[03:07:33] --- iljitsch has become available
[03:09:37] <iljitsch> virtual blue sheet: geoff and kurtis (chairs)
[03:10:17] <iljitsch> marcelo, jari, spencer, iljitsch, erik and two names that I didn't catch
[03:11:17] <iljitsch> geoff:
[03:11:18] <iljitsch> erik's document is useful but we need a functional decomposition
[03:14:54] <iljitsch> in what document should design decisions go?
[03:17:38] <iljitsch> keep design decisions in protocol doc for now?
[03:17:47] <iljitsch> erik: not everything is in there
[03:17:58] <iljitsch> geoff: part of it is in rfc 4177
[03:18:45] <iljitsch> geoff: keep design decisions in appendix for now
[03:18:57] <iljitsch> marcelo: also design decisions in other documents
[03:20:37] <iljitsch> geoff: hba document
[03:21:11] <iljitsch> geoff: think we're ready for last call
[03:21:20] <iljitsch> (wg last call)
[03:27:49] <iljitsch> geoff: we'll do a wg last call and then give it to the ads for security review etc
[03:29:28] <iljitsch> geoff: source address draft more operational
[03:46:07] <iljitsch> is this orthogonal to shim protocol? (discussion...)
[03:52:39] <iljitsch> do we want the huitema/marcelo ingress filtering draft as a working group document?
[03:53:01] <iljitsch> geoff: this is more about hosts telling the site routers how they want their packets to move out of the site
[04:01:53] <iljitsch> geoff: marcelo should generate a presentation pack about what he is trying to solve here as a presentation in vancouver
[04:02:07] <iljitsch> pekka enters the room
[04:03:55] <iljitsch> geoff: explain what THE problem is, not just part of it
[04:04:05] <iljitsch> next: address selection draft
[04:05:11] <iljitsch> geoff: apparently rfc 3484 has problems, maybe send this back to ipv6 wg?
[04:06:17] <iljitsch> marcelo: when correspondent is not shim enabled you need to explore source/dest combinations, not just dest
[04:06:25] <iljitsch> not related to shim6 protocol
[04:31:16] <iljitsch> marcelo and someone else are going to kickstart an update on rfc 3484 in another (!!) wg
[04:37:01] <iljitsch> referals draft is dormant right now
[04:39:34] <iljitsch> failure detection draft
[04:39:49] <iljitsch> merge with protocol? merge with reachability detect?
[04:53:49] <iljitsch> merge, pick one to make 01, don't care with, see about merging with protocol draft at ietf
[04:54:02] <iljitsch> pekka: maybe not merge, could reuse it for hip
[04:56:35] <iljitsch> geoff: remove nat/v4, iesg won't stand for it
[04:57:15] <iljitsch> geoff: yet unwritten document: interaction with upper layers / transports
[04:57:55] <iljitsch> pekka: can start this around the beginning of 2006
[05:03:35] <iljitsch> congestion: important but not now, how is interaction with apps people?
[05:03:55] <iljitsch> geoff: open drafts done
[05:04:56] <iljitsch> break for lunch
[05:05:06] --- iljitsch has left
[06:26:31] --- iljitsch has become available
[06:38:15] <iljitsch> erik's protocol draft up now, nearly everyone read it, nearly everyone has comments
[06:38:23] <iljitsch> erik:
[06:38:42] <iljitsch> decided to put context tag in flow label field
[06:38:51] <iljitsch> do uncoordinated shimn state removal
[06:39:01] <iljitsch> error message on lost state
[06:39:20] <iljitsch> us new ip protcol number for ccontrol (= parallel to icmp)
[06:41:43] <iljitsch> do we need protocol/flow label hack?
[06:41:58] <iljitsch> mtu problem doesn't seem too bad
[07:31:38] <iljitsch> disagreement on need to put in an 8 byte shim header when shim is active. iljitsch wants an option that a receiver can use during the negotiation that makes the sender NOT insert the header.
[07:37:37] <iljitsch> with flow label hack out the door we can use a bigger context identifier
[07:44:59] <iljitsch> do we need a checksum?
[07:47:03] <iljitsch> conclusion: checksum on every packet not necessary, but useful for signalling.
[10:28:59] --- iljitsch has left