18:00:37 #startmeeting gluon 18:00:38 Meeting started Wed Nov 2 18:00:37 2016 UTC and is due to finish in 60 minutes. The chair is bh526r. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:41 The meeting name has been set to 'gluon' 18:00:49 Hi guys 18:01:07 Hi Bin 18:01:07 #topic Roll Call 18:01:14 Hi Jeff 18:01:15 Hi everyone 18:01:20 #info Bin Hu 18:01:23 #info Georg Kunz 18:01:25 Hi Jin 18:01:26 #info Jeff Collins 18:01:28 #info Nikolas Hermanns 18:01:30 Hi Georg 18:01:36 #info JinLi 18:01:39 Hi Niko 18:01:44 #info Juanma Fernandez 18:01:58 Hi Juanma, welcome on board 18:02:03 #info Tom Hambleton 18:02:07 Hi Tom 18:02:13 Hi 18:02:20 #info Kamal Hussain 18:02:21 Hi 18:02:25 thanks 18:02:37 #topic Admin Update 18:02:42 Hi Kamal 18:02:47 Hi Bin 18:02:53 #info No update 18:03:08 #topic Summary of OpenStack Summit 18:03:36 #info We have had great demonstration and discussion last week in Barcelona 18:04:10 #info Many participants showed their proof-of-concept of Gluon with various SDN controllers in their booths 18:04:29 #info We also had good discussion in join Nova-Neutron session 18:04:49 #link https://etherpad.openstack.org/p/ocata-nova-neutron-session 18:05:21 #info above link is the etherpad of join Nova-Neutron session. And Gluon discussion is at the end 18:06:19 #info The feature "Add the information Nova requires for a port to the port itself" will be pursued 18:06:31 #link https://review.openstack.org/#/c/390513/ 18:06:51 Hi all 18:07:12 #info And another feature "Late allocation of IPs during VIF binding" is already supported in Newton 18:07:22 #link https://review.openstack.org/#/c/390511/ 18:07:34 #info we need to test and verify it 18:08:03 #info Ian will follow up, and also work on related pieces in Neutron 18:08:37 #info We also worked out our Ocata work plan on Friday 18:08:40 #link https://etherpad.openstack.org/p/ocata-gluon-work-plan 18:09:13 #info Above link gives tasks, comments and assigned resources 18:09:32 Anything more from summit you want to add? 18:10:09 Ok from my side 18:10:28 Great. Let's switch gear to review patches 18:10:34 #topic Patch Review 18:10:56 Bin, did you add sfc to the agenda? 18:11:09 #info Jin's patch "Test Case Proposal" 18:11:14 #link https://review.openstack.org/#/c/388821/ 18:11:32 I think Paul had some comments. Jin, will you be able to address Paul's comments? 18:12:00 Niko - at the end. 18:12:25 Brady also said he may not join us 18:13:16 #info Jin is not available. So move on to next patch 18:13:29 #info Tom's Patch "Example Shim Server Improvements" 18:13:40 #link https://review.openstack.org/#/c/390019/ 18:14:05 yes, I will respond to Pauls 18:14:16 #info It's sitting there for almos 2 weeks, no comments 18:14:24 Cool, thank you Jin 18:14:49 #info I suggest we merge it. I already gave +2 18:15:48 If no objection, can Niko or Kamal +2 and merge it? 18:16:10 which one? 388821? 18:16:19 390019 18:16:31 The Shim Example Code one 18:16:35 No objection from me :-) 18:16:56 388821 had comments from Paul, and -1. Jin will address Paul's comments 18:17:06 ok I will take a look at it today 18:17:15 Thank you Kamal 18:17:55 #info All agreed to merge https://review.openstack.org/#/c/390019/ "Example Shim Server Improvements". Kamal will take a look and take action accordingly. 18:18:43 yes 18:18:51 #info Next Patch "WIP: Add unit tests for gluon/objects/base.py" 18:19:00 #link https://review.openstack.org/#/c/391894/ 18:19:57 looks like no one had read it 18:20:01 I have looked it over 18:20:48 Great. Any comment? 18:20:49 It looks really good. I plan to call him to discuss it in more detail 18:21:02 Cool. 18:21:24 #info Tom has looked at it, and it looked really good 18:21:37 I looked at it also. I can do some simular test case to other packages 18:21:41 There is some functionality in the base object that we are not using. So, some code cleanup could be done. 18:21:46 #info Tom will call the author (Trevor) to discuss it in more detail before moving forward 18:22:35 Sure, Tom. Perhaps let us have a code review session with him in a conference call bridge. Who else wants to be there? 18:22:54 I want to be there 18:23:41 For "WIP: Add unit tests for gluon/objects/base.py" I will do a longer review tomorrow. 18:23:54 Thank you Niko. 18:24:44 #info Niko will also do a comprehensive review tomorrow 18:25:23 #info Move on to next patch "First draft of port and service binding model for Gluon" 18:25:33 #link https://review.openstack.org/#/c/392250/ 18:26:17 this is mainly a trigger for discussion 18:26:34 #info several comments indicated it is "Very crisp and concise document. Good job!" 18:26:35 don´t consider this a final solution, but add your thoughts to the patchset 18:26:51 let´s different approaches there 18:27:02 +discuss 18:27:10 :) 18:28:15 I assume everyone needs some more time to review it, and write comments 18:28:31 correct - for me atleast :) 18:29:02 Let's put this on the agenda for the voice conference next week. 18:29:11 #info Everyone needs some more time to review it abd give comments. 18:29:16 Yup 18:29:40 ok, agreed 18:30:11 I think that's all of the new patches 18:30:47 #info Jin will address Paul's comments on 38821 in a new patchset 18:31:36 #info Everyone will take more time for comprehensive review on 391894 and 392250 18:31:52 Good 18:32:14 #topic SFC Model 18:32:31 Niko, now it is SFC 18:32:41 Will you give a brief introduction? 18:32:53 Ok 18:33:19 Basicly I would just like to start the dicussion about the SFC model in Gluon. 18:33:48 Do you mean the SFC Proton? 18:33:50 I think we should come up with a model and discuss it in the community then 18:33:53 yes 18:34:03 Good idea, we should 18:34:14 I think the best way to do it is to file an rst in the gluon repo 18:34:33 in the same patch we would then have a model as well. 18:34:52 if that is merged the different shim could be started. 18:35:04 anybody has a different view? 18:35:27 Niko, are you thinking on having an SFC model based on the ietf draft? 18:35:35 or something different? 18:35:38 Sounds like a good plan. 18:36:29 we would start with the IETF SFC data model or the ETSI VNF Forwarding Graph 18:36:30 yes 18:36:35 juanmafg: yes 18:36:39 as far as I see there are quite a lot of models, one is suggested by IETF, another by ETSI and another completely different proposed by NEtworking SFC in Openstack 18:36:42 ok 18:37:24 it sounds like a good thing 18:37:24 For SFC, there are 2 different technologies: NSH-based (L2) and VPN-based (L3) 18:37:29 I think we need to understand them all and then find the right one for gluon 18:37:48 networking-sfc are they doing VPN-based? 18:38:00 juanmafg: can you give us more info there? 18:38:06 Those are implementation choices. What we really need is the SFC Proton that is agnostic of the underlying implementation 18:38:36 networking-sfc defines APIs, and those APIs are supposed to be agnostic of implementation 18:38:58 But now the implementation of networking-sfc APIs are based on NSH (L2) 18:39:28 Niko, can you be more specific? 18:39:36 more info about the different interfaces? 18:40:04 So whatever we propose in SFC Proton, we need to make sure the APIs can be implemented by NSH and VPN 18:40:18 yes, that is right 18:40:44 ok, it sounds like a good idea 18:40:47 I think we should not have a decision on that right now. I personaly don't have enough information for that 18:41:59 I have a talk in Austin regarding 2 different technologies in SFC: https://www.openstack.org/videos/video/service-function-chaining-technology-analysis-and-perspective 18:43:23 So Niko, perhaps when you work with Brady, Juan etc. on SFC Proton, just keep in mind the API agnostic of implementation 18:43:49 yes, that is the intention 18:44:08 Great 18:44:24 Do you have a draft proposal or just start to brainstorm? 18:44:45 I think Juan has 18:45:10 but I guess he will need more time for that. 18:45:16 that's right 18:45:34 I know all the technologies and the interfaces 18:45:42 I think best idea is that we give him a bit of time, filing the rst and then we can comment etc. 18:45:46 Cool. Of course to make it more mature 18:45:53 but still we have not done any comparisson or proposal for having a common API 18:46:01 this is just the first approach to Gluon 18:46:42 Cool. So SFC Proton is a good opportunity to have a common API 18:47:30 yes I think so as well. The main goal should be to get sfc more aligned 18:47:30 Take your time to think through it, and we look forward to see it 18:47:48 That's right, more aligned 18:48:27 It seems we are good 18:49:18 #info Niko will work with Juan, Brady etc to have a common API of SFC that can be implemented by different technologies, such as NSH-based and VPN-based 18:49:45 #info SFC Proton provides this good opportunity of alignment of SFC 18:50:33 #info It will take some time to think through it. Once the draft is ready, it will be submitted for review 18:50:54 Great, we had good discussion. 18:51:04 Anything else from anyone? 18:51:46 #topic AOB 18:51:55 #info Nothing else 18:52:10 If nothing else, we can adjourn the meeting 18:52:17 #info Meeting adjourned 18:52:36 Thank you everyone, and talk to you next time. 18:52:40 Bye 18:52:45 thanks! 18:52:45 Bye All! 18:52:45 bye 18:52:49 #endmeeting