[04:07:56] --- christian.dickmann has joined
[04:08:05] --- christian.dickmann has left
[04:52:33] --- hong_psl has joined
[04:52:48] --- hong_psl has left
[05:56:03] --- christian.dickmann has joined
[09:40:09] --- admcd has joined
[09:48:46] --- rbless has joined
[09:54:41] --- bew has joined
[09:57:54] --- hong_psl has joined
[10:02:06] --- lars has joined
[10:02:10] --- cgn has joined
[10:02:42] --- Magnus Westerlund has joined
[10:03:31] --- dlpartain has joined
[10:03:43] --- SAN has joined
[10:05:06] --- hong_psl has left
[10:06:22] --- peeskeepr has joined
[10:06:54] --- HannesTschofenig has joined
[10:06:55] <rbless> John Loughney is bashing the agenda
[10:07:02] --- hong_psl has joined
[10:07:14] <rbless> and gives WG Status update
[10:07:44] <HannesTschofenig> Roland is jabber scribe; Paulo Mendes & Andrew are going to be our meeting minute takers for this WG session.
[10:08:13] <rbless> GIST will be sent to Transport AD, QoS NSLP needs more discussion & a revision, QSPEC, Y.1541&RMD got only minimal review, need more review before we can progress
[10:08:15] <HannesTschofenig> John says that no further work will be added to the charter if the existing work is not completed.
[10:08:40] <HannesTschofenig> Henning suggests to have (at least) two designated reviewers being assigned to the documents.
[10:08:49] <rbless> Henning Schulzrinne: pick reviewers, name them
[10:09:33] <rbless> Hannes: if people ask for reviews they should be also review other documents...
[10:10:00] <HannesTschofenig> I just suggested that people asking whether there is support for their new work that they also have todo a review of the old documents.
[10:10:30] <rbless> Sorry, got that wrong...
[10:10:36] --- elwynd has joined
[10:11:32] <HannesTschofenig> John presents the Milestone update.
[10:12:55] <HannesTschofenig> Robert starts his presentation on the GIST update
[10:16:07] <HannesTschofenig> John says that he wants to provide an implementation report to the IESG.
[10:16:24] <HannesTschofenig> Experience showed that the implementations are actually quite simple.
[10:16:29] <rbless> Would make IESG happier
[10:16:29] <HannesTschofenig> Next: QoS NSLP
[10:17:26] <HannesTschofenig> Andrew said that the WGLC did not reveal structural problems but a number of clarifications.
[10:17:49] <rbless> Hannes: are the slides available?
[10:18:05] <Magnus Westerlund> https://datatracker.ietf.org/public/meeting_materials.cgi?meeting_num=65
[10:19:13] <rbless> Thanks for the pointer
[10:19:56] <rbless> John: talked to IESG, for v4 only single RAO value, so need to dig through protocol headers, IPv6 supports more values...
[10:20:34] <rbless> John: do we want this also for IPv4?
[10:22:40] <rbless> Aggregation marking issues are more general, not so QoS specific
[10:23:12] --- dlpartain has left
[10:23:55] --- peeskeepr has left
[10:24:16] <rbless> ACK flag: ongoing discussion
[10:24:41] <rbless> rerouting needs more dicsussion
[10:26:46] <rbless> draft needs at least one more update...
[10:27:08] <rbless> John: QSPEC stacking seems to be big issue
[10:27:44] <rbless> Jerry Ash: error codes and qspec related errors needs also clarification
[10:28:36] <rbless> Next up: QSPEC
[10:29:15] <rbless> John is presenting current status (updates/open issues)
[10:29:48] <rbless> need more discussion on error codes
[10:30:31] <rbless> Jerry Ash: also discussion on intermediate domains and mandatory parameters, reaction to missing parameters etc...
[10:31:54] <HannesTschofenig> John speaks about the QSPEC and QOSM idea.
[10:32:19] <HannesTschofenig> (and the problems that can appear in various situations)
[10:34:01] <rbless> (Missed Robert's comment)
[10:35:47] <rbless> Hannes: there are issues in the QSPEC draft that do not belong there because they have more impact on qos nslp
[10:36:01] <rbless> need clarification, will sent comments to list
[10:36:15] <rbless> Next up: Y1541-QosM draft
[10:38:06] <rbless> Hannes: reviewed Y1541 model, so far good, but some clarifications on minimum QoS, sent reviews to the list
[10:38:25] <rbless> Next Up: RMD-QOS model
[10:39:40] <rbless> David Black: three high level issues
[10:40:09] <rbless> RMD mainly for bandwidth management, so mainly EF, no AF or other DiffServ PHBs
[10:40:12] <HannesTschofenig> Problems with congestion management stuff (remarking of traffic)
[10:40:19] <HannesTschofenig> IANA consideration section missing.
[10:41:22] --- elwynd has left
[10:42:42] <rbless> Georgios: most comments should be resolved by clarification, remarking needs more work
[10:43:16] <rbless> David: Block 16 DSCPs way too much, one would be ok
[10:43:43] <HannesTschofenig> Paulo Mendes says that it is not clear how RMD can be used with AF.
[10:44:00] <rbless> Paulo Mendes: Title misleading, because mainly EF related, no AF support, so not for all DiffServ classes
[10:44:43] --- cedaoun@gmail.com has joined
[10:44:45] <rbless> Georgios: maybe clarify by applicability statement
[10:45:18] <rbless> Next Up: Controlled Load Servcie QoS model
[10:46:22] --- elwynd has joined
[10:50:05] --- Magnus Westerlund has left
[10:54:43] <rbless> John: People interested in CL?
[10:56:18] <rbless> Jerry Ash: Briscoe PCN doing CL w/o signaling, NSIS CL with signaling, how does it fit?
[10:56:41] <rbless> Robert Hancock: Brisco's approach is edge-to-edge signaling
[10:56:58] <rbless> e2e feedback signaling, sorry..
[10:57:38] --- lars has left: Logged out
[10:59:42] <rbless> David Black: mechanism like PCN in the network, not flow based like intserv
[11:01:36] <rbless> John: how many people are interesting in controlled load classic service and signaling with NSIS
[11:01:42] <rbless> half-dozen hands
[11:02:00] <rbless> Who is interested in the new PCN like signaling?
[11:02:15] <rbless> slightly more hands up
[11:02:28] --- cgn has left: Logged out
[11:02:34] <rbless> will take this to the list
[11:03:01] --- bew has left: Disconnected
[11:03:40] <HannesTschofenig> Martin starts with the NATFW NSLP presentation
[11:12:36] <rbless> Martin ready for WGLC
[11:12:51] <rbless> Next version at start of April
[11:13:20] <rbless> Robert Hancock: one open issue is RAO value, how do you fix it until April?
[11:13:29] <rbless> Martin: not relevant for WGLC
[11:13:50] <rbless> Next up: mobility draft
[11:15:31] --- cgn has joined
[11:15:45] --- Andreas has joined
[11:19:35] --- Magnus Westerlund has joined
[11:20:56] --- cedaoun@gmail.com has left
[11:22:57] --- Magnus Westerlund has left
[11:22:58] <rbless> Tad Hardie: clarifying question how is it different for NETLMM and IS-IS/OPSF change?
[11:23:35] <rbless> Hannes: NETLMM would look like a route change
[11:24:05] <rbless> Robert: signaling would probably not different
[11:24:40] <rbless> Ted: mobile node will not be aware that it moved...
[11:25:14] <rbless> Hannes: doesn't change signaling protocol
[11:26:57] <rbless> John: NETLMM is ongoing work, document should be finished before NETLMM, so document some general issues
[11:32:51] <rbless> Seong-Ho: need more feedback from protocol developers.
[11:33:26] <rbless> John: like to let the QoS NSLP authors consider this drafts
[11:34:04] <rbless> Martin: this document is only for QoS NSLP, but there are likely also issues for NAT/FW
[11:34:12] --- Magnus Westerlund has joined
[11:34:18] <rbless> need to work on that...
[11:34:51] <rbless> John: will skip extensibility draft discussion to afternoon
[11:35:41] <rbless> Next up: path-decoul´pled signaling discussion
[11:36:22] <rbless> Luis Cordeiro is presenting draft-cordeiro-nsis-hypath-00
[11:43:38] <rbless> John: is this something that partners in the EuQoS project want to deploy?
[11:44:29] <rbless> Luis: France Telecom and Telefonica are project partners and currently implementing things..
[11:45:23] --- bew has joined
[11:45:26] <rbless> Next Up: Robert Hancock with path-decoupled problem
[11:51:00] <rbless> Idea is to let Query message untouched, follows normal data path
[11:51:15] <rbless> Some node may redirect this query message to an off-path node
[11:51:32] <rbless> establishing a messaging association
[11:52:29] --- bew has left: Disconnected
[11:54:27] --- elwynd has left: Replaced by new connection
[11:54:29] --- elwynd has joined
[11:56:57] <rbless> goal of document is an applicability statement for GIST, no protocol changes to be done
[11:57:12] <rbless> WG adoption?
[11:57:57] <rbless> John: how many people thinks this approach is useful?
[11:58:01] <christian.dickmann> I think that this is useful ;)
[11:58:01] <rbless> dozens hands
[12:00:36] <rbless> Paulo Mendes: quite interested in this work, but have questions about applicability
[12:01:13] <rbless> Paulo: Is it expected that this will work with all NSLPs?
[12:01:48] <rbless> Paulo: relation to inter-domain model?
[12:03:19] <rbless> Robert: maybe some issues in the inter-domain off-path signaling can be covered by this approach, so this is a first steps and then see what's left, what functionality is missing
[12:04:52] <rbless> Robert: for first question. API is totally unchanged, but doesn't mean that signaling application can stay unaware, because applications need to influence elements on the forwarding path...
[12:05:11] --- HannesTschofenig has left
[12:06:09] <rbless> Jürgen Quittek: for me its a migration, but no protocol change, is it in the charter of the WG?
[12:06:45] <rbless> John: pretty much applicability statement, it is likely that no charter changes are required
[12:07:13] <rbless> Jerry Ash: thought that this will require recharter...
[12:07:51] <rbless> John: is possibly require rechartering, but no significant change
[12:08:09] <rbless> Jerry: relation to other SDOs?
[12:09:47] <rbless> John: Action point, activate liason to ITU-T to make them aware of this work
[12:11:15] <rbless> Hannes: we have ongoing discussions with ITU
[12:12:45] <rbless> Magnus: need to discuss this work
[12:15:17] <rbless> Ted Hardie: problem with the statement: "make the forwarding node do something interesting", otherwise you get a lot of last call comments that this is a pig in a poke...please state the design choices, otherwise we have problems with this applicability statement
[12:15:54] <rbless> Robert: need to be more precise in the applicability statement
[12:17:02] <rbless> Jerry: good to see ITU liason coming to NSIS, maybe announce to list
[12:18:32] <rbless> Hannes: for the statement "make the forwarding node do something interesting" (point number2 on slide), there are different protocols that could be used, like QoS-Diameter application
[12:19:15] --- elwynd has left: Replaced by new connection
[12:19:15] --- elwynd has joined
[12:19:15] --- elwynd has left
[12:19:37] <rbless> Jürgen: 2nd point: it's not in the draft just on the slides
[12:20:16] <Magnus Westerlund> I assume the liason that we are talking about is: https://datatracker.ietf.org/public/liaison_detail.cgi?detail_id=195
[12:21:37] <christian.dickmann> what is the result of the vote?
[12:21:54] <SAN> re-chartering it seems
[12:21:56] <Magnus Westerlund> 20+ for. No against
[12:21:57] <rbless> John: How many people would adopt this as WG document and work on it as soon as charter is cleared
[12:22:21] <christian.dickmann> thanks
[12:24:35] --- cgn has left: Logged out
[12:26:02] <rbless> next up: John on extensibility draft
[12:28:42] <rbless> Hannes: dependencies from current NSLPs to ext-draft?
[12:29:49] <rbless> John: yes need to be fixed. Want these issues in the NSLPs, because it would be easier for implementers.
[12:30:06] <rbless> John: would make it WG document, but need more input
[12:30:20] <rbless> Robert: should make this as WG doc asap
[12:32:06] <rbless> Robert: IANA registry?
[12:32:57] <rbless> Magnus: must have really good guidance for IANA in registry document
[12:33:48] <rbless> Robert: would like to keep QoS model stuff out of ext-doc
[12:34:54] <rbless> Al Morton: support for keeping QoS model interaction discussion in this draft, esp. in background of discussion with other SDOs
[12:35:48] <rbless> Martin Stiemerling: document should provide guidance for people in 10 years
[12:36:44] <rbless> Meeting closes now, next session in the 2nd in afternoon
[12:36:55] --- rbless has left
[12:36:58] --- christian.dickmann has left
[12:38:07] --- SAN has left
[12:40:08] --- Magnus Westerlund has left
[12:50:17] --- Andreas has left
[12:55:03] --- admcd has left
[13:43:55] --- hong_psl has left
[16:13:39] --- hong_psl has joined
[16:15:11] --- christian.dickmann has joined
[16:15:40] --- HannesTschofenig has joined
[16:16:43] --- lars has joined
[16:17:26] <HannesTschofenig> Juergen talks about Metering NSLP on the mike
[16:17:43] --- admcd has joined
[16:17:46] <HannesTschofenig> John speaks in favor of the work.
[16:17:50] <HannesTschofenig> Sorry, Henning
[16:18:00] <admcd> /subject NSIS
[16:18:02] --- SAN has joined
[16:18:05] <HannesTschofenig> Hening speaks in favor of the work.
[16:19:11] <HannesTschofenig> Henning will now talk about the "NSIS Operation over IP Tunnels".
[16:19:24] * admcd has set the topic to: NSIS
[16:22:18] <HannesTschofenig> John mentiones that the group already agrees that this work will be done within this WG.
[16:24:01] --- rbless has joined
[16:27:28] <rbless> anyone doing the scribe?
[16:29:01] <rbless> discussion on NAT traversal draft
[16:29:51] <rbless> Martin: assumption wrong that flow information is not carried in NSLPs, so fixing it in GIST only will cause synch problems
[16:32:52] <christian.dickmann> the NAT traversal draft requires very little code changes in normal GIST nodes
[16:34:43] --- sarolaht has joined
[16:37:19] <HannesTschofenig> Current presentation: "Interdomain QoS Model"
[16:39:39] --- lars has left: Logged out
[16:41:39] --- cgn has joined
[16:42:25] <HannesTschofenig> Jerry thinks that Y.1541 and RMD are inter-domain QoS models. He asks what the draft adds beyond the already available stacking mechanisms defined in the QoS NSLP.
[16:42:38] <HannesTschofenig> Response: The treatment of non-NSIS domains.
[16:42:54] <HannesTschofenig> Georgios mentions that the off-path signaling components are also covered in this draft.
[16:43:01] <HannesTschofenig> Response: That's not the focus of the draft.
[16:43:59] <HannesTschofenig> Georgios: When you have a centralized resource management then do you have to use a path-decoupled approach of signaling approach?
[16:44:30] <HannesTschofenig> Paulo Mendes: The advantage is clear to me: Separate what happens within the domain and what happens outside the domain.
[16:44:36] <HannesTschofenig> NSIS can fulfill this requirement.
[16:45:16] <HannesTschofenig> What has a domain to gain with this? With this approach the operator is still exposing its network. Every signaling message arriving in the network still has an impact on the network.
[16:45:26] <HannesTschofenig> It is a question whether this is the most feasible one.
[16:45:51] <HannesTschofenig> The operators don't want to expose their networks.
[16:48:49] <cgn> Hannes: operators do not want to expose network
[16:49:01] <cgn> what does this mean w.r.t to figures in presenration
[16:49:09] <cgn> Paolo:
[16:49:17] <cgn> PHB belong in networks not outside
[16:49:57] <cgn> Luis: Qos Model net yet defined
[16:50:17] <HannesTschofenig> Hannes: There is no indication that this is in the case.
[16:50:20] <cgn> Georgios: Diffser could be used end to end
[16:50:27] <HannesTschofenig> Georgios: DiffServ can also be used E2E.
[16:50:35] <HannesTschofenig> Paulo: This also relates to Fred Bakers draft.
[16:51:06] <HannesTschofenig> The meaning of the DiffServ Codepoints is specific to a domain. Not on an end-to-end basis.
[16:51:24] <HannesTschofenig> Georgios: It also depends on the SLAs
[16:51:28] --- sarolaht has left
[16:51:31] <HannesTschofenig> Paulo: SLS
[16:51:43] <HannesTschofenig> Georgios: The SLS is the technical part of a SLA
[16:54:49] <rbless> Hannes presenting diagnostic NSLP
[16:56:17] <rbless> Martin: applicability?
[16:57:12] <rbless> Jürgen Quittek: right direction, but why doesn't the NSLP support it? Why NSLP discovery not built into GIST?
[16:57:35] <rbless> Hannes: yes, looking for input to sort out different design options
[16:57:49] --- cgn has left: Logged out
[16:58:37] <rbless> Luis: useful application/tool, esp. for network operations and for testing
[17:00:02] <rbless> Robert: useful thing to have. how much do you want it for diagnostic purpose, e.g. installed states etc.
[17:01:35] <rbless> John: send comments to list please
[17:02:59] <christian.dickmann> which presentation is now on?
[17:03:21] <SAN> Additional NSIS Communication Patterns
[17:03:28] <christian.dickmann> thanks
[17:12:53] --- SAN has left
[17:13:25] --- hong_psl has left
[17:25:41] --- christian.dickmann has left
[17:39:38] --- HannesTschofenig has left
[17:42:46] --- rbless has left: Disconnected
[17:43:49] --- admcd has left