IETF
core@jabber.ietf.org
Friday, 9 November 2012< ^ >
stpeter has set the subject to: CORE WG | minutes at http://tools.ietf.org/wg/core/minutes | slides at https://datatracker.ietf.org/meeting/83/materials.html#wg-core
Room Configuration

GMT+0
[00:42:12] Michael Richardson leaves the room
[00:42:31] Michael Richardson joins the room
[01:49:24] Michael Richardson leaves the room
[01:49:30] Michael Richardson joins the room
[01:49:59] Michael Richardson leaves the room
[01:50:04] Michael Richardson joins the room
[01:51:45] Michael Richardson leaves the room
[02:07:44] Michael Richardson joins the room
[02:52:02] Michael Richardson leaves the room
[02:53:05] Michael Richardson joins the room
[03:00:13] Michael Richardson leaves the room
[03:01:05] Michael Richardson joins the room
[12:44:49] Michael Richardson leaves the room
[12:45:06] Michael Richardson joins the room
[12:48:07] Michael Richardson leaves the room
[12:59:10] Michael Richardson joins the room
[12:59:53] Michael Richardson leaves the room
[12:59:59] Michael Richardson joins the room
[13:55:36] Michael Richardson leaves the room
[13:56:19] Michael Richardson joins the room
[13:59:20] Michael Richardson leaves the room
[13:59:37] Michael Richardson joins the room
[15:00:13] Michael Richardson leaves the room
[15:00:22] Michael Richardson joins the room
[15:20:12] inesrob joins the room
[15:21:58] inesrob has set the subject to: CORE WG | minutes at http://tools.ietf.org/wg/core/minutes | slides at https://datatracker.ietf.org/meeting/85/materials.html#wg-core
[15:40:57] Michael Richardson leaves the room
[15:41:16] Michael Richardson joins the room
[15:42:27] Michael Richardson leaves the room
[16:19:36] Klaus joins the room
[16:24:27] седая борода joins the room
[16:24:40] Andrew McGregor joins the room
[16:25:31] Robert Cragie joins the room
[16:27:04] <Robert Cragie> Announcement re. COMAN activity
[16:27:04] Cullen Jennings leaves the room
[16:28:27] <Robert Cragie> Request to subscribe to mailing list: https://www.ietf.org/mailman/listinfo/coman and review and contribute
[16:28:35] jlcJohn joins the room
[16:30:24] <Robert Cragie> groupcomm, part deux in consolidated slides (http://tools.ietf.org/agenda/85/slides/slides-85-core-0.pdf)
[16:35:14] <Robert Cragie> 4/5 emphasis on PUT being allowed as idempotent
[16:37:14] <Robert Cragie> Questions
[16:38:13] <Robert Cragie> Regarding single request and three responses - Router-1 brokers the communication with the lights
[16:38:46] <Robert Cragie> Zach Shelby: What is the benefit over sending directly to lights?
[16:39:19] <Robert Cragie> A: Router is in a better position to perform DNS
[16:39:58] <Robert Cragie> ZS: Could configure light switches to join group and then send directly. If you don't know the group, could do it this way
[16:40:52] <Robert Cragie> ZS: This model is broken right now
[16:41:00] <Robert Cragie> Carsten at mic
[16:41:28] <Robert Cragie> Haven't considered proxying unicast req to multicast proxy yet
[16:42:02] <Robert Cragie> CB: would also prefer to see light switch control light directly
[16:42:52] <Robert Cragie> CB: Could ask proxy to aggregate responses for lights but that has issues (how long to wait)
[16:43:47] <Robert Cragie> Three comments: 1) Careful about terminology, esp. router - not always a complicated device. People below 30 think it is a complicated device
[16:44:14] <Robert Cragie> 2) Mentioned draft presented in TLS WG re. DTLS multicast ecurity
[16:44:17] <Robert Cragie> security
[16:44:53] <Robert Cragie> 3) Question for implementers - do you look at whether it came from multicast or unicast?
[16:45:32] <Robert Cragie> CB: Some OSes do not distinguish when it comes to multicasts
[16:45:38] <Robert Cragie> Cullen Jennings at mic
[16:45:56] <Robert Cragie> (some debate over this)
[16:46:10] <Robert Cragie> CJ: Interested in case where there is no DNS
[16:46:35] <Robert Cragie> HERFP (?) problem
[16:46:57] <Robert Cragie> Experience in SIP worth looking at
[16:47:47] <Robert Cragie> Aggregating of responses is challenging
[16:47:53] <Robert Cragie> Matt Gillmore at mic
[16:48:05] <Robert Cragie> What would happen if light 3 didn't get the packet?
[16:48:36] <Robert Cragie> AR: Light3 would not respond
[16:48:47] <Robert Cragie> As it is a NON, no expectation of a response
[16:48:56] <Robert Cragie> Not a critical error
[16:49:02] <Robert Cragie> Andrew McGregor at mic
[16:49:35] <Robert Cragie> How would recovery happen?
[16:49:46] <Robert Cragie> AR: Application would recover it
[16:50:30] <Robert Cragie> AR: Will do update of draft and request review on the mailing list
[16:50:59] <Robert Cragie> Next slides: Best Practices for HTTP-CoAP Mapping
[16:51:06] <Robert Cragie> Salvatore Loreto presents
[16:52:17] <inesrob> Thank you very much Robert
[16:54:47] <Robert Cragie> Rhetorical questions re. https to coaps - would it always map to a coaps session? What about multicast security?
[16:55:51] <Robert Cragie> Still to decide re. multicast
[16:57:01] <Robert Cragie> Don't talk about reverse proxy yet, i.e. CoAP client generating HTTP req
[16:59:52] <Robert Cragie> Carsten: Next steps slide - I would normally delete it but I think it is an important document
[17:00:37] <Robert Cragie> Questions from Carsten: Who has implemented HTTP-CoAP proxy (5 or so hands)
[17:00:49] <Robert Cragie> Who has done it differently?
[17:00:56] <Robert Cragie> Cullen raises hand is at mic
[17:01:23] <Robert Cragie> CJ: I find it difficult to implement from the document
[17:01:45] <Robert Cragie> CJ:What about HTTP proxy? Caches?
[17:02:20] <Robert Cragie> CJ: For the people who built systems - how did you get it to work because I am struggling?!
[17:02:55] <Robert Cragie> CB: This draft is trying to document common expectations
[17:03:36] <Robert Cragie> CJ: I totally agree we need a document but the approach used in current document doesn't seem to work
[17:03:43] <Robert Cragie> Zach at mic
[17:04:04] <Robert Cragie> ZS: It's getting better. There are too many options - need to narrow it down more
[17:04:20] <Robert Cragie> We know embedded mappings don't work
[17:04:26] yuichi.igarashi joins the room
[17:04:36] <Robert Cragie> Question feasibility of direct scheme mapping
[17:04:58] <Robert Cragie> Table 1 is a train wreck
[17:05:06] <Robert Cragie> Some weird codes
[17:05:17] <Robert Cragie> Not sure it even belongs in this document
[17:06:43] <Robert Cragie> Table suggests doing something server side is better than client side - not sure this recommendation should be in draft
[17:07:37] <Robert Cragie> Lots of reasons for putting a reverse proxy
[17:08:18] <Robert Cragie> Shouldn't recommend TCP is better over the internet for the reasons stated - firewall may be a valid reason
[17:08:55] <Robert Cragie> SL: Not had any feedback from people which is why it is where it is
[17:09:10] <Robert Cragie> Carsten at mic
[17:09:47] <Robert Cragie> CB: Need some text regarding how reverse proxy would work (demo on screen)
[17:10:09] <Robert Cragie> Need to agree on ways of doing it
[17:10:44] <Robert Cragie> SL: Appreciate there are priorities in WG but we need to keep working on this
[17:11:35] <Robert Cragie> CB: Would ask chairs to send out call for adoption as WG document
[17:12:08] <Robert Cragie> Next slides - Dealing with congestion issues in CoAP
[17:12:16] <Robert Cragie> Presented by Carsten Bormann
[17:12:32] <Robert Cragie> Goodput vs pps
[17:13:09] <Robert Cragie> pps - packets per second
[17:13:21] <Robert Cragie> next slide
[17:13:32] <Robert Cragie> 77
[17:15:08] <Robert Cragie> slide 78
[17:15:52] sal joins the room
[17:17:49] <Robert Cragie> Lars Eggert at mic
[17:18:11] <Robert Cragie> LE: Need to see some evaluation - just like with TCP
[17:18:17] <Robert Cragie> Andrew McGregor at mic
[17:18:56] <Robert Cragie> AM: CoAP does provide enough info for cc. This is a step in the right direction - show at least one algorith which works
[17:19:00] <Robert Cragie> LE at mic
[17:19:38] <Robert Cragie> LE: Need to show that competing application clients can be given fairness
[17:20:10] <Robert Cragie> LE: There are different TCP congestion controls
[17:20:39] <Robert Cragie> Want to make sure there is fairness between them
[17:21:02] <Robert Cragie> Scenarios are quite different but the way we evaluate cc is the same
[17:21:28] <Robert Cragie> CB: Would like to find simple but high performance cc
[17:21:35] yuichi.igarashi leaves the room
[17:21:57] <Robert Cragie> slide 79
[17:23:02] <Robert Cragie> Request for reviewers - who has enough knowledge? (one hand shown)
[17:23:24] <Robert Cragie> Gorry Fairhurst volunteers
[17:23:41] <Robert Cragie> Next slides: SOLACE
[17:23:55] <Robert Cragie> Slide 80
[17:24:04] <Robert Cragie> Waiting for Alper Yegin...
[17:24:22] <Robert Cragie> So moved onto CoRE Roadmap
[17:24:42] <Robert Cragie> Slide 93
[17:24:51] <Robert Cragie> slide 94
[17:25:30] <Robert Cragie> A lot of drafts - can't discuss them all
[17:25:41] <Robert Cragie> slide 95
[17:27:04] <jlcJohn> Apparently somebody turned off the audio feed... hopefully it will be reincarnated as 1230
[17:27:48] <inesrob> ok, thanks
[17:27:49] <Robert Cragie> OK will try to relay
[17:28:21] <Robert Cragie> CB: Who has looked at the draft? (about 7 - all think it is useful)
[17:28:30] <Robert Cragie> CB: Going to continue
[17:28:47] <Robert Cragie> Gorry Fairhurst at mic
[17:29:08] <Robert Cragie> GF: Useful to read - evolving document. What's the endpoint?
[17:29:27] <Robert Cragie> CB: It's a bit like a star!
[17:29:34] <Robert Cragie> Barry Leiba at mic
[17:29:59] <Robert Cragie> BL: Start as a Wiki page and morph into draft. Would be a better format
[17:30:37] <Robert Cragie> CB: A Wiki wouldn't describe all the docs in a similar way
[17:31:14] <jlcJohn> Audio back :^)
[17:31:17] <Robert Cragie> CB: Who was in saag meeting? (4 people)
[17:31:34] <Robert Cragie> @jlcjohn: Thanks for letting me know
[17:31:59] <Robert Cragie> Back to slide 80
[17:32:33] <Robert Cragie> slide 82
[17:32:39] <Robert Cragie> slide 83
[17:32:42] <Robert Cragie> slide 84
[17:33:14] <Robert Cragie> slide 85
[17:34:54] <Robert Cragie> slide 86
[17:35:49] <Robert Cragie> slide 87
[17:36:49] <Robert Cragie> slide 89
[17:36:53] <Robert Cragie> sorry 88
[17:37:35] <Robert Cragie> slide 89
[17:38:00] <Robert Cragie> No need to invent new protocols - we have plenty
[17:39:47] <Robert Cragie> Cullen at mic
[17:40:03] <Robert Cragie> Whose plan?
[17:40:13] <Robert Cragie> CB: The solace people
[17:40:53] <Robert Cragie> CB: Not sure this should be an IRTF activity
[17:41:07] <Robert Cragie> (Sorry that was CJ)
[17:41:22] <Robert Cragie> CJ: I plan on bringing solution to core WG
[17:41:29] <Robert Cragie> Zach at mic
[17:41:55] <Robert Cragie> ZS: When we started core we weren't sure what security solution would be
[17:42:08] <Robert Cragie> We went the public key route
[17:42:25] <Robert Cragie> Didn't we solve that?
[17:43:06] <Robert Cragie> CJ: The first slide said we need a key management protocol - what is that? TLS is a key mgmt protocol
[17:43:42] <Robert Cragie> The issue of how you get it to enrol plus all the other lifecycle steps - we don't have a way to do that
[17:43:53] <Robert Cragie> Need to address the problem
[17:44:02] <Robert Cragie> Behcet Sarikaya at mic
[17:44:19] <Robert Cragie> Bootstrapping is initial
[17:45:02] <Robert Cragie> Back to preso - side 91
[17:45:05] <Robert Cragie> slide 912
[17:45:12] <Robert Cragie> slide 91
[17:45:28] <Robert Cragie> Request for several contributions
[17:45:58] <Robert Cragie> Then look at contributions and extract structure
[17:46:19] <Robert Cragie> Cullen at mic
[17:46:36] <Robert Cragie> CJ: Agree with plan but we should do it here
[17:47:30] <Robert Cragie> Andrew McGregor: Are you aware what happened with HIP?
[17:47:49] <Robert Cragie> AM: Parallel research group set up with WG
[17:48:09] <Robert Cragie> AM: This seems to be a similar case
[17:48:40] <Robert Cragie> AM: Is anyone going to be satsified with a bootstrapping solution?
[17:48:52] <Robert Cragie> CJ: Charter says we should work on it
[17:49:21] <Robert Cragie> zach at mic
[17:49:53] <Robert Cragie> zs: would feel we can solidify what we need in this wg
[17:51:04] yuichi.igarashi joins the room
[17:52:24] <Robert Cragie> (sorry was at the mic mysefl
[17:52:46] <jlcJohn> no problem!
[17:52:49] <Robert Cragie> Hannes Tschofenig at mic
[17:54:09] <Robert Cragie> We may need to pick parts and reuse other parts in some cases, in others you can't. In other cases may talk to multiple parties. Even at the app layer there may be multiple authenticaton methods
[17:55:16] <Robert Cragie> CB: Encourage people to generate contributions
[17:55:40] <Robert Cragie> Alper Yegin at mic
[17:55:59] <Robert Cragie> Look at ETSI M2M which descibes bootstrapping
[17:56:37] <Robert Cragie> CJ: What's the conclusion?
[17:57:23] <Robert Cragie> AM: IRTF should go ahead with framework but WG should develop protocols
[17:58:19] <Robert Cragie> CJ: Have to meet the charter which says have to look at not just protocol security but how the keys get there
[17:58:48] <Robert Cragie> Would not want to see this to wait on IRTF group to do this work
[17:59:42] <Robert Cragie> Expecting to have some way that vendors can build devices which work together
[18:01:08] <Robert Cragie> Robert at mic
[18:01:35] <Robert Cragie> RC: Should be no dependency on IRTF group but framwork developed from that should be used if possible
[18:01:52] <Robert Cragie> Can do in conjuncton
[18:02:02] <Robert Cragie> Back to slides
[18:02:09] <Robert Cragie> Slide 92
[18:02:27] <Robert Cragie> Slide 96
[18:02:41] <Robert Cragie> Salvatore Loreto presenting
[18:03:30] <Robert Cragie> slide 97
[18:04:52] <Robert Cragie> slide 98
[18:06:53] <Robert Cragie> slide 99
[18:06:59] <Robert Cragie> slide 100
[18:09:35] <Robert Cragie> slide 101
[18:09:46] <Robert Cragie> Kostas ? at mic
[18:10:30] <седая борода> ? == Pentikousis
[18:10:42] <Robert Cragie> K: Don't agree that client/server model is not applicable to Intermittent Presence device
[18:11:21] <Robert Cragie> Gorry at mic
[18:11:25] yuichi.igarashi leaves the room
[18:11:48] <Robert Cragie> Are there more levels of sleep than the two presented
[18:12:20] <Robert Cragie> KP: Scaling between the sleeping stages
[18:12:36] <Robert Cragie> Hannes at mic
[18:12:53] <Robert Cragie> HT: This has been done for a long time - Wake on LAN for example
[18:13:19] <Robert Cragie> There is little we can do in the IETF on energy efficiency
[18:13:55] <Robert Cragie> SL: This is to see what we can do at the CoAP level only
[18:14:12] <Robert Cragie> Lars at mic
[18:14:34] <Robert Cragie> Surprised to see this come up - I thought it was always the intention that CoRE supports these devices
[18:15:03] <Robert Cragie> Could have a network where devices go away and come back and whether they are sleeping or not is irrelevant
[18:15:42] <Robert Cragie> Up to the implementer as to how this is done surely
[18:16:37] <Robert Cragie> carsten at mic
[18:17:27] <Robert Cragie> important to identify that there are "real" sleepy devices which do not handle client/server model
[18:20:38] <Robert Cragie> KP: If we can figure out how to maintain state in the network without polling it would be interesting
[18:20:47] <Robert Cragie> Andrew McGregor at mic
[18:21:27] <Robert Cragie> Observing a resource on a sleepy node probably needs a proxy
[18:21:42] <Robert Cragie> Lars at mic
[18:22:22] <Robert Cragie> When observing the network, it is not obvious what consititutes a sleepy device
[18:22:41] <Robert Cragie> Therefore these devices need to be taken into consideration
[18:23:34] <Robert Cragie> SL: Do we need any extra protocol work? Would like people to think about
[18:23:42] <Robert Cragie> Matthias Kovatsch at mic
[18:24:23] <Robert Cragie> MK: Need to consider the cases where nodes to sleep 99% of the time. Need mirroring or publish options if possible
[18:24:36] <Robert Cragie> IEEE concerned with MAC layer
[18:25:30] <Robert Cragie> Akbar Rahman presents
[18:25:38] <Robert Cragie> Slide 102
[18:26:11] <Robert Cragie> Slide 103
[18:27:15] <Robert Cragie> slide 104
[18:27:17] <Robert Cragie> slide 105
[18:27:27] <Robert Cragie> slide 106
[18:28:00] <Robert Cragie> slide 107
[18:28:17] <Robert Cragie> slide 108
[18:28:43] <Robert Cragie> slide 109
[18:29:26] <Robert Cragie> Lars at mic
[18:29:33] <Robert Cragie> Over engineering this
[18:29:59] <Robert Cragie> You might not be able to talk to the device even if it thinks its up
[18:30:43] седая борода leaves the room
[18:32:56] <Robert Cragie> Carsten draws attention to IPR declared on the draft
[18:33:09] <Robert Cragie> Lars at mic
[18:33:35] <Robert Cragie> It's a lot of effort to maintain these cariables and it could do more harm than good
[18:34:20] sal leaves the room
[18:34:53] <Robert Cragie> Meeting is adjourned
[18:34:59] <inesrob> Thank you very much for the notes
[18:35:08] Robert Cragie leaves the room
[18:35:29] Andrew McGregor leaves the room
[18:37:18] inesrob leaves the room
[18:37:27] Michael Richardson joins the room
[18:51:03] Andrew McGregor joins the room
[18:53:39] Andrew McGregor leaves the room
[19:11:42] Michael Richardson leaves the room
[19:13:10] Michael Richardson joins the room
[19:57:29] Michael Richardson leaves the room
[21:51:35] Klaus leaves the room
[21:53:56] Michael Richardson joins the room
[22:14:14] Michael Richardson leaves the room
[22:14:42] Michael Richardson joins the room
[22:15:06] Michael Richardson leaves the room
[23:05:09] Michael Richardson joins the room
[23:22:59] Michael Richardson leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!