[09:52:35] --- mafr has joined
[09:53:38] --- mafr has left
[11:54:47] --- magnus has joined
[11:57:18] --- admcd has joined
[12:00:15] --- sarolaht has joined
[12:01:04] --- cgn has joined
[12:01:04] --- lars has joined
[12:01:35] --- admcd has left
[12:01:49] --- admcd has joined
[12:02:50] --- mafr has joined
[12:02:50] --- mafr has left: Lost connection
[12:03:39] <admcd> slide available at: https://datatracker.ietf.org/public/meeting_materials.cgi?meeting_num=67
[12:06:05] --- rbless has joined
[12:13:36] --- toyokkie has joined
[12:14:07] <rbless> John Loughney: Need to get charter done. First priority
[12:14:35] --- loughney has joined
[12:14:39] <rbless> Currently: Robert Handock up presenting GIST (ntlp-11 draft)
[12:14:52] <rbless> s/Handock/Hancock
[12:15:01] <rbless> :-)
[12:18:48] <rbless> Now up: Path splitting issue
[12:20:14] <rbless> Now up: Routing State Errors
[12:22:01] <rbless> Now up Security I: GIST/NSLP LAyer Split
[12:22:26] <rbless> Now up: Security II
[12:24:41] <rbless> Hannes: General case is C-Mode option
[12:25:12] <rbless> Henning: can be done without breaking backwards compatability
[12:26:16] <rbless> Now up Security III
[12:27:57] <rbless> Now up: supplementary question
[12:28:47] <rbless> Hannes: normally intra-domain case you don't care about these security issues -> different trust model
[12:29:25] <rbless> Henning: Even within a domain you may want to ensure security for customers..
[12:30:09] <rbless> John: ADs suggested applicability statement for GIST
[12:30:18] --- mafr has joined
[12:30:53] <rbless> Henning: avoid dependency in Spec on app statement...
[12:31:21] <rbless> Now up: I hate router alert
[12:31:42] <rbless> Now up: a reminder
[12:33:11] <rbless> Now up Status for IPv6
[12:33:59] <rbless> (Concerning Router Alert Option)
[12:34:27] --- m_ersue has joined
[12:36:05] <rbless> Free BSD ROA implementation is broken, not such a problem for Linux (not optimal), Router implementations usually better
[12:37:20] <rbless> Now up: IPv4 RAO issues
[12:39:21] --- elwynd has joined
[12:40:25] <rbless> John: if packets are lost will GIST recover?
[12:40:36] <rbless> Reobert: yes.
[12:42:06] <rbless> Now up possible solutions
[12:42:45] <rbless> Slides are BTW: http://www3.ietf.org/proceedings/06nov/slides/nsis-1.ppt
[12:44:29] <rbless> Hannes: Henning Peters and Christian Dickmann made measurements with RAOs, usually GIST routed to DEV/NULL in DSL networks..
[12:50:56] <rbless> Henning: maybe for the next interop some router vendors could bring some boxes
[12:51:36] <rbless> Magnus Westerlund: should fix the SPECS...
[12:52:28] <rbless> Adrian Farrell: need to fix the specs or need to fix the implementations?
[12:53:18] <rbless> Robert: maybe not meant to FIX but to clarify the Specs referring to RAOs
[12:57:27] <rbless> Adrian Farrell: maybe better to use a non-UDP encapsulation Q-Mode, needs new IP protocol number
[12:58:06] <rbless> Henning Schulzrinne: concerned about the path forward...
[12:59:00] <rbless> HS: worried about last minute changes...not convinced that this is a good way forward..
[12:59:55] <rbless> Dave Oran: early review from routing AD was done by Alex Zinin
[13:00:21] <rbless> Dave Oran: fixing the spec doesn't help to get the implementation fixed
[13:02:05] <rbless> Dave Oran: must keep UDP encapsulation so there is any chance to get over NATs....
[13:03:22] <rbless> John: need to get summary of observed behavior that Hannes mentioned
[13:05:19] <rbless> Henning: even if later change to non-UDP encapsulation is made, this doesn't break the protocol totally
[13:05:50] <rbless> Adrian: don't see anything here that damages the Internet, mainly GIST damages itself
[13:08:13] --- m_ersue has left
[13:08:14] --- loughney has left
[13:08:21] <rbless> Dave Oran: ASIC vendors are usually ignoring all kind of IP options (not just RAO) and leave them in the forwarding path
[13:08:31] <rbless> Now up: QSPEC discussion
[13:09:45] --- loughney has joined
[13:12:48] <rbless> Jerry Ash presenting http://www3.ietf.org/proceedings/06nov/slides/nsis-7.ppt
[13:17:10] <rbless> Henning Schulzrinne: looks like RSVP TSPEC, differences?
[13:17:23] <rbless> Dave: same thing, the math is the math
[13:21:47] <rbless> Juergen Quittek: would the control plane be simpler than RSVP?
[13:31:23] --- mafr has left: Replaced by new connection
[13:35:29] <rbless> Dave: must discuss three different possilities on the list: roland's proposal, stacking, having recurse QoS NSLP sessions
[13:39:31] --- mafr has joined
[13:39:44] <rbless> Georgios: Alternative 3 is already there, we need one or two
[13:40:55] <rbless> Jukka Manner: clarification for stacking on slide 8, because previous slides said that there is no stacking
[13:42:30] <rbless> Jerry: previous slide mentioned stack removal, but latest discussion seems to not agree on that, so kept stacking
[13:44:42] <rbless> John: have a call on consensus
[13:45:55] <rbless> on the list and then update the draft
[13:46:24] <rbless> Now up: Jukka Manner and the QoS NSLP disscussion
[13:47:24] --- magnus has left
[13:50:42] --- magnus has joined
[14:02:10] <rbless> Dave: Maybe recusrion in NSLP is sufficient...
[14:03:06] <rbless> Dave: when one wants to decouple state machines for RMF admission, it may be good to use recursion for that...
[14:03:32] <rbless> John: is recursion a possible and sufficient solution for RMD?
[14:06:05] --- xiaomingfu has joined
[14:06:43] <rbless> Georgios: stacking is a different thing than recursion, sometimes they are the same sometimes not
[14:07:08] <rbless> John: would be good to have an example on the list
[14:07:20] <rbless> for all the three options...
[14:07:57] <rbless> and also to get feedback from implementors
[14:08:32] <rbless> Now UP: Martin Stiemerling with nat/fw NLSP
[14:09:17] <rbless> requirements in draft-bajko-nsis -FW-reqs-05, currently all covered by the SPEC
[14:14:25] <mafr> three!
[14:16:05] <cgn> Now I do know as well :-)
[14:16:35] <rbless> Now UP: Louis Cordeiro NSIS interop test results
[14:17:23] <rbless> s/Louis/Luís
[14:17:24] --- loughney has left
[14:21:51] --- sarolaht has left
[14:30:01] <rbless> John: no time to discuss any other QoS models now, need to wait for QSPECs
[14:31:35] --- cgn has left: Logged out
[14:31:39] <rbless> John: need to resolve any open issues from the charter and then move on
[14:32:03] --- lars has left
[14:32:31] <rbless> Meeting closed.
[14:32:39] --- rbless has left
[14:32:48] --- toyokkie has left
[14:32:53] --- xiaomingfu has left
[14:32:58] --- mafr has left
[14:33:04] --- toyokkie has joined
[14:34:17] --- toyokkie has left
[14:39:24] --- magnus has left
[14:50:20] --- admcd has left
[14:53:55] --- elwynd has left
[15:13:40] --- elwynd has joined
[16:16:07] --- elwynd has left: Replaced by new connection
[16:16:08] --- elwynd has joined
[17:14:45] --- elwynd has left
[17:55:10] --- elwynd has joined
[18:00:35] --- elwynd has left: Replaced by new connection
[18:00:36] --- elwynd has joined
[18:00:36] --- elwynd has left