IETF
dots
dots@jabber.ietf.org
Friday, April 8, 2016< ^ >
Yoav Nir has set the subject to: DOTS Meeting - IETF 94
Room Configuration
Room Occupants

GMT+0
[12:52:48] Meetecho joins the room
[12:54:50] christopher.morrow joins the room
[12:56:41] <christopher.morrow> howdy! - jabber scribing for dots this morning is me :)
[12:56:54] <christopher.morrow> please speak up if you have questions we should be starting in ~5 mins time.
[12:57:47] Tirumaleswar Reddy joins the room
[12:57:49] Gilbert Clark joins the room
[12:58:49] Hiroyuki Goto joins the room
[13:00:26] Kathleen Moriarty joins the room
[13:01:50] <Kathleen Moriarty> Good morning!  I just got the following error from the server when trying to connect to meetecho:
[13:01:54] <Kathleen Moriarty> ype Exception report
message
description The server encountered an internal error that prevented it from fulfilling this request.
exception
java.lang.NullPointerException
    com.meetecho.ajax.AjaXMPP.doPost(AjaXMPP.java:2167)
    com.meetecho.ajax.AjaXMPP.doGet(AjaXMPP.java:510)
    javax.servlet.http.HttpServlet.service(HttpServlet.java:624)
    javax.servlet.http.HttpServlet.service(HttpServlet.java:731)
    org.apache.tomcat.websocket.server.WsFilter.doFilter(WsFilter.java:52)
note The full stack trace of the root cause is available in the Apache Tomcat/7.0.68 logs.
[13:02:08] <Kathleen Moriarty> I'll start up audio only again
[13:02:13] Lorenzo Miniero joins the room
[13:02:30] <Meetecho> Kathleen Moriarty: can you try again? just tried and it works for me now
[13:03:04] Kathleen Moriarty_3812 joins the room
[13:03:15] <Kathleen Moriarty> OK, it's working now.  Thanks
[13:03:37] Lorenzo Miniero leaves the room
[13:04:03] <christopher.morrow> meeting start && agenda bashing underway
[13:07:11] <christopher.morrow> slide 3 coming, I believe after the initial usecases discussions.
[13:07:30] <christopher.morrow> now roland doing usecases drafts.
[13:07:44] <christopher.morrow> roland slide 2 now
[13:08:08] <christopher.morrow> slide 3
[13:08:59] <christopher.morrow> slide set for - draft-ietf-dots-use-cases-01 (Roland Dobbins, 10 min)
slide 4 now.
[13:09:22] <christopher.morrow> slide 5
[13:11:34] <christopher.morrow> slide 6
[13:12:39] <christopher.morrow> slide 7
[13:12:47] <christopher.morrow> questions ?
[13:13:18] Dan Wing joins the room
[13:13:33] <christopher.morrow> linda dunbar's question: "what exactly is a client, can you provide an example"
[13:13:46] <christopher.morrow> roland: "lots of examples on list, more < list provided >"
[13:14:07] <christopher.morrow> daniel - ericsson - 02 should have more feedback on this as well.
[13:14:22] <christopher.morrow> bob moskowitz - components and such will be outlined in -02
[13:14:59] <christopher.morrow> next at the mic: Kaname -  - draft-nishizuka-dots-inter-domain-usecases-01 (Kaname Nishizuka, 10 min)
[13:15:01] <christopher.morrow> slide 2
[13:16:02] <Kathleen Moriarty> Thanks for the detailed jabber scribing!  Much appreciated!
[13:16:11] <christopher.morrow> sure.
[13:16:13] <christopher.morrow> slide 3
[13:16:26] <christopher.morrow> (I wish there were slide numbers on these slides)
[13:16:36] <christopher.morrow> 'categorization of usecases' is the title.
[13:16:50] <christopher.morrow> slide 4
[13:17:33] <christopher.morrow> slide 5
[13:17:40] <christopher.morrow> 'protection methods'
[13:17:43] Tirumaleswar Reddy leaves the room
[13:17:43] Tirumaleswar Reddy_3188 joins the room
[13:19:20] <christopher.morrow> slide 6 - attack telemetry
[13:19:57] Alexander Lyamin joins the room
[13:20:03] <christopher.morrow> slide 7
[13:20:06] <christopher.morrow> 'ddos protection status'
[13:21:28] <christopher.morrow> slide 8 - 'inter-domain usecases'
[13:21:59] <christopher.morrow> slide 9 - usecase 1
[13:22:12] <christopher.morrow> slide 10 - usecase2
[13:22:24] <christopher.morrow> slide 11 - usecase 3
[13:23:17] <christopher.morrow> slide 12 - usecase 4
[13:23:27] <christopher.morrow> oops, sorry 12 is delegation model.
[13:23:35] Tirumaleswar Reddy joins the room
[13:23:39] <christopher.morrow> (and it's the current slide - 12)
[13:23:58] Tirumaleswar Reddy_3188 leaves the room
[13:24:25] <christopher.morrow> andreas - question at the mic
[13:25:46] <christopher.morrow> 'how much of the delegation model is in scope here?' (and some quesitons about labeling on the slide... relay/server/client, please clarify)
[13:26:24] <christopher.morrow> kaname: free to include more info in the signaling should work fine. (more on list)
[13:27:04] <christopher.morrow> roland at the mic - referring back to 'protection methods' slide - all of the data bits listed probably shouldn't be in the dots proto.
[13:27:19] <christopher.morrow> they are useful for mitigation..
[13:28:01] <christopher.morrow> daniel at the mic - same as roland... also not sure about how to do the sharing between domains.
[13:28:34] KShiomoto joins the room
[13:28:53] <christopher.morrow> chair (tobias) jumps in for coordination conversations (with i2nsf/etc0
[13:29:39] Tirumaleswar Reddy_4696 joins the room
[13:29:56] <christopher.morrow> linda dunbar at the mic - 'flowspec no workie? why?'
[13:30:08] Tirumaleswar Reddy leaves the room
[13:30:26] <christopher.morrow> kaname - seems there was some confusion... it's possible that we misunderstood eachother (more on-list/off-mic later)
[13:30:46] <christopher.morrow> skipping slides to centalized architecture
[13:30:51] <christopher.morrow> and now last slide 'next steps'
[13:31:17] <christopher.morrow> kaname's next preso:  - Use cases of draft-nishizuka-dots-inter-domain-mechanism-00 (Kaname Nishizuka, 10 min)
[13:32:00] <christopher.morrow> chairs say: there's 2 preos for this, one for usecases (this one) one for protocol (from frank later)
[13:32:03] <christopher.morrow> slide 1
[13:32:04] <christopher.morrow> slide 2
[13:32:09] <christopher.morrow> 'overview'
[13:32:46] <christopher.morrow> slide 3
[13:32:56] <christopher.morrow> 'architecture of cooperatove ddos protection'
[13:33:30] <christopher.morrow> slide 4 distributed arch
[13:35:28] Alexander Lyamin leaves the room
[13:35:56] <Kathleen Moriarty> Chris Morrow speaking?
[13:36:16] <christopher.morrow> (thnx!)
[13:36:27] <christopher.morrow> frank - clarifying some.
[13:36:36] <christopher.morrow> outlining scenarios more clearly is useful.
[13:36:53] Tirumaleswar Reddy_4696 leaves the room
[13:37:10] Tirumaleswar Reddy joins the room
[13:38:05] <christopher.morrow> roland at the mic - topology isn't super important here.
[13:38:59] <christopher.morrow> roland - different networks/operators may have different gear/capabilities - so prescritptive solutions are not as good. 'lowest common denonimator is important'.
[13:39:31] <christopher.morrow> andreas - disagree - possible at l7 for problems to require much more assitance/info.
[13:40:11] <christopher.morrow> alexander lemon - diff tech will bring diff results (especially l7)
[13:40:20] <christopher.morrow> tobias - hat-off at mic
[13:40:25] Tirumaleswar Reddy leaves the room
[13:40:41] <Dan Wing> I agree that layer 7 has HTTPS today, and more tomorrow.  But "tomorrow" layer 4 may have TCPINC.
[13:40:45] Tirumaleswar Reddy joins the room
[13:41:11] <christopher.morrow> more coordination would be beneficial if possible.
[13:41:34] <christopher.morrow> also, my 'andreas' previously is ... fleming. (sorry!)
[13:41:53] <christopher.morrow> speaker back!
[13:42:09] <christopher.morrow> slide 5 - centralized architecture.
[13:43:23] <christopher.morrow> slide 6 - challenges for interodomain cooperative ddos protection
[13:44:37] <christopher.morrow> fyi, all the slides appear to be online at:
https://datatracker.ietf.org/meeting/95/materials/
(search for dots)
[13:45:35] <christopher.morrow> dave - sandvine - questions about usecase info requiring info about inbound sources/etc.
[13:46:52] Ida Leung joins the room
[13:46:58] <christopher.morrow> dave again at mic
[13:47:36] <christopher.morrow> re-asking his question for clarification
[13:47:48] <christopher.morrow> frank  at mic - responding - seems like a good question dave.
[13:49:14] <christopher.morrow> liquan - at the mic
[13:49:21] <christopher.morrow> tobias -pushign for list discussion
[13:49:27] <christopher.morrow> alexander lemon - at the mic
[13:49:42] <christopher.morrow> dots requirements preso - andrew
[13:49:55] <christopher.morrow> - draft-ietf-dots-requirements-01 (Andrew Mortensen, 10 min)
[13:50:05] <christopher.morrow> slide 2
[13:50:11] <christopher.morrow> (changes outline)
[13:50:23] <christopher.morrow> slide 3 revies intro section
[13:50:35] <christopher.morrow> slide 4 added terminology
[13:50:49] <christopher.morrow> slide 5 updated terminology
[13:51:12] <christopher.morrow> slide 6 reqiurements intro
[13:51:57] <christopher.morrow> slide 7 general requirements
[13:52:16] Hiroyuki Goto leaves the room
[13:52:25] Hiroyuki Goto joins the room
[13:53:39] <christopher.morrow> slide 8 operational requirements (cont'd)
[13:54:41] <christopher.morrow> frank at the mic - question about notification/conflict detection
[13:55:17] <christopher.morrow> andrew: more discussion is required here... let's talk at the end of this preso.
[13:55:28] <Gilbert Clark> is the intent that support for mitigation lifetime be *required*?
[13:55:37] <Gilbert Clark> not mic, just curious
[13:55:41] <christopher.morrow> fleming at the  mic: ttl ignoring?
[13:56:01] <christopher.morrow> I think 'not required' as much...per request.
[13:56:07] <christopher.morrow> but protocol required, make sense?
[13:56:27] <Gilbert Clark> think so, yup.  thanks
[13:57:07] <christopher.morrow> bob at the mic trying to answer fleming as well.
[13:57:49] <christopher.morrow> roland at the mic - peering for dots not require dns/resolution.
[13:58:11] <christopher.morrow> daniel - mailing-list question plus timestamp for lifetime quesiton
[13:58:17] <christopher.morrow> (sounds like gilbert's question sort of as wlel)
[13:58:39] <christopher.morrow> slide 9 data channel requirements
[13:59:13] <christopher.morrow> slide 10 - security requirements.
[13:59:19] <christopher.morrow> slide 11 next steps
[13:59:46] <christopher.morrow> slide 12 - done!
[13:59:48] <christopher.morrow> oops
[13:59:57] <christopher.morrow> dave - sandvine - missing methods-o-mitigation/etc
[14:00:15] <christopher.morrow> andrew - out of scope most likely?? discuss!
[14:00:55] <christopher.morrow> slide 12 for realz now!
[14:01:18] <christopher.morrow> andrew for next slide set as well: - draft-mortensen-dots-architecture-00 (Andrew Mortensen, 15 min)
[14:01:23] <christopher.morrow> slide 1
[14:01:38] <christopher.morrow> slide 2 overview
[14:02:33] <christopher.morrow> slide 3 - basic architecture
[14:03:16] <christopher.morrow> slide 4 establishing signaling sessions
[14:03:52] <christopher.morrow> slide 5 direct signaling
[14:04:14] <christopher.morrow> slide 6 relayed signaling
[14:04:43] <christopher.morrow> slide 7 relayed signaling issues
[14:05:39] <christopher.morrow> slide 8 dots agent relationships
[14:05:52] <christopher.morrow> slide 9 multihomed dots client
[14:06:18] <christopher.morrow> slide 10 additional signaling modes
[14:07:00] <christopher.morrow> slide 11 next steps
[14:07:34] <christopher.morrow> slide 12 thankyou! (in blue)
[14:07:58] <christopher.morrow> roland at the mic - privacy concerns /etc
[14:08:11] <christopher.morrow> sharing is probably required.../etc.
[14:08:57] Dan Wing leaves the room
[14:09:08] Dan Wing joins the room
[14:09:16] <christopher.morrow> dan wing at the mic - end all relays! </pitchfork>
[14:09:30] <Tirumaleswar Reddy> hummmmm
[14:09:56] <christopher.morrow> presenting dan wing -  draft-reddy-dots-transport-03 (Dan Wing, 10 min)
[14:10:09] <christopher.morrow> slide 1
[14:10:18] <christopher.morrow> slide 2 - changes to draft
[14:10:35] <christopher.morrow> slide 3 happy eyeballs-like technique
[14:10:39] <christopher.morrow> oops, not yet.
[14:11:18] <christopher.morrow> asking for feedback on CoAP vs http/etc.
[14:11:25] <christopher.morrow> (asking for on-list discussion)_
[14:11:28] <christopher.morrow> slide 3 for real now
[14:12:38] <christopher.morrow> slide 4 - happy eyeballs for dots signal channel
[14:13:06] <christopher.morrow> slide 5 dots signal channel
[14:13:31] <christopher.morrow> slide 6 dots data channel
[14:13:56] <christopher.morrow> slide 7 'why CoAP for dots?'
[14:14:12] Gilbert Clark leaves the room
[14:15:03] Gilbert Clark joins the room
[14:15:05] <christopher.morrow> slide 8 'running code'
[14:15:09] <christopher.morrow> daniel at the mic
[14:15:31] Gilbert Clark leaves the room
[14:15:40] <christopher.morrow> (does json/cbor/etc make sense woth coap? expand on why)
[14:16:04] <christopher.morrow> nik at mic - clarification on proto/channel
[14:16:51] <christopher.morrow> slide 8 again
[14:17:11] Gilbert Clark joins the room
[14:17:21] <christopher.morrow> slide 9 use cases
[14:17:36] <christopher.morrow> slide 10 performance cosniderations
[14:18:29] <christopher.morrow> slide 11 draft-reddy-dots-transport-03
[14:18:37] <christopher.morrow> andrew at the mic.
[14:18:53] <christopher.morrow> disconect in draft for coap vs flowspec/rest/etc...
[14:19:04] <christopher.morrow> tobias at mic
[14:19:24] <christopher.morrow> concensus questions should go to the mic - says dan.
[14:20:00] <Kathleen Moriarty> count?
[14:20:11] <christopher.morrow> minimal count on reading :(
[14:20:17] <Kathleen Moriarty> Thanks
[14:20:25] <christopher.morrow> dave sandvine at the mic
[14:21:02] <christopher.morrow> dan - the data model sort of needs shaking out here... happy eyeballs means some paralellism.
[14:21:44] Ida Leung leaves the room
[14:21:50] <christopher.morrow> nik at the mic
[14:22:20] <christopher.morrow> fleming at the mic
[14:23:59] <christopher.morrow> questions about asynchronicity in the conversation between client/server...
[14:24:38] <christopher.morrow> liquan  at the mic
[14:25:00] <christopher.morrow> daniel at the mic
[14:25:11] Gilbert Clark leaves the room
[14:25:47] <christopher.morrow> perhaps  the simplicity rquirements argue against coap...
[14:25:49] <christopher.morrow> (from dan)
[14:26:08] <christopher.morrow>             - draft-reddy-dots-transport-03 (Dan Wing, 10 min)
[14:26:23] <christopher.morrow> oh copy/paste..
[14:26:30] <christopher.morrow> frank slide 1 of:  Protocol of draft-nishizuka-dots-inter-domain-mechanism-00 (Frank Xialiang, 10 min)
[14:26:55] <christopher.morrow> slide 3 - merits of interdmain cooperatove ddos.
[14:27:53] <christopher.morrow> slide 4 - challenges for interdomain cooperatove ddos protection
[14:28:15] Gilbert Clark joins the room
[14:28:17] <christopher.morrow> slide 5 - distributed architecture
[14:28:28] <christopher.morrow> slide 6 centralized architecture
[14:28:53] <christopher.morrow> slide 7 interdomain ddos protocol
[14:30:00] <christopher.morrow> slide 8 provisioning stage protocol
[14:31:10] <christopher.morrow> slide 9 signaling stage protocol
[14:31:49] Gilbert Clark leaves the room
[14:32:18] <christopher.morrow> slide 10 signaling stage protocol (cont)
[14:32:39] <christopher.morrow> slide 11 implemenation related
[14:33:22] <christopher.morrow> slide 12 next steps
[14:33:23] Gilbert Clark joins the room
[14:33:52] <christopher.morrow> slide 13 thanks!
[14:34:04] <christopher.morrow> roland at the mic, roll back to protocol data slide
[14:34:05] Einar Lanfranco joins the room
[14:35:41] <christopher.morrow> andrew at the mic
[14:36:21] <christopher.morrow> (referencing slide 9 btw)
[14:36:33] <christopher.morrow> liquan  at the mic
[14:37:27] <christopher.morrow> asking for the chair to clarify scoping
[14:37:34] <christopher.morrow> tobias a the mic - as chari
[14:37:36] <christopher.morrow> chair
[14:39:13] <christopher.morrow> roman asking for clarification of which usecase goes in to which draft.
[14:40:27] <christopher.morrow> lots of back/forth with frank/roman... about coverage/etc.
[14:41:41] <christopher.morrow> tobias asking for running code examples/etc
[14:41:58] <christopher.morrow> nik - at the mic - pointing at some code exisitng and talking about: "lots of existing code/etc"
[14:42:29] <christopher.morrow> kaname - at mic
[14:42:48] <christopher.morrow> kaname states they have code intenrally for the draft he talked to previously.
[14:42:57] <christopher.morrow> andrew at the mic
[14:44:31] Dan Wing leaves the room
[14:48:12] Gilbert Clark leaves the room
[14:48:17] Gilbert Clark joins the room
[14:49:56] Dan Wing joins the room
[14:51:55] <Tirumaleswar Reddy> media streams RTP uses UDP, WebRTC uses UDP for data channels, QUIC uses UDP
[14:52:00] KShiomoto leaves the room: Replaced by new connection
[14:52:14] KShiomoto joins the room
[14:52:52] <Gilbert Clark> except GRE wouldn't eliminate the need for UDP - UDP or TCP would need to be tunneled through GRE to make that work
[14:53:05] <Gilbert Clark> it's not a transport protocol in and of itself
[14:54:12] christopher.morrow joins the room
[14:54:26] <christopher.morrow> (sorry, laptop decided to vacation from the internet)
[14:55:19] <christopher.morrow> bob presenting - - Introducing session layer considerations (Robert Moskowitz, 5 min)
[14:55:45] <christopher.morrow> slide 4 - what ssls services not needed
[14:55:47] <Kathleen Moriarty> he's pretty easy to understand/hear, no worries
[14:55:52] <christopher.morrow> :)
[14:56:14] <christopher.morrow> slide 5 - so what is ssls providing
[14:56:25] KShiomoto leaves the room
[14:58:03] <christopher.morrow> slide 6
[14:58:48] <christopher.morrow> slide 7 discussion.
[14:59:17] <christopher.morrow> roland at the mic
[15:00:17] <christopher.morrow> tobias at the mic
[15:00:23] <christopher.morrow> <nohat>
[15:00:28] christopher.morrow leaves the room
[15:00:32] <christopher.morrow> ...question for roland...
[15:00:53] <christopher.morrow> roland - yes, we should add usecase data for this
[15:01:20] <christopher.morrow> roland at the mic -
[15:01:49] <christopher.morrow> times up.
[15:02:01] <Kathleen Moriarty> thank you, Chris!
[15:02:09] <christopher.morrow> sure thing
[15:02:10] christopher.morrow leaves the room
[15:02:15] <Gilbert Clark> thanks :)
[15:03:05] Tirumaleswar Reddy leaves the room
[15:03:12] Einar Lanfranco leaves the room
[15:03:13] Kathleen Moriarty_3812 leaves the room
[15:03:20] Gilbert Clark leaves the room
[15:03:21] Hiroyuki Goto leaves the room
[15:04:30] Meetecho leaves the room
[15:07:57] Dan Wing leaves the room
[15:25:22] Kathleen Moriarty leaves the room
[15:46:58] KShiomoto joins the room
[15:47:04] KShiomoto leaves the room