[00:35:09] Ue joins the room [00:48:04] Ue leaves the room [01:03:06] Dave Thaler joins the room [01:03:15] Dave Thaler has set the subject to: MIF IETF 79 [01:03:55] behcet joins the room [01:04:06] bernie joins the room [01:04:07] arifumi joins the room [01:04:08] Zhen Tsao joins the room [01:04:15] Ue joins the room [01:04:30] Dave Thaler is now known as Dave Thaler (Jabber Scribing) [01:04:49] agenda bashing [01:05:04] dudisaki joins the room [01:05:29] agenda slides at http://www.ietf.org/proceedings/79/agenda/mif.txt [01:05:34] thomson joins the room [01:05:41] oops http://www.ietf.org/proceedings/79/slides/mif-6.ppt [01:05:59] don't know if PDF is online [01:07:50] slide: draft-ietf-mif-problem-statement IESG review (no slide numbers) [01:07:56] thomson is now known as Tomasz Mrugalski [01:08:21] now on slide 8 in the deck (no number on slide) [01:09:00] Tomasz Mrugalski leaves the room [01:09:52] tomasz.mrugalski joins the room [01:09:56] ted lemon: reaction to Ralph's feedback... agree... want comments now? [01:10:26] marc blanchet: we agree with the comments, revising doc [01:10:58] marc: need feedback on whether revision sufficiently addresses the comments [01:11:33] slide: draft-ietf-mif-current-practices-02 AD Review [01:11:39] angga joins the room [01:12:19] margaret: if we don't get more info from contributors we sent email to, we'll remove those sections in a week or so [01:12:54] shinmiyakawa joins the room [01:13:04] marc: i wrote macos6 part but based on my own testing, agree it's not good but better for the implementers to supply text [01:13:35] margaret: all the apple ones, blackberry, android, nokia, and a couple of the microsoft sections incomplete [01:14:10] gigix73 joins the room [01:14:19] (?#1): move to appendix instead of removing [01:14:49] Jari Arkko: partial information is not very useful, cut it if don't have full picture [01:15:37] Andrew Sullivan: less information after we're done seems to be the wrong direction if looking for understanding of current practices, so opposed to cutting [01:16:03] Marc: disagree. At least for my text people could conclude the wrong thing. Either verify it's correct or delete [01:16:37] Margaret: some sections are quite weak and perhaps incorrect and written in hope implementers would supply correct info [01:16:52] danwing joins the room [01:17:00] Andrew: then document it as gossip and say we don't understand how it really works [01:18:43] Jari: I don't disagree, but some text gives wrong picture [01:19:09] Margaret: maybe create a section for the gossip ones with appropriate warning text? [01:19:14] wmhaddad joins the room [01:19:52] Jari: no, have to remove confusion. Can make it clear what parts you do understand, need actual text changes, doesn't matter what section [01:20:18] euijong hwang joins the room [01:20:34] Ue leaves the room [01:20:46] Ue joins the room [01:21:04] margaret summarizing result - we either made it clear or remove it [01:22:10] Ted: there's another draft on current practice with a better format. it systematically identifies areas to analyze (no data in the draft though). Point is to standardize the set of problems to analyze, then it's clear and won't mislead people [01:22:24] satoru.matsushima joins the room [01:23:06] next presentation, Teemu on split DNS http://www.ietf.org/proceedings/79/slides/mif-0.ppt [01:23:16] slide 2 [01:23:19] Joonhyung Lim joins the room [01:24:03] Joonhyung Lim leaves the room [01:24:07] Joonhyung Lim joins the room [01:24:26] slide 3 [01:25:31] [note: (?#1): == Bernie Hoeneisen] [01:25:40] slide: changes since -03 [01:26:53] slide 5 [01:27:21] Jari: lack of this option means it's default nameserver, yes? [01:28:32] Teemu: this doesn't deprecate existing option [01:29:27] Dave Thaler: agree with jari [01:29:56] Teemu: will update draft [01:30:43] Ted Lemon: this seems like a really bad idea, the security implications are mind-bogling. The security considerations section just says it's insecure and it's missing mitigations. You should require use of DNSSEC [01:31:21] Teemu: this doesn't make things worse [01:32:00] Ted: makes it possible for attacker to make it look secure by not taking over most names, just a target name not secured by DNSSEC [01:32:10] Teemu: like RFC 4191 for routing [01:32:28] (er, for router selection rather) [01:33:28] Ted: if you use DNSSEC you don't need this, client just queries multiple DNS servers and will get the correct answer from the correct one [01:33:51] discussion about whether can have local names in DNSSEC... [01:34:26] Ted: notion of preferred interface doesn't make sense [01:35:20] Stuart Cheshire: DNSEC will give you authoritative non-existence from one, and authoritative existence from another network [01:36:01] Ted: then could argue DHCP option to tell client to go authenticate since at a hotspot [01:37:18] (?#2): repeating Teemu's point about more general issue not specific to DNS [01:37:56] Teemu: might trust only what you get over a trusted interface like VPN [01:38:09] Ted: would be a lot less upset if it was restricted to trusted interfaces [01:39:22] (?#2 = Gaetan?) [01:39:53] Jari: there are some cases where user might trust one network more like trusting ietf-1x more than the hotel network or something [01:41:16] Ted: can dynamically learn which interface to use by using whichever returns DNSSEC validated positive answer [01:41:23] Hyon-Young Choi joins the room [01:42:11] Stuart: ted said it returns NXDOMAIN or drops or what? (Ted: drops) [01:42:58] [note: ?#2 = Gaetan Feige] [01:43:12] satoru.matsushima leaves the room [01:43:33] Stuart: you're describing a new state of zones in DNS. things where want to drop and let client timeout [01:45:08] Dave Thaler: just tried it and see NXDOMAIN right now in this case [01:47:08] fujiwara joins the room [01:48:21] Olafur: most OSs assume there's just one DNS, and that would be nice but not world today. main issue is need to tie DNS and use of addresses to same interface. Want to use same DNS server as for network will use for communication to the answer. and dnssec doesn't help everything. [01:48:38] nobody states his name on mic, including me X( [01:49:21] satoru.matsushima joins the room [01:49:47] Andrew: looking for scope within which you prefer an interface. don't know how to figure out what network context one is in. [01:49:49] shtsuchi joins the room [01:50:29] Gaetan: also may get different answers based on geographical location even if both same org [01:51:33] Andrew: many TLD servers have answer dependent on what routing you have, that's how anycast works. [01:52:41] Paul Carlos-..: these problem spaces interact [01:53:02] Hui (as wg chair): we won't modify server behavior here [01:53:18] slide 6 [01:54:17] ogud joins the room [01:55:24] Andrew joins the room [01:55:29] slide 7 [01:56:34] duan.chen joins the room [01:57:26] updated draft after maastricht but list was silent [01:57:37] Hui asking Ted to commit to sending review to list [01:57:54] Jari: maybe organize a conf call in a month to discuss [01:57:58] Teemu: will set up [01:58:11] ACTION ITEM: Ted to send review to list [01:58:19] ACTION ITEM: Teemu to set up conf call next month [01:58:29] slide 8 [01:59:12] slide 9 [01:59:39] Hui asking to update draft before conf call [01:59:56] gigix73 leaves the room [02:00:08] ACTION ITEM: Teemu to update draft in response to feedback before conf call [02:00:30] next presentation [02:00:55] slide 3 of DHCP route option draft slides [02:01:00] behcet leaves the room [02:01:01] http://www.ietf.org/proceedings/79/slides/mif-1.ppt i think [02:01:13] Sarah Tarrapey joins the room [02:02:24] slide 4 [02:02:46] any comments? [02:03:07] how many have read? maybe 30? [02:03:18] how many to adopt? maybe 15 [02:03:26] how many to not adopt? 0 [02:03:40] will confirm on mailing list [02:03:55] next up Yuri on API [02:04:02] http://www.ietf.org/proceedings/79/slides/mif-2.ppt [02:05:17] g.e.montenegro joins the room [02:05:24] slide 2 (no slide numbers): MIF API Requirements [02:05:57] slide 3: General Considerations [02:06:19] Ue is now known as Tatsuji Ue [02:07:34] slide 4: Scope... [02:07:59] ogud leaves the room [02:07:59] ogud joins the room [02:09:08] slide 6: Examples... [02:09:19] er 5: Examples... [02:10:49] slide 6: MIF API Open Issues [02:13:12] Ted: mp draft about dtls. We ran into needing to talk about groups of interfaces not just individual interfaces [02:16:59] ogud leaves the room [02:17:08] ogud joins the room [02:18:33] tanizawa joins the room [02:19:12] above Ted is Ted Hardie (not Ted Lemon) [02:19:22] Ted Lemon: want to see more discussion on list, haven't seen any [02:19:51] Dave Thaler: socket option question should be directed to Austin Group, and they'll probably say wrong question it should not be socket options when you need type safety [02:19:59] slide 8 [02:20:31] gigix73 joins the room [02:20:56] (The Austin Group owns the POSIX standard) [02:21:21] next is http://www.ietf.org/proceedings/79/slides/mif-4.pptx [02:21:41] Gaetan Feige presenting [02:21:43] bad, this is not consistent with the order on the agenda [02:22:56] (it's consistent with what's in the agenda slides at http://www.ietf.org/proceedings/79/slides/mif-6.ppt) [02:23:22] slide 2 [02:24:34] S7365468EEE7F6 joins the room [02:29:56] slide 3 [02:30:05] (3: Problem Statement II) [02:31:04] S7365468EEE7F6 leaves the room [02:31:24] Teemu: propose removing conflict text from dns draft completely and say it's a more general problem [02:34:38] slide 4 [02:36:44] slide 5 [02:37:34] slide 6: Summary [02:37:54] behcet.sarikaya joins the room [02:38:55] slide 7: possible scope [02:41:16] Gabor Bajko: WiFi Alliance has ongoing work on the L2 problem on slide 2 [02:41:55] Gabor: on slide 4, also issue when have link-layer connectivity but broken network-layer connectivity, also looking at in WFA [02:42:01] satoru.matsushima leaves the room [02:42:53] Gaetan: don't think it has relation to routing part like which DNS servers to use. L2 part is in WFA but how handle that at IP layer is not in WFA. and same in 3GPP [02:43:37] Margaret: need some notion of up to send IP packets but not for general internet service [02:44:33] satoru.matsushima joins the room [02:44:51] Gabor: WFA is just trying to solve problem of knowing which wifi network to connect to [02:46:37] Yuri: different scopes of session. for some need protocol interactions, some need apis [02:48:52] Yuri: are virtual interfaces standardized in how they're implemented? [02:49:03] satoru.matsushima leaves the room [02:49:44] Gaetan: linux has a per-interface routing in addition to general routing [02:52:08] Gaetan: we don't want every app to reimplement policy control [02:52:18] Yuri: take offline, there are other use cases [02:53:44] Andrew Sullivan: circularity argument [02:54:21] next is Chen Cao on http://www.ietf.org/proceedings/79/slides/mif-5.pdf [02:55:00] slide 2 [02:55:05] er Zhen Cao [02:55:07] gigix73 leaves the room [02:55:27] slide 3 (no slide numbers): An Example: FTP [02:57:37] slide 4 [02:58:31] Yuri: what about solutions from shim6, sctp, etc? [03:03:42] Dave Thaler: strong host model doesn't have this problem, and hosts today use strong host. XP IPv4 (but not IPv6) was last major OS version to use weak host model. [03:04:40] margaret speaking about what WG should do in general [03:04:47] about ongoing connections [03:05:36] should we add something to problem statement? (not asking now about whether to produce a solution) [03:05:59] ogud leaves the room [03:06:12] Simon Perrault: reformulate problem in terms of host models - weak or strong [03:06:13] wmhaddad leaves the room [03:06:22] ogud joins the room [03:06:44] margaret: what about problem of networks up for service but not in any meaningful way [03:07:22] Juan Carlos: in favor of amending problem statement doc [03:10:02] Ted Hardie: good idea but scope carefully. If deal with interface groups, then notions of add/delete from such a group is special, and exposing is useful. But this is wrong group for saying how sessions are constructed on top of. [03:10:34] Ted Hardie: they may want to express a desire for weak/strong host within an interface group [03:11:59] ogud leaves the room [03:13:18] asking group- how many think should work on? [03:13:27] unanimously for [03:13:39] about 12-15 for [03:14:11] ACTION ITEM: Marc to add text [03:14:58] next is MIF current practice analysis [03:15:01] not updated since IETF 78 [03:15:21] http://www.ietf.org/proceedings/79/slides/mif-3.pdf [03:16:10] ogud joins the room [03:17:20] margaret (as individual): would like to wait since will have more problems to analyze [03:19:29] people shouldn't assume we don't do solutions until we know all problems though [03:19:56] WE ARE DONE [03:20:08] Tatsuji Ue leaves the room [03:20:15] Hyon-Young Choi leaves the room [03:20:16] tomasz.mrugalski leaves the room [03:20:17] Sarah Tarrapey leaves the room [03:20:29] duan.chen leaves the room [03:20:30] danwing leaves the room [03:20:37] bernie leaves the room [03:20:39] Dave Thaler (Jabber Scribing) leaves the room [03:20:46] thomson joins the room [03:20:48] Zhen Tsao leaves the room [03:20:59] angga leaves the room [03:20:59] Joonhyung Lim leaves the room [03:21:00] euijong hwang leaves the room [03:21:05] g.e.montenegro leaves the room [03:21:29] ogud leaves the room [03:21:36] dudisaki leaves the room [03:21:38] shinmiyakawa leaves the room [03:26:56] shtsuchi leaves the room [03:26:59] Andrew leaves the room [03:27:51] shtsuchi joins the room [03:28:01] behcet.sarikaya leaves the room [03:28:06] shtsuchi leaves the room [03:34:24] bernie joins the room [03:36:02] tanizawa leaves the room [03:39:08] Andrew joins the room [03:39:29] arifumi leaves the room [03:39:42] danwing joins the room [03:39:58] Andrew leaves the room [03:40:34] thomson leaves the room [03:40:42] bernie leaves the room [03:41:18] danwing leaves the room [03:44:23] arifumi joins the room [03:46:59] arifumi leaves the room [03:49:31] arifumi joins the room [03:57:40] angga joins the room [04:11:42] arifumi leaves the room [04:18:06] thomson joins the room [04:19:09] danwing joins the room [04:25:23] Joonhyung Lim joins the room [04:30:26] angga leaves the room [04:39:45] bernie joins the room [04:42:26] thomson leaves the room [04:49:30] Joonhyung Lim leaves the room [04:51:46] shtsuchi joins the room [04:51:58] shtsuchi leaves the room [04:52:20] thomson joins the room [04:55:02] Joonhyung Lim joins the room [04:56:10] danwing leaves the room [05:02:32] g.e.montenegro joins the room [05:03:17] g.e.montenegro leaves the room [05:05:21] danwing joins the room [05:06:26] danwing leaves the room [05:06:36] wmhaddad joins the room [05:06:52] danwing joins the room [05:15:00] Joonhyung Lim leaves the room [05:17:36] thomson leaves the room [05:22:41] danwing leaves the room [05:33:55] bernie leaves the room [05:38:19] bernie joins the room [05:42:28] bernie leaves the room [05:44:08] bernie joins the room [06:10:04] wmhaddad leaves the room [06:25:43] bernie leaves the room [06:45:35] bernie joins the room [06:47:37] bernie leaves the room [06:49:37] bernie joins the room [07:19:20] bernie leaves the room [07:26:28] bernie joins the room [07:28:25] bernie leaves the room [07:30:32] bernie joins the room [08:51:48] fujiwara leaves the room [08:53:47] bernie leaves the room [09:50:07] bernie joins the room [09:50:56] bernie leaves the room [09:50:59] bernie joins the room [09:52:13] bernie leaves the room [09:52:57] bernie joins the room [11:51:15] bernie leaves the room