15:00:29 <ad_rien_> #startmeeting massively_distributed_clouds
15:00:30 <openstack> Meeting started Wed Oct 25 15:00:29 2017 UTC and is due to finish in 60 minutes.  The chair is ad_rien_. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:31 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:34 <openstack> The meeting name has been set to 'massively_distributed_clouds'
15:00:49 <parus> Good Morning/Afternoon!
15:00:57 <ad_rien_> #chair parus ad_rien_
15:00:58 <openstack> Current chairs: ad_rien_ parus
15:01:04 <ad_rien_> #topic roll call
15:01:11 <ad_rien_> Please update the etherpad ;)
15:01:30 <ad_rien_> #link https://etherpad.openstack.org/p/massively_distributed_ircmeetings_2017 line 1413
15:01:33 <rcherrueau> o/
15:01:36 <ansmith> o/
15:01:39 <jamemcc_> hello
15:02:28 <ad_rien_> anyone from FBK, Ericsson ?
15:03:10 <ad_rien_> ok
15:03:12 <ad_rien_> seems not
15:03:13 <dancn> yes, FBK, but in another meeting also, partial ttention
15:03:21 <ad_rien_> ok thanks dancn
15:03:32 <ad_rien_> so a few of them are on vacations on Inria too.
15:03:43 <ad_rien_> Actually I should have been off too but…
15:03:55 <ad_rien_> #topic news
15:04:19 <ad_rien_> as usual please go to the pad, read/amend/complete
15:04:29 <ad_rien_> If you have questions or some points you would like to discuss please do it ;)
15:05:44 <ad_rien_> ok let's switch to the next topic so ;)
15:05:53 <ad_rien_> #topic ongoing-actions
15:06:04 <ad_rien_> we have two important points to discuss today
15:06:11 <ad_rien_> (at least from my side):
15:06:19 <ad_rien_> 1./ the F2F meeting
15:06:28 <ad_rien_> 2./ the use-cases' slides
15:06:34 <ad_rien_> I propose to start with 1./
15:07:01 <ad_rien_> as you can see, I already added a couple of links toward the different sessions/presentations that are scheduled for the Sydney meeting.
15:07:24 <ad_rien_> I would like to ask anyone to complete this list with presentations that you believe are relevant for our discussions
15:07:38 <ad_rien_> (i.e. please explain what you/we can expect from the discussion).
15:07:53 <ad_rien_> I will open a dedicated etherpad
15:08:23 <ad_rien_> Regarding the agenda: Can you please add your expectations (i.e. points you would like to discuss).
15:08:34 <jamemcc_> Thanks for the long list of related sessions - I was able to mine a couple of them to add to the Suggested LCOO working group sessions - thanks
15:08:42 <ad_rien_> I remind that I will not attend the summit this time. dpertin and parus will chair the session
15:09:13 <ad_rien_> so parus maybe you would like to highlight some points in particular?
15:09:17 <ad_rien_> regarding the agenda
15:09:42 <parus> do we agreee that the next IRC session will be in late novement?
15:09:50 <parus> November?
15:10:07 <ad_rien_> Nov 22 ?
15:10:10 <ad_rien_> If I'm right
15:10:28 <parus> That works for me.
15:10:58 <parus> Of course, we will create an etherpad for the f2f so anyonne interested can find out what happens.
15:11:12 <jamemcc_> Communication proposal - I'm assuming we will use the ongoing working pad - even for the F2F session - alternative is we would setup a standalone pad just for that session.  In either case we should let the speakersupport@openstack.org knw to change the text description here https://www.openstack.org/summit/sydney-2017/summit-schedule/events/19510/fogedgemassively-distributed-clouds-working-group to include hyperlink to that pad.
15:11:35 <jamemcc_> 22 ok for me
15:11:51 <ad_rien_> @parus done
15:12:15 <ad_rien_> @jamemcc_ good point
15:12:27 <jamemcc_> If the F2F is likely to bemore of a birds of a feather session - people coming and attending who are not usual collaborators - maybe standlone makes more sense
15:13:05 <jamemcc_> FYI - I will not be travelling to Sydney either
15:13:34 <ad_rien_> guys would you please put your name once again on the pad (i.e. the F2F meeting if you attend the summit)
15:14:07 <ad_rien_> ok so that's all from my side right now.
15:14:24 <jamemcc_> I can work with Parus to set it up - not a big debate needed here I think
15:14:39 <ad_rien_> parus:  we can iterate offline to make progress on the agenda. dpertin will also give an overview of the ongoing activities of the WG/SIG/Team
15:14:53 <ad_rien_> great let's move forward so
15:15:14 <parus> OK. I think there was an item about Neutron also that we were planning to address at sidney. Right?
15:15:42 <ad_rien_> parus could you please elaborate a bit more ?
15:15:56 <ad_rien_> we have a new PhD student that should start by december (hopefully)
15:16:13 <ad_rien_> but right now we didn't do anything from our side around Neutron. Did I miss(understand) something
15:16:16 <ad_rien_> ?
15:16:17 <parus> Maybe my memory is fuzzy. but I thought we mentionned something like that in past meeting about cell v2 and neutron.
15:16:25 <ad_rien_> ok
15:16:43 <ad_rien_> there is indeed some discussions about how including similar concepts of Cell V2 in neutron
15:16:50 <ad_rien_> I do not know what are the current progress?
15:17:02 <ad_rien_> that can be one point to put in the agenda indeed
15:17:05 <ad_rien_> good point
15:17:05 <ad_rien_> thanks
15:17:17 <parus> I thought the action was to contact people from Neutron... and sydney might be a good time.
15:17:55 <parus> thanks.
15:18:20 <ad_rien_> #link https://etherpad.openstack.org/p/FEMDC-F2F-meeting-sydney-summit let's start to work on the agenda
15:19:16 <ad_rien_> ok
15:19:25 <ad_rien_> may I suggest to move on the use-cases bullet ?
15:19:31 <parus> yes
15:20:14 <ad_rien_> #topic use-cases' presentation
15:20:25 <ad_rien_> #link https://docs.google.com/presentation/d/1pAAdrzn1kgQCCAffdngiE-NvzsDn8sN4Uj6NFV6NgjQ/edit?usp=sharing
15:20:41 <parus> Thanks for anyone to provide comments to this draft.
15:21:00 <ad_rien_> ok
15:21:06 <ad_rien_> May I propose to spend 20 min on this point
15:21:19 <ad_rien_> I think it is important that we have a common understanding ?
15:21:20 <ad_rien_> Is it ok
15:21:21 <ad_rien_> ?
15:21:25 <parus> +!
15:21:26 <parus> +!
15:21:28 <parus> +1
15:21:40 <ad_rien_> So slide 3, I will prepare the content
15:21:45 <dancn> +1 FBK will add something right after the fog congress next week
15:22:03 <ad_rien_> Slide 4
15:22:12 <ad_rien_> What do you mean by Service Scenarios ?
15:22:15 <ad_rien_> Please parus
15:22:31 <parus> Service Scenarios = VoLTE or uCPE or ....
15:23:03 <parus> some service that telcos offer to entreprise or consumer.
15:23:28 <parus> or some service that an enterprise builds.
15:23:47 <ad_rien_> ok
15:24:07 <ad_rien_> my point is that it is difficult to see the diff between Deployment scenarios vs service scenarios
15:24:15 <ad_rien_> but it seems you know what you are talking about
15:24:16 <ad_rien_> ;)
15:24:20 <ad_rien_> so let's move forward
15:24:37 <parus> Deployment scenario would be cinder on edge , or cinder on central node.
15:24:51 <ad_rien_> whereas services are more about applicaitons, correct?
15:24:54 <parus> yes
15:25:09 <ad_rien_> ok
15:25:10 <parus> Application is a good word. Let's change service for application.
15:25:30 <ad_rien_> ok maybe putting deployment scenarios at that time is a bit sooner
15:25:30 <ad_rien_> ?
15:25:38 <ad_rien_> too early
15:25:38 <ad_rien_> sorry
15:26:04 <jamemcc_> Parus - thanks I see the VM machines are low profile and include non Linux.
15:26:23 <ad_rien_> maybe we can talk about different sandboxing technologies
15:26:35 <ad_rien_> i.e. VM, containers, unikernel
15:26:36 <ad_rien_> ..
15:26:37 <ad_rien_> ?
15:27:02 <parus> I am happy to. serverless too.
15:27:30 <parus> jamecc: does that address your point?
15:28:55 <ad_rien_> Slide 7 ?
15:28:57 <ansmith> Slide 7 - the term 'Edge Node' is not clear to me, the slide refers to relationships/attributes of app, frameworks, premises, etc.
15:28:58 <ad_rien_> can I?
15:29:04 <ad_rien_> yes
15:29:06 <ad_rien_> good point
15:29:23 <ad_rien_> Edge sites vs edge nodes (i.e. site= a micro/nano DC vs nodes = a server
15:29:24 <ad_rien_> )
15:29:47 <ansmith> +1
15:30:02 <ad_rien_> They may be different
15:30:07 <ad_rien_> OpenStack / Kubernetes
15:30:16 <ad_rien_> not sure it should be presented in that way
15:30:31 <parus> What way do you suggest?
15:30:34 <ad_rien_> (once again don't get me wrong, this is your presentation parus, so just some feedbacks from our side)
15:30:36 <ad_rien_> ;)
15:30:52 <parus> I am open to suggestion... can you be more specific?
15:31:04 <ad_rien_> A mCPE vs a C-RAN
15:31:07 <ad_rien_> what do it mean?
15:31:16 <ad_rien_> maybe two slides
15:31:22 <ad_rien_> the first one hardware oriented
15:31:27 <ad_rien_> the second one software oriented
15:31:32 <ad_rien_> i.e. slide 6 hardware oriented
15:31:57 <ad_rien_> i.e. what's make an edge site different than a normal cloud  from the hardware viewpoint/requirements
15:32:13 <ad_rien_> and then Slide 7: what make en edge site different… from the software viewpoint
15:32:31 <ad_rien_> you can run Kubernetes on top of OpenStack (and actually a lot of end-users do that)
15:32:43 <ad_rien_> e.g., the FBK use-case for instance
15:32:47 <parus> OK I get it. This is a good point.
15:32:59 <parus> I think it is three slides:
15:33:19 <parus> one to define edge, one to show hw options, one to show sw options
15:34:29 <ad_rien_> yes
15:34:46 <ad_rien_> I mean how you provide then containers, unikernel, VM, baremetals
15:34:48 <ad_rien_> is another question
15:35:06 <ad_rien_> "They may be in different administrative domains"
15:35:15 <ad_rien_> this is an important assumption
15:35:26 <ad_rien_> just to let you know that we do not take it right now
15:35:33 <ad_rien_> as it makes the global pictures much more difficult
15:35:41 <ad_rien_> (i.e. you need to think about peering agreement ....)
15:36:06 <ad_rien_> in other way, having a global Openstack to supervise/orchestrate a (large) couple of edge sites might be relevant
15:36:22 <ad_rien_> (whatever if you are leveraging a broker or a fully distributed approach)
15:36:38 <ad_rien_> but supervising several systems looks definitely a real challenge
15:36:47 <ad_rien_> I'm not sure even ONAP is targetting that
15:36:52 <ad_rien_> billing, ….
15:36:59 <parus> Are you suggesting I leave this point out?
15:37:03 <ad_rien_> this will become a nightmare
15:37:12 <ad_rien_> I think this point is more relevant to the guys that deal with federation challenge
15:37:13 <parus> We want to focus on things that Openstack people can act on.
15:37:25 <ad_rien_> see for instance the current efforts that are done in NIST
15:37:26 <parus> That is the audience.
15:37:41 <ad_rien_> There is a difference from my viewpoint between federations
15:37:48 <ad_rien_> and edge computing
15:37:56 <ad_rien_> is it clear ?
15:38:34 <ad_rien_> #link https://www.openstack.org/summit/sydney-2017/summit-schedule/events/20475/supporting-general-federation-for-large-scale-collaborations  Federation objective
15:38:44 <parus> it is different... but there are a lot of federating questions around edge that are specific.
15:38:51 <ad_rien_> here you really target to interconnect different domains
15:38:58 <ad_rien_> yes
15:39:14 <ad_rien_> there are some overlaps indeed  but
15:39:23 <ad_rien_> the evil in the details :-P
15:39:37 <ad_rien_> the evil is  ;)
15:39:37 <ad_rien_> sorry
15:39:52 <ad_rien_> we discussed a lot with Craig Lee during the first year of the WG
15:40:04 <ad_rien_> for instance you have to deal with Keystone federation challenges
15:40:21 <parus> I am happy to leave the topic out. I would really like people to walk out of the presentation and think about how edge impact Neutron, how it impact cell v2, how it may impact AMQP ... etc.
15:40:23 <ad_rien_> where are in a edge infrastructure operated by the same telco this question can be non relevant
15:40:30 <ad_rien_> yes
15:40:32 <ad_rien_> +1
15:40:39 <ad_rien_> so maybe you can add that in the software slide
15:40:44 <ad_rien_> (i.e. according to what we discuss before)
15:40:56 <ad_rien_> from the software view point it is different than the federation challenge
15:41:03 <ad_rien_> from the hardware also
15:41:15 <ad_rien_> because it is a single domain (at least in the vision we have/share)
15:41:19 <ad_rien_> butonce again this is your deck ;)
15:41:38 <ad_rien_> so please take it as feedback not an obligation (I do not want to be impolite)
15:42:13 <parus> How do you feel about the question on privacy?
15:42:18 <ad_rien_> I really like the latest slides
15:42:22 <ad_rien_> I think they put the right questions
15:43:01 <ad_rien_> coul dyou please elaborate a bit more ?
15:43:08 <ad_rien_> @rcherrueau is our privacy expert ;)
15:43:09 <parus> 11:42
15:43:18 <ad_rien_> He made his Phd on it.
15:43:27 <parus> even though it might be the same domain.
15:43:40 <parus> One driver of edge, is that private content can stay local.
15:43:45 <ad_rien_> Privacy is something rather large. Could you please precise your point?
15:43:50 <ad_rien_> yes
15:43:55 <parus> This is a new attribute for openstack.
15:43:56 <ad_rien_> indeed
15:43:59 <ad_rien_> yes location
15:44:14 <parus> Should we open that topic with this audience?
15:44:18 <ad_rien_> we should be able to reify the location at the end-users/developers API level
15:44:22 <ad_rien_> yes
15:44:25 <ad_rien_> may make sense
15:45:15 <ad_rien_> Maybe asking the question or at least mentioning sounds good
15:45:25 <parus> I will add that as a bullet.
15:45:37 <ad_rien_> ok
15:45:38 <parus> Maybe on side 14.
15:45:47 <ad_rien_> ok
15:45:51 <ad_rien_> 11:45
15:46:02 <ad_rien_> I think we should close that point
15:46:07 <ad_rien_> We can iterate offlne for sure
15:46:12 <parus> Can you please make sure to mark comments on the presentation? I will review offline.
15:46:27 <ad_rien_> we can iterate by mail. It will be easier
15:46:58 <parus> Great. Thanks for the input.
15:47:00 <ad_rien_> Ok any other comment ?
15:47:23 <parus> did we address jamecc's point?
15:47:54 <ad_rien_> ok I propose to switch to AMQP to have a short overview of what has been done (unless dancn wants to add something regarding the fog congress) ?
15:48:18 <ad_rien_> (jamemcc_ can you double check that your comment(s) is/are in the slide please)
15:48:21 <ad_rien_> ?
15:48:30 <ad_rien_> #topic AMQP
15:48:31 <dancn> no update, we are working on the demo right now... :-)
15:48:36 <ad_rien_> ok thanks
15:48:40 <ad_rien_> I will put it in the pad
15:48:45 <ad_rien_> ansmith:  ?
15:49:14 <ansmith> not much to update, we are scheduling a meeting in early november to discuss testing details
15:49:43 <ad_rien_> ok
15:49:49 <ad_rien_> anything else ?
15:51:39 <ad_rien_> ok
15:51:41 <ad_rien_> thanks
15:51:58 <ad_rien_> #topic open discussions
15:52:13 <ad_rien_> so as I wrote I have to send the offical email for switching to a SIG
15:52:19 <ad_rien_> jamemcc_:  did you already do it for LCCO ?
15:53:33 <ad_rien_> jamemcc_:  are you still there?
15:54:01 <ad_rien_> (by the way this is the only point I have from my side, so please feel free to add any comment/remark if you have some)
15:54:42 <parus> There is a call on white paper starting in 5 min.
15:55:05 <ad_rien_> yes
15:55:15 <ad_rien_> so if there is nothing else. I propose to end the meeting now?
15:55:20 <parus> +1
15:55:26 <ad_rien_> ok
15:55:31 <ad_rien_> so thanks everyone
15:55:43 <ad_rien_> Feel free to complete/amend the pad
15:55:44 <ad_rien_> .
15:55:48 <ad_rien_> #endmeeting