[05:07:07] --- reh has joined
[09:42:13] --- Martin Stiemerling (NSIS WG chair) has joined
[09:46:42] <Martin Stiemerling (NSIS WG chair)> Slides for the session are here https://datatracker.ietf.org/meeting/69/materials.html
[09:47:16] --- toyokkie has joined
[09:51:24] --- jugi has joined
[09:51:32] --- jugi has left
[09:51:47] --- Jukka has joined
[09:55:23] --- reh has left
[09:55:40] --- reh has joined
[09:56:21] --- admcd has joined
[10:00:25] --- lars.eggert@googlemail.com has joined
[10:02:31] <lars.eggert@googlemail.com> is the audio OK?
[10:03:10] <admcd> ok for me
[10:03:17] <lars.eggert@googlemail.com> ok, good
[10:03:21] --- magnus westerlund has joined
[10:03:23] <Martin Stiemerling (NSIS WG chair)> For me quite low level
[10:03:27] <Martin Stiemerling (NSIS WG chair)> but acceptable
[10:03:31] <magnus westerlund> ok
[10:03:33] --- sarolaht@jabber.org has joined
[10:03:51] --- sarolaht@jabber.org has left
[10:04:02] --- PasiS has joined
[10:04:39] <PasiS> jukka doing wg update
[10:04:46] --- m_ersue has joined
[10:05:15] <PasiS> gist status
[10:05:20] <PasiS> all the open issues have been fixed
[10:05:30] <PasiS> according to robert who is not in the meeting
[10:05:52] <Martin Stiemerling (NSIS WG chair)> BTW: Thanks to pasis for taking jabber notes!
[10:06:29] <PasiS> lars: should have a half a page email to iesg with the draft update to highlight the key changes
[10:07:13] <Martin Stiemerling (NSIS WG chair)> John is doing a write-up end of this week!
[10:07:30] <PasiS> gist will be resent to magnus
[10:08:25] <PasiS> nsis interop report, roland bless
[10:08:48] <PasiS> listing participants (on slides)
[10:09:38] <PasiS> mainly focused on testing gist
[10:10:06] <PasiS> gist test cases from christian dickmann
[10:10:20] --- bxmina has joined
[10:10:58] <PasiS> also tested qos-nslp but no time to test firewall implementations
[10:11:10] <PasiS> tested all major protocol features
[10:11:25] <PasiS> nat traversal was not tested
[10:11:54] <PasiS> explaining christian's draft structure
[10:12:37] <PasiS> i.e., tests in more detail
[10:13:09] --- reh has left: Replaced by new connection
[10:13:10] --- reh has joined
[10:13:29] <PasiS> conclusions
[10:13:44] <PasiS> works quite well, not too complex to implement
[10:14:06] <PasiS> have different independent implementations
[10:14:21] <PasiS> only some clarifications required
[10:14:29] <PasiS> robert will be updating these to ver-14
[10:14:38] <Martin Stiemerling (NSIS WG chair)> Changes are already in -14
[10:14:47] <reh> at least, i hope so
[10:15:01] <Martin Stiemerling (NSIS WG chair)> I believe ;-)
[10:15:52] <PasiS> helpful to have automated testing
[10:15:57] <Martin Stiemerling (NSIS WG chair)> Question to Roland
[10:16:02] <PasiS> ok
[10:16:06] <Martin Stiemerling (NSIS WG chair)> Are those test programs available for free?
[10:16:21] <PasiS> ill ask
[10:16:24] <Martin Stiemerling (NSIS WG chair)> thanks!
[10:17:05] <PasiS> (missed was lars said hile walking to mic)
[10:17:28] <Martin Stiemerling (NSIS WG chair)> he was forwarding a rumour from some interop participants
[10:17:35] <Martin Stiemerling (NSIS WG chair)> about code base of GIST implementations
[10:18:58] <PasiS> martin: yes
[10:19:16] <Martin Stiemerling (NSIS WG chair)> Thanks to Uni Göttingen for making this open source! :)
[10:20:04] <PasiS> if someone has a test case he thinks is missing just send to christian
[10:20:24] <PasiS> jukka: univ helsinki is going to do a c implementation, aiming for full implementation
[10:20:32] <PasiS> cannot say about the timeline
[10:21:00] <PasiS> univ of (?) implementation is in java
[10:21:06] <Martin Stiemerling (NSIS WG chair)> uni of Coimbra
[10:21:10] <PasiS> thx
[10:21:35] <PasiS> next: qos nslp presented by jukka
[10:21:55] <PasiS> wglc ended in june, many comments most favourable
[10:22:04] <PasiS> no serious objections
[10:22:40] <PasiS> will update the draft based on comments the send the spec forward
[10:22:48] <PasiS> any comments? no
[10:23:38] <PasiS> now up: nat/fw nslp presented by hannes
[10:25:16] <Martin Stiemerling (NSIS WG chair)> A comment to NATFW: We can run the WGLC around end of August/early September
[10:25:26] <Martin Stiemerling (NSIS WG chair)> A comment to NATFW: We can run the WGLC around end of August/early September
[10:25:45] <PasiS> no slides
[10:25:45] <PasiS> a security review was made, some minor comments
[10:25:45] <PasiS> differences between the last two versions were just summarising the security issues and some editorial things
[10:25:45] --- Jukka has left: Replaced by new connection
[10:25:45] <PasiS> jukka: what's next?
[10:25:45] <PasiS> hannes: depends on the chairs
[10:25:45] <PasiS> hannes: would also appreciate comments
[10:25:58] <Martin Stiemerling (NSIS WG chair)> What about my comment?
[10:26:00] <PasiS> will say
[10:26:04] <Martin Stiemerling (NSIS WG chair)> thx!
[10:26:44] <Martin Stiemerling (NSIS WG chair)> Jep, we need a new WGLC :)
[10:26:47] <PasiS> hannes: sounds good to me
[10:27:14] <PasiS> hannes showing the table of contents on screen
[10:27:22] --- kenichi has joined
[10:28:11] <PasiS> next up: mobility applicability
[10:28:21] <PasiS> changes between -06 and -07
[10:28:39] <PasiS> terminology cleaned up
[10:28:50] <PasiS> further study section cleaned up
[10:28:57] <PasiS> editorial fixes
[10:29:05] <PasiS> open issue section created
[10:29:12] <PasiS> open issue (1/4)
[10:29:28] --- Kungro has joined
[10:29:39] <PasiS> about tearing direction
[10:30:11] <PasiS> open issue (2/4) about mip interaction
[10:30:23] <PasiS> how should be cleaned up?
[10:30:29] <PasiS> open issue (3/4)
[10:30:34] <PasiS> nat/fw interaction
[10:30:49] <PasiS> when and how should nsis tear down pinhole?
[10:31:20] <PasiS> ms may not know whether nat is nsis aware or not
[10:31:26] <PasiS> open issue (4/4)
[10:31:34] <PasiS> security considerations
[10:31:37] <PasiS> schedule
[10:31:52] <PasiS> discuss open issues by mid oct
[10:32:00] <PasiS> early nov submit -08
[10:32:15] <PasiS> mid nov discuss wglc comments
[10:32:22] <PasiS> no comments about the draft
[10:32:44] <PasiS> next up: rmd-qosm -- georgios
[10:32:54] <PasiS> main updates since ietf-68
[10:33:11] <PasiS> updates qspec and qos nslp draft changes
[10:34:12] <PasiS> some editorial changes
[10:34:16] <PasiS> next step
[10:34:25] <PasiS> waiting to be sent to iesg for review
[10:34:43] <PasiS> any comments from ads?
[10:34:46] <PasiS> no
[10:35:07] <PasiS> now up: qspec
[10:35:13] <PasiS> no one is talking about it
[10:35:28] <PasiS> skipping it
[10:36:12] <PasiS> hannes: on qspec -- have sent mail to the list, use qspec parameters in different context in the dime wg
[10:36:46] <PasiS> have dime meeting on friday, ops ads plan to get the documents finished within next couple of months
[10:36:50] <PasiS> looking for reviewers
[10:37:10] <PasiS> if the qspec doesn't go forward it will block the other document
[10:37:24] <PasiS> hoping the qspec to go forward pretty soon
[10:37:50] <PasiS> y.1541 now
[10:37:52] <PasiS> no slides
[10:37:59] <PasiS> editorial updates
[10:38:08] <PasiS> terminology changes
[10:38:19] <PasiS> wglc was asked
[10:38:32] <PasiS> but andrew mcdonald sent comments to the list
[10:38:40] <PasiS> don't know id action was required
[10:39:07] <PasiS> wants wglc
[10:39:54] <PasiS> roland: jerry retired, what does it mean to the document
[10:40:08] <PasiS> A: should find the source of the file
[10:40:36] <PasiS> roland: auth48 requires that the author agrees on document, hopefully jerry is still reachable
[10:41:26] <PasiS> lars: can invoke override procedure on auth48 if jerry is not reachable, but would be easier if he could be contacted at the time
[10:41:34] <PasiS> A: believe he is reachable
[10:41:54] <PasiS> next: nsis operation over ip tunnels
[10:42:00] <PasiS> no one coming to present it
[10:42:05] <PasiS> skipping to state machine
[10:42:11] <PasiS> no one presenting that either
[10:42:37] <PasiS> hannes: giving a presentation to state machine is challenging
[10:42:53] <PasiS> hannes: no changes, just resubmitted, nothing to present
[10:43:01] <PasiS> gist over sctp?
[10:43:09] <PasiS> christian dickmann
[10:43:27] <PasiS> doc not yet published
[10:43:50] <PasiS> sctp as transport would benefit from multi-streaming, multi-homing, etc.
[10:44:06] <PasiS> adds a protocol-id to gist
[10:44:31] <PasiS> did a lot of updates:
[10:44:46] <PasiS> added new ma-protocol-id for "dtls"
[10:44:49] --- batfli has joined
[10:44:56] <PasiS> open issues
[10:45:09] <PasiS> is it ok to have both "sctp" and "dtls" in same doc
[10:45:43] <PasiS> magnus: i cannot see good reason to split the documents
[10:45:54] <PasiS> A: dtls could also be used with dccp
[10:46:18] <reh> not clear to me why you need a separate id for DTLS anyway
[10:46:30] <PasiS> want me to proxy you?
[10:46:42] <reh> yes please
[10:46:45] <PasiS> (are you robert h.??)
[10:46:47] <PasiS> :-)
[10:46:51] <reh> yes :-)
[10:46:55] <reh> i promise to think about it
[10:47:00] <reh> (the dtls question)
[10:48:25] <PasiS> magnus: what is status of dtls over sctp?
[10:48:37] <PasiS> A: not ready yet
[10:49:12] <PasiS> magnus: what's the relationship between normal sctp and pr-sctp
[10:50:00] <PasiS> detailed text for dtls still has to be sorted out
[10:50:12] --- Martin Stiemerling (NSIS WG chair) has left
[10:50:13] <PasiS> need help on reviewing
[10:50:15] --- Martin Stiemerling has joined
[10:50:31] <PasiS> next up some individual drafts
[10:51:02] <reh> i think the underlying question is: if there are multiple ways to use transport layer security with SCTP, should the gist/sctp document define a single one which is mandatory to support (still allowing negotiations within [d]tls and sctp to choose options of course), or are there multiple useful configurations for which the negotiation has to start within GIST.
[10:51:14] <reh> but, i see the topic has moved on; we can do it on the mailing list.
[10:51:42] <PasiS> magnus: how about implementation report, it is on charter?
[10:51:43] --- jnob has joined
[10:52:04] <Martin Stiemerling> Which document/milestone
[10:52:05] <Martin Stiemerling> ?
[10:52:13] <PasiS> last milestone, dec 2006
[10:52:17] <Martin Stiemerling> wait
[10:52:24] <lars.eggert@googlemail.com> Dec 2006    Submit 'General Internet Signaling Transport protocol implementation document' to the IESG for publication as an Informational RFC
[10:52:39] <PasiS> people suggested to add that to state machine document a long time ago
[10:52:45] <PasiS> (that was said by hannes)
[10:53:07] <PasiS> magnus: are there people who are willing to work on that?
[10:53:40] <Martin Stiemerling> I assume that there was the wish to fix some implementation guidelines
[10:53:48] <PasiS> hannes: doesn't have to be on state machine document, it is a bit stretch to have impl. issues there
[10:53:48] <Martin Stiemerling> in the past with this milestone
[10:54:22] <PasiS> going to mic
[10:55:32] <Martin Stiemerling> need to re-check this milestone togehter with the state machine draft
[10:55:33] <PasiS> hannes: many things have changes since then, couple of implementations are ready
[10:57:30] <PasiS> magnus: sounds that group does not need the document right now, i'm ok to kill a milestone that is not needed
[10:58:54] <PasiS> hannes: Q to ADs and chairs. most changes have been minor editorial updates. would it be possible to simply ship the documents? i would push them forward
[10:59:04] <PasiS> you can change editorial changes forever
[10:59:18] <PasiS> there is not much sense in waiting forever
[10:59:36] <PasiS> magnus: gist wasn't ready, so cannot ship the document before gist is fixed
[11:00:07] <PasiS> hannes: just want to see some stuff leaving the wg
[11:00:21] <PasiS> for example could ship qos
[11:00:41] <Martin Stiemerling> Chair's question: We started with the state machine draft and now discuss QoS WGLC?
[11:00:42] <PasiS> things have been fixed
[11:00:46] <Martin Stiemerling> I'm somewhat lost...
[11:03:27] <PasiS> hannes: about pcn, missed the comment
[11:03:44] <PasiS> (i guess i'm lost as well)
[11:04:20] <PasiS> magnus: pcn needs to show that they can come up with something work. let's charter that work then
[11:04:25] <PasiS> too early at this point
[11:04:45] <PasiS> NOW... moving to individual drafts
[11:05:13] <PasiS> hong cheng: multihoming support
[11:05:54] <PasiS> motivation: multihoming supported on ip layer
[11:06:10] <PasiS> transparent to the app on end host
[11:06:16] --- jnob has left
[11:06:21] <PasiS> no predefined flow spearation
[11:07:10] <PasiS> scenarios for nsis multihoming
[11:08:41] <PasiS> issues:
[11:08:47] <PasiS> session binding information
[11:09:31] <PasiS> same mri for multiple flows in route optimization cases
[11:09:57] <PasiS> home agent support in merging and splitting of signaling
[11:10:24] <PasiS> fast path selection
[11:10:41] <PasiS> path relationship indication in mobile environment
[11:11:04] <PasiS> showing example picture on slide
[11:12:38] <PasiS> next step
[11:12:46] <PasiS> encourage discussion
[11:13:22] <PasiS> relationship with mobility applicability draft
[11:13:34] <PasiS> not all multihoming cases are mobility related
[11:13:55] <PasiS> wants to take this as a wg item
[11:14:16] <PasiS> questions?
[11:14:34] <Martin Stiemerling> It is too early to take this as WG item
[11:14:45] <Martin Stiemerling> as we need to get the other items done first
[11:14:48] <PasiS> lars: strongly pushing back any new item until the current items are done
[11:15:22] <PasiS> hong: fine with not taking it as wg item right now, but later
[11:15:48] <PasiS> jukka: these presentations are for informational purpose for now, for future consideration
[11:16:18] <PasiS> lars: can open discussion on gist-bis in future, the design phase is over now
[11:17:05] <PasiS> hannes: didn't see any problems with multihoming and gist earlier when going through the different issues
[11:17:17] <PasiS> lars: performance optimizations to a non-deployed protocol are useless
[11:18:05] <PasiS> next: hypath -- luis cordero
[11:18:22] <PasiS> implemented and tested in european research project euqos
[11:18:27] <PasiS> changes
[11:19:58] <PasiS> hypath approaches: mrm and tlv
[11:20:59] <PasiS> on working issues
[11:21:05] <PasiS> choose the best approach
[11:21:09] <PasiS> security issues
[11:21:16] <PasiS> scalability
[11:21:18] <PasiS> implementation
[11:21:24] <PasiS> questions? no
[11:22:22] <PasiS> next: inter-domain reservation agregation for qos nslp -- roland bless
[11:22:32] <PasiS> just to make you aware of the related problems
[11:22:48] <PasiS> end-to-end qos requires interdomain signaling
[11:23:04] <PasiS> control plane scalability required
[11:23:17] <PasiS> need aggregation concept
[11:25:05] <PasiS> provide mechanisms for dynamic creation of inter-domain aggregations
[11:25:09] <PasiS> problems
[11:25:22] <PasiS> determination of aggregator and deaggregator
[11:25:58] <PasiS> signaling between aggregator and deaggregator
[11:26:55] <PasiS> a priori determination of flow's pth
[11:27:16] <PasiS> route change detection -- explained with a figure of slide
[11:28:06] <PasiS> proposed changes
[11:28:17] <PasiS> not changes to existing documents
[11:28:24] <PasiS> but for future extensions
[11:28:46] <PasiS> qos nslp: new route-record object to find deaggregators
[11:29:41] <PasiS> gist: new af-mrm - can be used for route check
[11:30:12] <PasiS> soliciting input from anyone interested
[11:30:48] <PasiS> hannes: interesting stuff. have you worked on an overall architecture?
[11:31:02] <PasiS> seems to extend rmd beyond one domain
[11:31:14] <PasiS> A: totally decoupled from rmd
[11:31:35] <PasiS> would require many changes
[11:31:46] <PasiS> and extensions to qos-nslp
[11:32:08] <PasiS> some things are already in qos-nslp
[11:32:31] <PasiS> georgios: agree with hannes. if we do rechartering would like to have this included
[11:33:13] <PasiS> next: authorization for nslp -- jukka manner
[11:33:25] <PasiS> nslp would need to distinguish users
[11:33:32] <PasiS> status
[11:33:37] <PasiS> draft has been stable for long time
[11:34:51] <PasiS> hannes: i think it is useful (as a coauthor) helps to tie sip signaling with qos signaling
[11:35:42] <PasiS> luis: can see some interest in this work. started to implement this
[11:35:53] <PasiS> will try to give some comments
[11:36:07] <PasiS> roland bless: thinks this is important to get this done
[11:36:16] <PasiS> once other items are finished
[11:36:51] <PasiS> jukka: did some implementation of this
[11:37:03] <PasiS> seems to work
[11:37:23] <PasiS> next presentation: jukka: peering data for nslps
[11:37:35] --- m_ersue has left
[11:37:45] <PasiS> for making peering decisions on reliable or secure sessions
[11:39:00] <PasiS> could be used for getting metering information
[11:39:29] <PasiS> hannes: one thing that confuses me is the use of word "peering2
[11:39:44] <PasiS> jukka: peering comes from gist spec
[11:41:13] <PasiS> the proposal
[11:41:27] <PasiS> add new gist object: peering information object
[11:41:59] <PasiS> allows new types of interactions and nslp
[11:42:25] <PasiS> roland: question: why isn't the current mechanisms for providing nslp data sufficient data
[11:43:11] <PasiS> jukka: you need a number of messages to set up the state
[11:44:11] <PasiS> that was the last presentation
[11:44:11] --- m_ersue has joined
[11:44:15] <PasiS> any further comments?
[11:44:31] <Martin Stiemerling> Thanks to all and especially to Jukka!
[11:44:41] <reh> thanks to all too
[11:45:03] <PasiS> lars: 1) need to get gist used on something before doing extensions to it
[11:45:26] <PasiS> should push gist to users of rsvp
[11:46:10] <PasiS> 2) three new rsvp related drafts in tsvwg. nsis people might be interested to participate
[11:46:14] <PasiS> end of meeting
[11:46:17] <Martin Stiemerling> Bye bye!
[11:46:23] <PasiS> bye
[11:46:26] --- lars.eggert@googlemail.com has left
[11:46:38] --- Martin Stiemerling has left
[11:46:48] --- admcd has left
[11:46:52] --- m_ersue has left
[11:47:07] --- bxmina has left
[11:47:16] --- PasiS has left
[11:47:47] --- Kungro has left
[11:48:23] --- toyokkie has left
[11:51:40] --- kenichi has left
[11:52:39] --- magnus westerlund has left
[11:53:14] --- reh has left
[12:41:09] --- batfli has left