyone is an OPNFV member, not necessarily open to becoming one at this time 14:42:43 #undo 14:42:44 Removing item from minutes: 14:42:44 once approved, project participants break the higher use case into blueprints, feature requests, project proposals in upstream projects 14:42:48 #info mkoderer notes not everyone is an OPNFV member, nor necessarily open to becoming one at this time 14:43:13 And those should go through internal review in the project, plus submission as WIP in specs repos upstream for upstream revision & review 14:43:14 sgordon, ? 14:43:15 mkoderer : Openstack may fail at that because Telco use cases don't get enough attention in core projects 14:43:18 That's the ideal 14:43:46 dneary: agree - in OPNFV you simply propose a project, and if you fill in the forms right and convince the technical steering committee it's in scope and not conflicting with other OPNFV projects then you are a go 14:44:05 #info any linux foundation account holder can participate in OPNFV as volunteer 14:44:06 dneary: I think we need any requirements coming from OPNFV to pass through this WG first. If we don't get this in place we will end up in the same place we were in a year ago 14:44:06 matrohon: that's correct.. but OPNFV will fail too if they don't listen to their customers :) 14:44:08 mkoderer, sgordon , the only place where all stakeholders are represented and where normative req's are created is in ETSI. Therefore, we need a process to ensure OPNFV can take the ETSI req's and use cases and map them to upstream req's and blueprints 14:45:05 don't see anything in OPNFV processes yet which sets out how to interact with OpenStack community 14:45:13 adrian-hoban : +1 14:45:25 adrian-hoban_, In terms of project requirements, or blueprints? 14:45:44 #info ETSI use cses were for ETSI PoC and now for nfv pase 2 and not for OPNFV or Openstack 14:46:05 adrian-hoban_, To date, no requirements project has "completed" - that is, tested code, merged upstream 14:46:18 dneary, yeah but blueprints have been submitted 14:46:26 in a similar state to those we were getting 1+ years ago 14:46:36 :) 14:46:39 so i can see where adrian-hoban_ is going with that line of thought 14:46:59 now sure "we can fix that" but it would have been better to do it before submission 14:47:06 sgordon, What happened (as you know) is that the OPNFV project got busy when under the impression that Nova blueprints needed to be submitted in early March, after the Liberty branch was opened 14:47:11 fzdarsky: I think the flow is IFA --> TST --> OPNFV --> Telco WG --> OpenStack Projects 14:47:28 adrian-hoban, +1 14:47:29 adrian-hoban: that's a monster 14:47:43 dneary: I got disconnected and missed your question 14:47:44 OK lets agree that Blue Prints coming from OPNFV requiremnts be whetted by teclowg 14:47:49 mkoderer, but that is probably necessary, if you want full representation 14:48:06 adrian-hoban: we won't to build an environment in 2015 and not in 2030 :) 14:48:14 sgordon, We're now post-processing those to ensure they're useful to OpenStack 14:48:22 There are some short-cuts 14:48:25 The telcowg then manage the passage of BP to approval in openstack 14:48:33 Many of us have folks in all these projects 14:48:40 adrian-hoban, Indeed 14:48:55 Much redundancy, many repeated conversations 14:49:12 adrian-hoban, yes, but from experience, those people do not communicate well in all companies 14:49:24 well/at all 14:49:33 A BP in openstack does it have a hold status so that telcowg can weekly clear or correct what OPNFV comes up with? 14:49:48 until someone actually approves it 14:49:53 it's in a hold status by default ;) 14:50:16 So we ask OPNFV to start a BP with Hold staus 14:50:40 what i want to get back to is mkoderer's question though 14:50:48 i thought we are trying to collaborate with OPNFV, but seems going in other drection 14:50:55 about where use cases are tracked particularly if you are not part of OPNFV 14:50:59 vks, how so? 14:51:17 And let telcowg rework it to meet the openstack module requirments? 14:52:01 Plus there are cases where cross module BPs are required and that is not the capability of OPNFV let teclowg suggest how to handle that 14:52:22 sgordon, so going by above conversation this group participation in BP's will be zero 14:52:56 Which group? you mean telcowg ? Not at all 14:53:03 yes 14:53:21 vks, i dont see where you got that from 14:53:53 sgordon : the Use case repo makes sense for people like mkoderer : OPNFV teams can still review them or even register new Use cases in it 14:54:14 i think what we keep coming around on in this conversation is that the use case definition is the primary overlap between opnfv and this group 14:54:15 One of the reasons we set up this Telco WG was a request from various PTLs and core devs for the Telco community to collaborate and agree on an approach before broadening the review requests. I don't think that fundamental has changed 14:54:28 See let each requirement project come to this meeting from OPNFV to telcowg, present gt advice from telcowg before they initiate even BP 14:54:28 blueprints although discussed in groups are ultimately owned by an individual 14:54:41 we can collaborate on making sure they are likely to succeed and meet everyone's needs though 14:54:54 matrohon, +1 14:54:54 #info "One of the reasons we set up this Telco WG was a request from various PTLs and core devs for the Telco community to collaborate and agree on an approach before broadening the review requests. I don't think that fundamental has changed" 14:55:12 adrian-hoban, big +1 - and though i see opnfv as a possible step towards handling this it's clearly not there yet 14:55:18 Yes agreed but there needs be communication or join meet between the two in Openstack telcowg to help that happen 14:55:32 rprakash, isnt that how i started the meeting? 14:55:42 i was also on the opnfv call on this topic y'day 14:56:18 what we need to get to is a little clearer division of responsibilities though 14:56:33 sgordon: +1 14:56:37 May be I missed that but then we must get dneary who has been leading OPNFV side to work with you on that 14:57:00 sgordon, +1 14:57:48 sgordon, re: Where are use cases tracked, seems like OPNFV is the current best place. 14:57:58 dneary, unless your not in OPNFV 14:58:01 I understand the "not everyone is in OPNFV" argument 14:58:05 im still concerned i am seeing no answer to that 14:58:06 sgordon, +1 . kind od RACI table. with owner and deadline... 14:58:18 But people who are not paying members can still participate (unlike ETSI NFV) 14:58:41 And it's sufficiently telco focussed that we have a decent chance that use-cases as defined will be useful to that audience 14:58:41 technically yes, but is that politically realistic 14:58:58 sgordon, Depends 14:58:58 dneary, you don't have to pay to participate in ETSI NFV :) 14:59:02 OK let us ask each requirments project in OPNFV to present their cases to telcowg, and then jointly plan what BPs are required in which module to make that happen 14:59:02 dneary: OPNFV is seen as vendor drivern maybe? 14:59:22 #action sgordon to kick off thread of conversation to try and define a clearer framework for who does what (OPNFV versus telcowg) 14:59:27 IMHO the best place for telco usecases is our review process :) 14:59:49 it's a tranperant process and I don't need any commitee agreement 14:59:49 sgordon, If there is an impression that requirements are being "done wrong" by OPNFV, to the detriment of non-OPNFV participants, the incentive to get involved and "fix" them grows - especially if the requirements coming from OPNFV result in code 15:00:18 dneary, i think the incentive at the moment is low because they are done wrong enough that they and code associated with them wont be accepted 15:00:19 At the end of the day, if neither OPNFV or ETSI NFV results in code being written and merged into OpenStack, it doesn't matter which we talk to 15:00:30 Lets ask how can telcowg help OPNFV requirment folks to translate their requirments to BP, can some one propose a process 15:00:30 dneary: +1 15:00:50 meeting time ends :) 15:00:51 * dneary has to drop for another meeting 15:01:07 mkoderer, right - if we did do something with OPNFV for use cases i would like to see them at least move to similar technology/tooling for the reviewing 15:01:08 not clear now 15:01:13 we have hit our hard stop 15:01:22 let's continue in #openstack-nfv or on the mailing list 15:01:32 thanks all - like i said this was the hot topic in the room yesterday 15:01:36 wanted to bring everyone up to date 15:01:40 #endmeeting