[07:04:45] thomas.heide.clausen@gmail.com joins the room [07:05:00] teco.boot joins the room [07:05:40] Alex Petrescu joins the room [07:05:42] Agenda [07:06:10] Thanks Teco, when you want to speak on the mike I could take over scribing. [07:07:13] history [07:09:54] velt joins the room [07:10:37] Jari: to difficult to describe MANET general case. Therefore rechartered. [07:11:03] RFC5889 update [07:12:33] Issues raised. What are next steps? [07:12:44] Jari: [07:14:00] [Alex, Teco: Thanks for this....] [07:16:44] Ulrich Herberg joins the room [07:19:12] Alex Petrescu leaves the room [07:20:23] Alex Petrescu joins the room [07:20:54] Changes (2/4) [07:21:53] Charlie P.: Title: addressing is not only routers [07:22:10] CP: discussing on title: it doesnt pertain only to routers... the scope of LLs is intertwinned to neighbordhoodness** [07:22:29] CP: routers or non routers impacted by this difficulty in maintaining neighbohoodness of scope [07:22:47] CP: not persuaded to change the name of the doc [07:23:06] JA: personally also of the oppinion it would be good to add "A" not sure about the word " router" [07:23:12] Erik Nordmark [07:23:45] EN: "A" may be useful, but "router"... ther are applications, avoid the path of goinf only "routers"... [07:23:58] JA: limited but describining everything... [07:24:20] ThomasClausen: get to the point of understanding this is MANET interfaces of routers, not ifaces of hosts or... [07:24:53] CP: possible of MANET where many nodes do forwarding, others dont forward, even those who dont forward... straightforward host to have addresses... [07:25:08] TC: different than hosts having off...? [07:25:27] CP: not mobile, not Mobile IP... [07:25:36] TC: topologically correctness [07:26:07] CP: not about topo correctenss.. if you have aMANET with many ptp connectivity, then you could move a node from a place to another, keep sale addresse... [07:26:10] CP:... [07:26:25] JA: interesting, but what kinds of changes reqed to doc? [07:26:33] Teco Boot is TB [07:27:15] TB: new title is better addressing the document is about... because routing protocol, host, we need in this proporsal the rproto to advertise the prefixes...*TB: this is indeed a better title [07:27:44] TC: in SD we spent two hours on presentation... we do understand what a router is... not need to rewrite host apps and protos [07:28:01] TC: reopening new issues we WG we got ok? [07:28:09] TC: reevelauate and challenge? [07:28:16] JA: do you agree or disagree? [07:28:22] TC: I cant suggest changes [07:28:33] TC: doc as done approved by IESG satisfies... [07:28:38] EN: recommendation... [07:28:57] Changes (3/4) [07:30:44] Changes 4/4: back to original text [07:32:19] EN: issue is the doc still says the is fla... and that comes to the conclusion [07:32:30] EN: link locals have routability scope and uniqueness scope [07:32:37] EN: in case of.. different. [07:33:00] EN: yes we can not test for that uniqueness; but we cant test the uniquess of same address. [07:33:15] JA: rfc4862? [07:34:05] (listening) [07:40:17] There are LL addresses that are intended to be globally unique, and LL addresses that have link-scope uniqueness [07:41:41] Jari: Now we don't want to change to doc, unless it is really in error [07:42:05] General opinion, conclusion in document is OK [07:43:41] AP / JA agreed that there is no difference in ND uniqueness test for LL and global [07:44:51] TC: Change Section 5 (from Erik N) on screen [07:45:03] No, bottom section 6.1 [07:47:03] Top 6.1: this is important change [07:48:30] This is the text on modified EUI-64 [07:48:55] Chris Dearlove had comments on this [07:49:46] Ronald int Velt: drop text : from xxxx [07:50:56] CP: against title change [07:51:37] H.Rogge: Link local change is OK. [07:52:44] TC: we are about configuring router interfaces [07:52:57] UH: agreed on what TC said [07:53:16] CP: repeat objections to title change [07:55:20] TC: (can't type that fast...) [07:57:31] CP: I want to there is something... we gonna config routers, some routers dont route, so wilingness 0, routers that dont route... that's a lot more broken thahn current title. [07:57:53] TC: point is in IETF conflicting viezs: routning protocol, forwards packets, responds to RS... [07:58:01] TC: there are routers who do that [07:58:26] TC: MANET inface behaviors, wrong length, broken way to do things, roamaing [07:58:29] Teco Boot [07:58:50] TB: this doc now describes how router ifaces being configured, maybe some applies to hosts, for sure I know there are some proble,ms [07:59:11] TB: some aspects, e.g. rt protocol does not run, maye the node runs, the rt protocol si a host [07:59:36] TB: if you want to support hosts in a MANET and consider addresss for them. we could work for, but this doc has not annalywed this [07:59:45] TB: this doc is not about Host address mmodel [07:59:54] TB: I would like to change the title as currently suggested [08:00:04] TB: thats my oppinion. [08:00:19] Ryuji Wakikawa, the old title is not broken, just the new is more accurate [08:00:39] RW: this modif is not big (the upper on slide 13 white on black) [08:01:04] RW: I think this is just qualifictations; fix these issues only by RFC Editors, I am fine with them. [08:01:23] CP: return this issue: do we allo addresses for non routers by using mechanisms in AUTOCONF [08:01:40] CP: if specific to routing only then some may bring up solutions for hosts too, then out of scope [08:02:01] CP: what I thought... we should addressing model applicable to hosts as routers, clearer to mechanisms [08:02:12] CP:... [08:03:48] Ulrih HErberg is UH [08:04:06] UH: support title change ro Routers, with hosts respect.. we shouldnt delay the doc [08:04:14] JA: defined elsewhere how to do hosts [08:04:25] Samita Chakrabarti [08:05:04] SC: regarignt dhe link local addresses... another spin off, the change in ll addresses text is very much better, (the NEW of upper half ot slide 13 black on white) [08:05:32] TC: isolated IP hop, model, keep, then easy task. [08:05:53] Henning Rogge, [08:06:00] HR: I support the title change to Routers [08:06:13] HR: maybe another doc for Hosts [08:06:37] HR: dhcp or so, second draft, not all hosts need to be configured this way to communiate in a MANET [08:06:50] HR: restricting this doc to a router ifaces is a good idea [08:06:55] JA: call for voting in room [08:07:04] JA: in a moment [08:07:21] JA: will also send this on list [08:07:30] JA: consensus call [08:07:32] JA: is [08:08:27] JA: group says title change; dont do change to recommendation, we have top NEW text the "world is flat" change, slide 13 black on white [08:08:58] JA: say y or n [08:09:07] JA: y counted 16 [08:09:14] TC: 16 for, 0 against [08:09:28] JA: room oppinion is recorded as support for that [08:09:44] JA: instruct RFC editor to publish with changes [08:10:00] TB: option? text from? [08:10:10] TB: collect all in one place place [08:10:15] JA: hmmm... [08:10:32] JA: the consensus call will include a complete list of changes... [08:10:40] TC: dont before Monday [08:10:59] (there we go) [08:11:20] RW: 10 days or a week to finish the consensus call. [08:11:28] Rechartering !!!! [08:11:34] yeah... [08:11:47] What is the name of the cat ? [08:11:57] kitten [08:17:16] TB: this proposal is also using DHCP for address uniqueness and mgmt [08:17:35] TB: on a slide it says that address uniqueness guaranteed by central node alone [08:17:43] TB: unique token by request in nodes [08:18:04] TB: confused.. dhcpv6 adjusted or is dhcpv6 as usual? [08:28:19] TB: DHCPv6 could be used tho provide aligned addresses, this helps address compression in RFC5444 (packetbb) [08:28:51] Restart of DHCP server is problem, it starts to provide same addresses [08:29:10] DAD as being done today doesn't work on MANET [08:30:47] TC: when routing protocol provides DAD basis, this can help [08:32:21] Carlos: solution or problem space ?? [08:32:34] Problem space has affect on solutions [08:34:42] TC: Problem Space is important [08:35:24] JA: yes, problem space on distributed mechanisms are important [08:40:36] AP: Problem space and / or problem statement? [08:41:00] TC: yes, one before the other. Start with problem space analysis [08:41:21] TC: 2 in parallel [08:41:27] 1: DHCP based [08:41:39] 2: Problem space [08:42:22] Then recharter: problem statement, solution space analysis [08:43:23] Carlos: Problem space: look what is out there, and what problems are? [08:45:09] TC: DHCPv6 doesn't all. So problem space analysis is needed to get all issues [08:47:25] carlos: danger that analysis takes years [08:47:36] EB: agreed with Carlos, fears [08:47:57] EB: take the ROLL approach [08:49:23] HR: multiple gateways, so service discovery [08:49:40] UR: Yes, but now now discuss service discovery [08:50:08] Carlos: Is gw selection part of routing protocol [08:50:28] HR: end-node should select GW [08:51:48] TC: good discussions, part of problem analysis and should be in document [08:53:02] AP: lots of discussions on list on DHCP in MANET [08:53:40] HR: DHCP server no need for run on border [08:58:06] teco boot at the mic [08:58:34] market will go the other way, stateless [08:58:54] DHCPv6 useful for other parameters [08:59:41] let's not make the decision based on some existing IPv4 implementations [09:00:35] TC: do you think there is sufficient on problem space to make progress with the alternative solution soon? [09:00:39] Alex Petrescu leaves the room [09:01:13] Alex Petrescu joins the room [09:01:21] TB: re-iterates his problems with DHCP based [09:02:11] TB: for V6 we can make use of unique tokens of devices [09:03:09] RW: doing DHCP first does not rule out doing stateless later [09:03:28] TB: [09:03:43] JA: we are doing this in parallel [09:04:59] TB: proposes to replace first step of item 1 with SLAAC based address [09:05:23] nordmark joins the room [09:07:38] JA: DHCPv6 can be used without protocol extensions, SLAAC based approach requires protocol changes [09:16:12] Alex Petrescu leaves the room [09:16:24] Alex doesnt present, all time for Carlos [09:17:41] Teco, should you not rename your protocol, say to "Teco's Configuration Protocol" - it makes for a catchy acronym ;) [09:18:33] I'll rename to BOOT protocol [09:18:41] hehe, good one ;) [09:18:51] Oeps, BOOTP already exists [09:19:12] so TCP it is, then ;) [09:19:31] Why are TC's out there in OLSR? [09:19:49] Well....;) [09:25:54] TC: Goals are in charter, so likely this activity belongs to WG [09:28:11] TC: Document outdates fast, but have elements that keep value over time [09:32:09] teco.boot leaves the room [09:32:18] thomas.heide.clausen@gmail.com leaves the room [09:32:42] Ulrich Herberg leaves the room [09:41:13] velt leaves the room [09:51:39] nordmark leaves the room [11:00:52] nordmark joins the room [11:03:52] nordmark leaves the room [13:32:17] thomas.heide.clausen@gmail.com joins the room [13:47:23] thomas.heide.clausen@gmail.com leaves the room