[10:39:28] --- suz has joined
[10:39:36] --- suz has left
[10:53:18] --- ikob has joined
[10:53:39] --- ikob has left
[10:57:44] --- venaas has joined
[10:59:28] --- sleinen has joined
[10:59:41] --- frodek has joined
[11:00:34] --- raj_uoa has joined
[11:04:28] --- suz has joined
[11:09:02] --- ikob has joined
[11:09:24] --- lixia has joined
[11:15:14] --- lixia has left: Replaced by new connection
[11:32:51] --- jschmid has joined
[11:34:45] --- jschmid has left
[11:36:11] --- gvandeve has joined
[11:38:19] --- lixia has joined
[11:38:48] --- lixia has left
[11:39:07] --- gvandeve has left: Logged out
[11:48:10] --- ikob has left
[11:50:44] --- okabe has joined
[11:51:19] --- okabe has left
[12:00:22] --- ThOr101 has joined
[12:05:26] --- ThOr101 has left
[12:06:20] --- ikob has joined
[12:20:22] --- dthaler has joined
[12:22:00] --- dthaler has left: Replaced by new connection
[12:23:46] --- dthaler has joined
[12:24:06] <dthaler> anyone understand what problem this is solving and can explain?
[12:26:55] <dthaler> ok why does this involve BGP at all, as opposed to pure PIM?
[12:28:33] <venaas> I think provider of core wants to have no state in their routers. they do however run bgp and don't want additional protocols
[12:28:55] <dthaler> they want multicast with no state?
[12:29:03] <venaas> using bgp is rather sick imo
[12:29:05] <dthaler> doesn't that mean broadcast?
[12:29:56] <venaas> not sure if I understand this, but I thought idea was to signal between edge routers, and that they don't need state in the core itself. maybe I'm wrong
[12:30:20] <dthaler> how do you construct a multicast forwarding tree then with no state
[12:33:48] <venaas> hmm ok, if idea is to actually do multicast in the core, and not encapsulate packets and unicast them between the PE routers, then you are right
[12:34:17] <dthaler> I still don't get why BGP is involved at all
[12:35:21] <venaas> I suppose mainly because they feel comfortable with BGP. Also PIM doesn't quite do what they want, but better adapt PIM than try to make it work with BGP
[12:35:54] <dthaler> well BGMP was meant for those familiar with BGP but no one wants that either.
[12:39:22] <venaas> it's mostly a BGP community that wants to do multicast and thinks PIM is complicated I think. They probably haven't yet realized that using BGP will end up at least as complicated once doing the details. Oh well, I don't really understand this more than you (:
[12:39:24] <dthaler> toerless is right, the advantage of a separate TCP session is that your high-frequency multicast state updates don't block high-priority route convergence updates
[12:40:09] <venaas> Not sure if it's good idea to have possibly bursty PIM signalling over a tcp session
[12:41:22] <dthaler> agree (same problem as in MSDP)
[12:41:49] <venaas> seen proposal for doing SA signalling with BGP as well
[12:44:03] <venaas> believe this will be discussed in l3vpn tomorrow, should be interesting
[12:47:32] --- gregory has joined
[12:49:07] <gregory> i guess there is no jabber scribe today? Just side comments?
[12:49:10] --- psavola has joined
[12:49:32] <gregory> (stuck in another wg, but would like to keep tabs on mboned)
[12:50:41] <venaas> almost done now
[12:51:05] <venaas> experience with multicast vpn is next up I think
[12:51:19] <gregory> ok
[12:51:47] <venaas> for 5 min. then 5 min on mib and that's all (according to agenda)
[12:52:36] <psavola> the ad will probably have a few choice words afterwards, or so DaveM warned
[12:53:44] <gregory> anyone care to scribe the VPN discussion for me?
[12:54:25] <dthaler> objectives slide up now
[12:55:32] <dthaler> what's MDT?
[12:55:46] <dthaler> current deployement is GRE encapsulation, PIM protocol, MDT SAFI
[12:56:16] <gregory> anyone know where the preso is posted? It's not at "prelim & interim materials" page?
[12:57:14] <dthaler> don't know
[12:57:36] <dthaler> hard to hear him, but I think he said MDT aggregates multiple VPN trees into one underlying multicast tree
[13:00:04] <venaas> btw, the previous presentation was Label Switched Multicast Solutions, which was pretty much the same as in IDR yesterday I think
[13:00:34] <dthaler> I-D being presented now is draft-ycai-mvpn-experience-00.txt
[13:01:41] <dthaler> MVPN reference is apparently draft-ietf-l3vpn-2547bis-mcast-00.txt
[13:04:00] <dthaler> some comment from John Meylor which I didn't follow
[13:04:10] <dthaler> now on to next presentation
[13:04:41] <dthaler> (no discussion after previous except Meylor's comment)
[13:05:04] <dthaler> now up draft-mcwalter-ip-mcast-mib-00
[13:10:09] <dthaler> discussion on where this should be done
[13:10:25] <gregory> hmm. I would have expected more interest, comment, discussion on the live experience w/ mvpn. Hmmm.
[13:10:47] <dthaler> I think there wasn't enough info for this WG to understand it
[13:10:49] <gregory> Do we think that perhaps it is so underused that in reality, people don't have much to say?
[13:10:57] --- raj_uoa has left
[13:11:15] <psavola> does any operator in the room do or think of doing mvpn's ? :)
[13:11:20] <dthaler> Kessens: dave meyer wants to step down as WG chair... should we close the WG?
[13:11:39] <gregory> Whoa
[13:11:49] <dthaler> if it stays open, he prefers chairs that are operators running real multicast networks
[13:12:09] <gregory> thanks for scribing dt. keep it up here!!
[13:12:43] <dthaler> meylor, dino favor keeping wg open otherwise multicast status will worsen
[13:13:52] <gregory> look like any operators are stepping up?
[13:14:12] <dthaler> not yet but there's people at the mike line
[13:15:16] <dthaler> joel: so far WG has been addressing problems we thought users would have 10 years ago as opposed to problems they have now
[13:16:06] <dthaler> meylor: wants notes to reflect thanks to dave meyer for all the work done so far
[13:16:39] <dthaler> eric rosen(?): first 45 minutes of meeting showed very few people interested in drafts here
[13:17:11] <dthaler> meyer: actually that was a bout expired drafts, so not surprising, not indicative
[13:18:25] <dthaler> Lenny Juliano (sp?) : people want to see it succeed, but deployments are not growing, need to understand why
[13:20:00] <dthaler> meylor: we are seeing active multicast deployment, they're just more walled garden approaches today
[13:20:19] <dthaler> kessens: this isn't a walled garden task force
[13:20:45] <dthaler> peanut gallery: so we have the L2VPN WG?
[13:21:00] <dthaler> typo L3VPN
[13:21:37] --- gregory has left: Replaced by new connection
[13:21:52] --- gregory has joined
[13:22:46] <gregory> Lenny Giuliano: Juniper consulting systems engineer who specializes on mcast. he eats and breathes the stuff daily w/ the customers deploying mcast.
[13:23:48] <dthaler> kessens: general sense seems to be to continue WG
[13:24:42] <dthaler> kessens: should we do charter review on the list or at a BOF?
[13:24:47] <gregory> perhaps it makes sense to hold a workshop with carriers, some of their enterprise & edu customers, and some key IETFers to see what people are using, what they want to use, and what are the barriers?
[13:25:22] <dthaler> agree, there was one maybe 5-6 years ago but not since the IP Multicast Initiative/forum went away
[13:25:27] <gregory> perhaps we could include an IAB sponsor in the workshop?
[13:26:04] <gregory> how many people in the room,?
[13:26:17] <dthaler> my guess is around 60?
[13:26:44] <dthaler> more like 80
[13:28:06] <dthaler> kessens: will require at least one chair to be an operator running a mcast network
[13:28:07] --- newcat has joined
[13:28:35] <dthaler> about 20 people in room just raised hand about consider themselves operators
[13:28:52] <dthaler> kessens is asking for nominations
[13:29:07] <dthaler> (not here)
[13:29:26] <dthaler> all done now
[13:29:36] --- venaas has left: Logged out
[13:29:39] --- newcat has left
[13:29:57] <dthaler> ok maybe not all done now
[13:30:54] <gregory> did anyone get nominated?
[13:30:54] <dthaler> ok now all done
[13:31:10] <gregory> Did they ask to see hands of anyone that would consider taking the job?
[13:31:12] <dthaler> no, nominations should be done to kessens (not during meeting)
[13:31:14] --- suz has left
[13:31:32] <gregory> ok. Thanks a ton for scribing. Very helpful.
[13:31:50] <dthaler> no problem
[13:32:11] --- dthaler has left
[13:32:29] --- gregory has left
[13:33:07] --- ikob has left
[13:47:25] --- frodek has left: Disconnected
[13:48:39] --- sleinen has left: Disconnected
[13:57:19] --- psavola has left: Disconnected
[15:27:14] --- gregory has joined
[15:27:22] --- gregory has left