IETF
intarea
intarea@jabber.ietf.org
Tuesday, November 5, 2013< ^ >
marka has set the subject to: INTAREA: Where to Go for Discussion
Room Configuration
Room Occupants

GMT+0
[00:00:11] <Wes George> @sarikaya - can you relay mic comments on this presentation (if any) to the jabber room? I'm in another session and not on audio
[00:00:15] <sarikaya2012> motination: design low-cost unified approach to transition
[00:00:29] <sureshk> I can do that Wes
[00:00:42] <sarikaya2012> openv6 in a nutshell - 1
[00:01:29] <sarikaya2012> openv6 in a nutshell -2
[00:02:03] <sarikaya2012> @Wes, sure
[00:02:21] <sarikaya2012> China telecom use case
[00:03:16] <sarikaya2012> new protocol work
[00:04:03] <sarikaya2012> Deployments
[00:04:42] <sarikaya2012> Bits-N-Bytes
[00:04:46] <sarikaya2012> summary
[00:04:59] <sarikaya2012> Alain Durand on mic
[00:05:38] <sureshk> Alain Durand wants to know how the device gets configured
[00:05:40] <sarikaya2012> device dynamically learn config or push it to devices?
[00:06:01] <sureshk> Guillaume mentions that the configuration is flow based, and not netconf based
[00:06:46] Dave Thaler joins the room
[00:06:48] narten leaves the room
[00:07:04] <sureshk> Alain mentions that there are two different aspects here
[00:07:18] <sureshk> one is the static part of the config and the other is the dynamic flow part
[00:07:43] <sarikaya2012> Ralph, what are you advocating?
[00:07:52] <sureshk> Ralph wants to know what aspects of the solution scales with the number of flows rather than the number of CPEs
[00:07:55] <sarikaya2012> Jim, it is flow setup
[00:08:28] <sarikaya2012> Andrew on mic
[00:08:51] <sureshk> Alain prefers having one technology per device
[00:09:15] <sureshk> Guillaume mentions that is more applicable to transition between technologies
[00:09:35] <sureshk> Alain mentions then the multiple technology case is not concurrent
[00:09:59] <sarikaya2012> Yong on mic
[00:10:12] <sureshk> Yong wants to know what is the definition of flow according to the authors
[00:10:16] Jinmei Tatuya joins the room
[00:10:35] Lee Howard leaves the room
[00:10:38] <sureshk> Guillaume mentions that it has not solidified yet
[00:10:47] <sureshk> but is a combination of info in L2-L4
[00:10:52] <sarikaya2012> Jim, flows not yet defined
[00:11:16] <sureshk> Erik Kline mentions that this work requires a rollout of a new CPE.
[00:11:36] <sarikaya2012> I think the speaker's name is Guillaume
[00:12:03] <sureshk> @behcet: yes
[00:12:06] narten joins the room
[00:12:51] <sarikaya2012> Tina on mic
[00:13:19] Joe Hildebrand leaves the room
[00:13:31] <sarikaya2012> Ted on DHCP config applications
[00:13:34] <sureshk> Tina Tsou mentioned that she got some info from Qiong Sun from CT about an interop test between two vendors
[00:13:55] <sureshk> regarding openv6 implementations
[00:14:04] <Wes George> you can decide whether to relay this to mic or not, but my initial reaction to this draft was to say "so the problem with IPv6 transition is that there are too many transition technologies, so we propose making *another* one to make it easier. does this seem to anyone else like it'd be WAAAY easier to simply bypass the transition technologies and just deploy IPv6?"
[00:14:36] Joe Hildebrand joins the room
[00:14:54] <sureshk> @Wes: the proposal is not up for adoption here and the presenter left the room in a hurry
[00:15:01] <Andrew Sullivan> @Wes: I believe there is an xkcd for this :)
[00:15:02] <sureshk> They will be at the Bits n Bytes
[00:15:13] <sarikaya2012> @Wes, we have Ted now speaking
[00:15:33] <sureshk> @Andy: it is on my speed dial
[00:15:34] <sureshk> http://xkcd.com/927/
[00:16:04] <Wes George> heck with xkcd, this is an xzibit meme - yo dawg, I heard you like transition technologies, so I made you a transition technology that lets you transition while you transition
[00:16:07] <Wes George> thanks :-)
[00:17:46] <sarikaya2012> @Wes, I think we need to put any "blame" on China telecom O:)
[00:18:50] <Wes George> perhaps, but CT still has enablers.
[00:20:51] Brian Carpenter leaves the room
[00:21:19] narten leaves the room
[00:21:21] Brian Carpenter joins the room
[00:22:29] idiliod leaves the room
[00:23:10] Brian Carpenter leaves the room: Away
[00:23:22] Joe Hildebrand leaves the room
[00:23:40] <sarikaya2012> Stuart at mic
[00:26:18] <sarikaya2012> big queues at the mic
[00:26:55] <sarikaya2012> @suresh, pls take over for a few minutes
[00:31:08] narten joins the room
[00:31:13] Brian Carpenter joins the room
[00:33:23] <sarikaya2012> Dave and Andrew Sullivan at mic
[00:33:41] Joe Hildebrand joins the room
[00:35:54] Joe Hildebrand leaves the room
[00:37:14] <sarikaya2012> again a big lineup
[00:37:59] <SM> split-view
[00:38:54] <sarikaya2012> Dave Thaler at mic
[00:39:00] Wes George leaves the room
[00:41:25] <sarikaya2012> Mikael at mic
[00:41:53] Abhay Roy joins the room
[00:43:54] <Andrew Sullivan> I quite agree with Dave, actually.  The main thing I was trying to say is that the FQDN could well give you better results than IPs
[00:44:38] <Dave Thaler> I figured that, I was disagreeing only with the part where you were guessing what Ted was trying to say, rather than the part about what you were trying to say :)
[00:44:40] <Andrew Sullivan> and either you think that options are ok (in which case you need to explain what trade off you're making) or else you think DHCP options are wrong
[00:44:43] <SM> Deployment-wise, FQDNs are easier.
[00:44:44] <Andrew Sullivan> yes
[00:44:48] Abhay Roy leaves the room
[00:44:50] Abhay Roy joins the room
[00:44:51] narten leaves the room
[00:44:59] <Andrew Sullivan> It seems to me that Ted's position is _actually_ that options are evil and wrong
[00:45:16] <Dave Thaler> maybe but I would disagree with that myself
[00:45:21] <Andrew Sullivan> that's fine, but there is no path to the conclusion "so use IP addresses" from "options are evil and wrong"
[00:46:06] <Dave Thaler> perhaps so, we haven't gotten to the slides about fqdn vs ip
[00:46:14] <Andrew Sullivan> there's that :)
[00:47:30] <Dave Thaler> btw if you haven't readyit, the dhc-topo-conf draft says the DHCP server resolves the DHCP server itself at the time of the DHCP request and hands the resulting IP's to the client in the DHCP response, so the option carries IP addresses... the DNS query is just done by the DHCP server rather than the client.
[00:47:48] narten joins the room
[00:48:08] <sarikaya2012> Ted finally continues
[00:48:38] <Andrew Sullivan> Right, and the point is that if you got back two A records in that case, you'd have two IP addresses
[00:48:45] <Dave Thaler> yes
[00:49:10] Lee Howard joins the room
[00:49:18] narten leaves the room
[00:50:09] <sarikaya2012> Ted: DHCP can give both, IP and FQDN
[00:50:31] <sarikaya2012> Client should decide which one to use
[00:50:41] Joe Hildebrand joins the room
[00:51:22] <sarikaya2012> Ted: Way forward, dhc should decide
[00:52:37] Abhay Roy leaves the room
[00:53:12] Lee Howard leaves the room
[00:53:30] narten joins the room
[00:53:57] mike joins the room
[00:56:07] mikemlb joins the room
[00:59:54] mike leaves the room
[01:01:11] narten leaves the room
[01:02:19] Guillaume Leclanche joins the room
[01:07:03] <Guillaume Leclanche> sureshk: I did not leave the room in a hurry, I just happen to be seating at the back :)
[01:07:31] <Andrew Sullivan> You don't actually need a new way to do this
[01:07:44] <Andrew Sullivan> you use the (ip address) config from DHCP to get your DNS server
[01:07:56] <Andrew Sullivan> then you look everything up in the DNS as srv records
[01:08:10] <Andrew Sullivan> using, in effect, dns-sd
[01:09:59] <Dave Thaler> ... which is protected via DNSSec
[01:11:02] <Andrew Sullivan> or anyway, can be
[01:11:18] <Andrew Sullivan> +1000 to what Joe is saying
[01:12:23] Jinmei Tatuya leaves the room
[01:12:24] <sarikaya2012> Suresh, thanks for all the discussions
[01:12:24] Joe Hildebrand leaves the room
[01:12:29] <sarikaya2012> meeting adjurned
[01:12:44] SM leaves the room
[01:13:05] Guillaume Leclanche leaves the room
[01:14:26] pebersman leaves the room
[01:15:58] Andrew Sullivan leaves the room
[01:17:02] Andrew Yourtchenko leaves the room
[01:24:56] sureshk leaves the room
[01:28:56] Brian Carpenter leaves the room
[01:29:56] Dave Thaler leaves the room
[01:30:14] Dave Thaler joins the room
[01:33:18] Guillaume Leclanche joins the room
[01:33:28] Lee Howard joins the room
[01:34:35] Andrew Yourtchenko joins the room
[01:35:27] Guillaume Leclanche leaves the room
[01:35:59] neilkatin leaves the room
[01:36:41] Jinmei Tatuya joins the room
[01:40:23] Ole Troan leaves the room
[01:41:31] Brian Carpenter joins the room
[01:42:00] Jinmei Tatuya leaves the room
[01:44:57] Joe Hildebrand joins the room
[01:45:31] Joe Hildebrand leaves the room
[01:46:17] narten joins the room
[01:47:44] Brian Carpenter leaves the room: offline
[01:56:33] narten joins the room
[01:57:04] Jinmei Tatuya joins the room
[01:58:07] Andrew Yourtchenko leaves the room
[01:58:10] narten leaves the room
[01:58:22] Jinmei Tatuya leaves the room
[01:59:04] narten leaves the room
[02:22:25] sureshk joins the room
[02:43:39] sureshk leaves the room
[03:05:13] sarikaya2012 leaves the room
[03:32:36] Lee Howard leaves the room
[03:53:05] Dave Thaler leaves the room
[04:10:30] Dave Thaler joins the room
[04:10:36] Dave Thaler leaves the room
[16:03:23] mikemlb leaves the room