15:02:27 <ad_ri3n_> #startmeeting fog-edge-massively-distributed-clouds
15:02:28 <openstack> Meeting started Wed Feb 14 15:02:27 2018 UTC and is due to finish in 60 minutes.  The chair is ad_ri3n_. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:02:32 <openstack> The meeting name has been set to 'fog_edge_massively_distributed_clouds'
15:02:33 <ad_ri3n_> #chair  ad_ri3n_
15:02:34 <openstack> Current chairs: ad_ri3n_
15:02:39 <ad_ri3n_> #chair parus
15:02:41 <openstack> Current chairs: ad_ri3n_ parus
15:02:45 <ad_ri3n_> #topic roll call
15:02:56 <ad_ri3n_> so let's see who join today
15:02:56 <ad_ri3n_> ;)
15:03:08 <parus> Please remind me of the etherpad address!
15:03:11 <kgiusti> o/
15:03:23 <ad_ri3n_> #info agenda
15:03:24 <avankemp> o/
15:03:28 <ansmith> o/
15:03:35 <ansmith> parus, https://etherpad.openstack.org/p/massively_distributed_ircmeetings_2018
15:03:38 <ad_ri3n_> #link https://etherpad.openstack.org/p/massively_distributed_ircmeetings_2018 (line 202)
15:04:19 <ad_ri3n_> So B-yond / Redhat / Inria
15:04:27 <ad_ri3n_> anyone from the DragonFlow team?
15:05:12 <ad_ri3n_> #topic announcements
15:05:31 <ad_ri3n_> so a few news from my side mainly related to the Vancouver summit
15:05:52 <ad_ri3n_> the presentations that have been submitted under the umbrella of the FEMDC SiG
15:06:00 <ad_ri3n_> are the ones I put in the etherpad
15:06:09 <ad_ri3n_> if you are aware about additional ones please complete the pad ;)
15:06:37 <parus> I have submitted three abstracts. I will add to the pad.
15:07:01 <ad_ri3n_> ok cool
15:07:05 <parus> I must have missed the part where I had to refer to FEMDC
15:07:13 <ad_ri3n_> that's ok
15:07:47 <ad_ri3n_> The idea was just to add something like: '' this work is related to the FEMDC activities…" or something similar
15:08:00 <ad_ri3n_> but please parus complete the pad.
15:08:12 <ad_ri3n_> so that's all
15:08:14 <ad_ri3n_> from my side
15:08:43 <ad_ri3n_> (I'm going to discuss about what we are doing in the edge sessions but I think we can now consider it as an on going actions as it is redundant every weeks)
15:08:59 <ad_ri3n_> so if nothing more from your side guys let's go on.
15:09:00 <ad_ri3n_> ?
15:09:23 <ad_ri3n_> #topic ongoing-actions edge sessions
15:09:55 <ad_ri3n_> So as you probably alreadly know there is a weekly meeting supervised by that OpenStack foundation every tuesday at 15. UTC
15:10:43 <msimonin> o/
15:10:48 <ad_ri3n_> Currently we are mainly discussing ''state of the art'' solutions  that enable administrators/end users to deal with multiple regions deployment scenarios
15:11:25 <ad_ri3n_> I put some links but roughly yesterday we discussed about oaktree/shade and Tricircle before splitting
15:12:07 <ad_ri3n_> We also briefly went through the gap analysis we are consolidating with Ericsson folks (BTW, thanks parus for your comments, I started to answer a few of them but I saw this morning you added more ones).
15:12:23 <ad_ri3n_> so I don't know whether you would like to have more details or not.
15:12:34 <ad_ri3n_> if not we can move to the next topic ?
15:12:39 <parus> It is a great analysis. I want to congratulate you on it.
15:12:54 <ad_ri3n_> it is clearly not completed but at least we started finally ;)
15:13:26 <parus> I struggled a bit with terminology (e.g. VMI)... but it made sense in the end.
15:13:45 <ad_ri3n_> BTW kgiusti ansmith do you have news regarding possible interests of other RedHat folks on the edge challenges?
15:14:05 <parus> The main comment is about defining the node where "you are". did you address that?
15:14:20 <ad_ri3n_> it would be great to get feedbacks on use-cases or expectations related to administrators/developpers requirements
15:14:34 <ad_ri3n_> parus:  not yet, gonna look
15:15:39 <kgiusti> ad_ri3n_: what's the meeting details - I'll see if I can get anyone from the openstack team to attend
15:16:09 <ad_ri3n_> #link https://etherpad.openstack.org/p/edge-gap-analysis
15:16:23 <kgiusti> ad_ri3n_: thanks, ansmith ^^^
15:16:25 <ad_ri3n_> ^ this is the link related to the gap analysis we are doing
15:16:30 <ad_ri3n_> and the general link
15:16:32 <ad_ri3n_> is
15:16:51 * ad_ri3n_ is looking for another etherpad link … too many links … definitely
15:17:16 <ad_ri3n_> #link https://etherpad.openstack.org/p/2017_edge_computing_working_sessions general link for all edge working sessions
15:17:34 <ad_ri3n_> ^ that one should be the good entrypoint ansmith kgiusti
15:17:44 <kgiusti> ad_ri3n_: excellent, thanks
15:17:49 <ad_ri3n_> Once again, it would be really great to have more folks of redhat involved in our action
15:17:59 <ad_ri3n_> We know that in France RedHat has some interests
15:18:35 <kgiusti> ad_ri3n_: if you have names of the folks in france let me/ansmith know
15:18:38 <ad_ri3n_> it would be great to have additional persons to help us to move forward
15:18:43 <ad_ri3n_> faster and in the right direction ;)
15:18:48 <parus> ad_ri3n_: are you still having meetings with ericsson to develop the gap analysis further?
15:18:50 <ad_ri3n_> @msimonin did you remind?
15:18:54 <kgiusti> ad_ri3n_: agreed
15:18:59 <ad_ri3n_> parus:  yes next friday
15:19:09 <ad_ri3n_> 14- > 17
15:19:15 <ad_ri3n_> Paris Time
15:19:38 <parus> those are face to face . Right?
15:19:46 <ad_ri3n_> no
15:19:56 <ad_ri3n_> professional skype
15:20:03 <ad_ri3n_> one is in US
15:20:11 <ad_ri3n_> the other one is in Sweden
15:20:17 <ad_ri3n_> and we are in Rennes and Nantes from our side
15:20:53 <parus> I am happy to join if that helps.
15:21:36 <ad_ri3n_> I will ask Joao to invite you
15:21:45 <ad_ri3n_> but the discussion right now is more on a prototype
15:21:51 <ad_ri3n_> i.e. ok there is no solution right now
15:22:00 <ad_ri3n_> what can be the solution with the goal of mitigating the effort
15:22:54 <ad_ri3n_> kgiusti:  ansmith
15:23:01 <ad_ri3n_> Bertrand Rault
15:23:16 <ad_ri3n_> is the one that validates the involvement of RedHat within the Discovery initiative
15:23:27 <ad_ri3n_> #link https://beyondtheclouds.github.io
15:23:32 <ad_ri3n_> see bottom of the page
15:23:47 <ad_ri3n_> ok
15:24:12 <kgiusti> ad_ri3n_: thanks
15:24:21 <ad_ri3n_> It would be great if we can attract more people
15:24:46 <ad_ri3n_> interested by other challenges such as : how sharing VM Images across distinct sites in an efficient manner.
15:24:58 <ad_ri3n_> Bertrand told us that RedHat is also interested by these challenges
15:25:25 <kgiusti> ad_ri3n_: I'm surprised RH doesn't have more folks at the table for this.
15:25:56 <ad_ri3n_> Me too  :-P
15:26:07 <kgiusti> ad_ri3n_: Bertrand is part of the Openstack team,  I'll reach out.
15:26:11 <ad_ri3n_> so I ask the ones I know (i.e. probably the best ones ;))
15:26:26 <ad_ri3n_> who are you guys ;)
15:26:45 <kgiusti> ad_ri3n_: the best you know in *Enterprise Messaging*, which is a separate group for Openstack, unfortunately :P
15:26:45 <ad_ri3n_> ok  perfect moment to switch to the AMQP update ;)
15:27:00 <ad_ri3n_> :-P
15:27:07 <parus> Before you do
15:27:12 <ad_ri3n_> parus:  please ?
15:27:19 <parus> you may want to mention white paper
15:27:24 <parus> #link https://www.openstack.org/edge-computing/
15:27:52 <ad_ri3n_> excellent
15:27:52 <parus> just saw it on the web site
15:27:53 <ad_ri3n_> Thanks parus
15:27:54 <ad_ri3n_> ;)
15:28:52 <ad_ri3n_> Ok I invite all of you to give a look
15:29:17 <ad_ri3n_> This is the output of a couple of exchanges/telcos/debates to deliver some foundations for our future discussions
15:30:12 <ad_ri3n_> a short paper but that defines at least the limit of what we call edge computing infrastructures.
15:30:24 <ad_ri3n_> Ok so let's switch to the next topic
15:30:36 <ad_ri3n_> #topic ongoing-actions amqp evaluations
15:30:42 <kgiusti> I'm still working on a dragonflow AMQP 1.0 driver POC
15:30:47 <ad_ri3n_> ok
15:31:17 <kgiusti> and I'll have a presentation regarding extending the oslo api to better suite such use cases at PTG
15:31:30 <kgiusti> that's it from me - ansmith? msimonin?
15:31:42 <ansmith> nothing to add
15:31:46 <msimonin> I think I'll be at the oslo session  :)
15:32:02 <ad_ri3n_> +1
15:32:09 <msimonin> From our side I've put some notes on the e-pad
15:32:12 <kgiusti> msimonin: if you bring something to throw, try to make it *soft*, like rotten fruit....
15:32:13 <ad_ri3n_> parus:  finally you will come to the PTG event?
15:32:21 <ad_ri3n_> :-)
15:32:29 <msimonin> msimonin: melted cheese ?
15:32:36 <msimonin> kgiusti:*
15:32:49 <parus> I am making travel arrangements to be at PTG. Arriving monday morning and leaving on wednesday.
15:33:10 <parus> Looking forward to see all of you!
15:33:20 <msimonin> same here parus !
15:33:48 <msimonin> So, to make it short regarding AMQP
15:34:02 <msimonin> Javier is making a lot of progress on the framework code base !
15:34:42 <msimonin> We are now able to automate the tests of the major oslo messaging patterns
15:35:14 <msimonin> I've make some initial tests to scale to 5-10K agents
15:35:17 <msimonin> made*
15:35:45 <msimonin> This is very promisinfg
15:35:49 <msimonin> promising*
15:36:38 <msimonin> that's all from my side
15:36:47 <ad_ri3n_> ok thanks for the update
15:36:55 <ad_ri3n_> I looking forward to seeing your presentation guys ;)
15:37:18 <msimonin> Oh yes we submitted a presentation with kgiusti and Javier
15:38:03 <kgiusti> bring grilled cheese.... :)
15:38:11 <ad_ri3n_> :-)
15:38:22 <ad_ri3n_> BTW is Javier attending the meeting today ?
15:39:00 <msimonin> kgiusti: don't dare me
15:39:04 <msimonin> kgiusti:  :)
15:39:12 <msimonin> yes Javier is sitting next to me
15:40:14 <ad_ri3n_> ok
15:40:18 <ad_ri3n_> Javier welcome
15:40:32 <ad_ri3n_> I should definitely arrange a telco with you ;)
15:40:35 <ad_ri3n_> Sorry for that
15:40:52 <ad_ri3n_> Ok so we can move to the next topic which is Open discussion
15:40:59 <ad_ri3n_> #topic open-discussions
15:41:14 <parus> I have a question regarding PTG
15:41:19 <ad_ri3n_> Please ?
15:41:29 <parus> Do we have an agenda on how we are going to use our time?
15:41:38 <parus> for this team.
15:41:43 * msimonin must leave, thanks for the meeting !
15:41:59 <parus> Are we going to dive into the Vancouver presentations?
15:42:24 <ad_ri3n_> partially
15:42:54 <ad_ri3n_> there is a dedicated etherpad to prepare our discussion:
15:42:57 * ad_ri3n_ is looking for
15:43:19 <ad_ri3n_> #link https://etherpad.openstack.org/p/edge-ptg-dublin
15:43:42 <ad_ri3n_> on monday afternoon we will finalize the agenda for tuesday morning
15:43:58 <ad_ri3n_> the goal of the discussion on tuesday morning is to discuss with core developpers of other projecs.
15:44:57 <parus> Which other projects are we targeting? Neutron?
15:45:03 <ad_ri3n_> so parus the more materials we are going to put on this pad
15:45:06 <ad_ri3n_> the better it will be
15:45:16 <ad_ri3n_> K N G N
15:45:23 <ad_ri3n_> Keystone Nova Glance Neutron
15:45:33 <parus> OK! I will make this a priority this week.
15:45:34 <ad_ri3n_> these are the ones you must have to be able to launch VMs
15:45:52 <parus> I think you might need horizon and heat as well.
15:45:53 <ad_ri3n_> last week we discussed a lot of keystone
15:46:03 <ad_ri3n_> Horizon is a WEB server
15:46:13 <ad_ri3n_> but …
15:46:20 <ad_ri3n_> yes why not actually
15:46:31 <ad_ri3n_> Heat is also a plus
15:46:43 <parus> Changes are pretty simple but how will information be presented on Horizon if you have 1000 nodes.
15:46:44 <ad_ri3n_> i.e if you look to L1/L2/L3/...
15:46:51 <ad_ri3n_> Heat comes in the loop later
15:47:01 <ad_ri3n_> parus:  the problem is already present
15:47:03 <ad_ri3n_> in a large DC
15:47:11 <ad_ri3n_> so
15:47:14 <ad_ri3n_> I think you are right
15:47:18 <ad_ri3n_> there will be specific questions
15:47:27 <ad_ri3n_> my point is that Horizon is stateless
15:47:29 <parus> To some degree... maybe what has been done for AZ can be done for edge.
15:47:39 <ad_ri3n_> in the sense that it is just an entry point to the different APIs
15:47:55 <ad_ri3n_> the hard part right now
15:48:07 <ad_ri3n_> is related to the DBs, the communication bus
15:48:24 <ad_ri3n_> and the possibility to be able to call a remote service
15:48:46 <ad_ri3n_> (i.e. Nova from one site wants to contact a remote Glance for collecting the VM image for instance that is needed to boot the VM)
15:48:50 <ad_ri3n_> how can you do that
15:48:55 <ad_ri3n_> if you have indepent openstand
15:49:14 <parus> I don't disagree. putting in place the structure that allows the performance to be achieved is important.
15:49:21 <ad_ri3n_> openstack. Nova from site A should be authentified on site B to be able to contact Glance on B
15:49:41 <ad_ri3n_> Actually there is another action that would be interesting to do
15:49:46 <ad_ri3n_> (maybe for the last cycle)
15:49:59 <parus> But putting in place the data models and the abstractions to have a useful interaction with the system should not be understimated.
15:50:05 <ad_ri3n_> it is an analysis of what is behind the scene of ONAP
15:50:16 <ad_ri3n_> parus:  fully agree
15:50:47 <ad_ri3n_> We have one Phd and two postdoctoral researchers at Inria that investigates such challenges (but from the academic viewpoint)
15:51:06 <ad_ri3n_> what is difficult is to make the link between academic PoCs and all the opensource solutions that exist
15:51:32 <ad_ri3n_> I think there is also orchestration engines/services within ONAP
15:51:44 <parus> Yes there is.
15:51:51 <parus> a Master orchestrator
15:51:54 <ad_ri3n_> what is the added value w.r.t. HEAT for instance is something I would be interested to better understand
15:52:21 <ad_ri3n_> (actually that is something that should be valuable for our community to understand not only for me ;) sorry  :-P )
15:52:29 <parus> Let me take an action here... and maybe I can come back with some input.
15:52:49 <parus> Or if someone wants to work on this with me... that would be even better.
15:53:12 <ad_ri3n_> @dpertin
15:53:12 <ad_ri3n_> ?
15:53:14 <ad_ri3n_> are you there
15:53:16 <ad_ri3n_> seems not
15:53:30 <ad_ri3n_> Ok we have one engineer that will work on that for the next 3 months
15:53:33 <ad_ri3n_> Dimitri Pertin
15:53:41 <ad_ri3n_> he chaired the sydney meeting with you
15:53:43 <ad_ri3n_> in Sydney
15:53:47 <ad_ri3n_> so you know him
15:53:52 <parus> I do1
15:53:58 <parus> I do !
15:54:09 <parus> I will reach out.
15:55:10 <ad_ri3n_> ok
15:55:18 <ad_ri3n_> anything else you would like to discuss guys?
15:55:47 <parus> I would be afraid to get another action !
15:56:00 <ad_ri3n_> ;-)
15:56:16 <ad_ri3n_> ansmith:  kgiusti: I just added one point regarding a possible ping from your side
15:56:28 <ad_ri3n_> in order to attract additional redhat folks (i.e. from the OpenStack group ;))
15:56:34 <ad_ri3n_> so that's all from my side
15:56:39 <ad_ri3n_> Thanks you for attending the meeting
15:56:40 <kgiusti> +1
15:56:51 <ad_ri3n_> and see you in Dublin ;)
15:57:19 <ad_ri3n_> #endmeeting