[08:32:24] james woodyatt joins the room [10:38:54] tsavo_work@jabber.org/Meebo joins the room [11:01:01] paventhan joins the room [11:01:14] shengjiang joins the room [11:01:23] behcet.sarikaya joins the room [11:03:19] pselkirk@jabber.isc.org joins the room [11:03:19] swb joins the room [11:04:09] geir joins the room [11:04:17] geir leaves the room [11:04:22] t.k joins the room [11:04:30] Pierrick to present problem statement and current practices drafts [11:04:53] no slide numbers in his presentation [11:05:26] connection manager replaced by session manager [11:05:52] malc joins the room [11:07:09] now current practices draft [11:07:16] Jari on mic [11:08:25] Tomasz Mrugalski joins the room [11:08:44] Dave Thaler on mic [11:09:26] Bing.Liu joins the room [11:13:52] rvdp joins the room [11:15:02] Brian Carpenter on mic [11:15:14] Andrew joins the room [11:15:53] Ted Lemon on mic [11:15:54] Dave Thaler joins the room [11:16:47] Thomas is on mic [11:17:06] scott brim on mic [11:19:20] Jari on mic [11:22:09] Ted on mic [11:22:37] g.e.montenegro joins the room [11:24:00] Scott on mic [11:24:33] Pierrick back on his presentation [11:26:18] Marc Blanchett is on mic [11:28:21] arifumi joins the room [11:31:46] Tim Chown on mic [11:32:31] pselkirk@jabber.isc.org leaves the room [11:34:37] chairs on analysis draft [11:35:29] should we do analysis draft? [11:35:57] Ted thinks analysis draft is similar to current practices draft [11:36:14] Margareth : no [11:36:41] tsavo_work@jabber.org/Meebo leaves the room [11:37:15] Brian on mic [11:40:26] Jari: we need vonlunteers to do the work [11:41:09] Margaret: who wants it? [11:41:26] Margaret: strong consensus to do the work [11:41:51] Margaret: if interested in doing the draft, send mail to the chairs [11:42:27] Teemu on dns server selection [11:42:45] slide 2 [11:42:53] slide 3 [11:44:15] slide 4 [11:46:16] cjbernardos joins the room [11:46:19] slide 5 (guessin from remote) [11:46:30] yes slide 5 [11:46:34] now 6 [11:47:34] Simon Perreault joins the room [11:47:56] andy sullivan (?) on mic [11:48:56] andrew. [11:50:03] My mother still calls me Andy sometimes :) [11:50:32] ops, sorry for that :) [11:50:58] no worries. I haven't been called it in so long that I usually don't respond if someone calls me Andy. [11:51:55] cjbernardos leaves the room [11:54:24] pselkirk@jabber.isc.org joins the room [11:56:27] Andy Sullivan on mic [11:56:48] andy you're the best [11:56:59] slide 7 [11:57:41] love, mom [11:58:18] (sorry that was intended to be private) [11:58:30] hehehe [12:12:29] Dapeng on mic [12:12:56] Andy on mic [12:16:07] james woodyatt leaves the room [12:16:26] I don't buy any real scenario where there's one interface that wants to configure disjoint DNS servers for different namespaces. [12:16:37] +1 [12:17:22] If you want to do that, you have to deal with hosts that don't yet do whatever this new option would be, and to do that you'd hand out the IPs of a DNS server inside that ISP that then uses the more detailed information when it does recursive queries for example. [12:17:35] and once you have that, you don't need DNS IPs in the DNS suffix list. [12:18:12] Hui: ten minutes for eachpresentation [12:18:43] Woj on DHCPv6 route option draft [12:18:47] slide 2 [12:19:41] slide 4 [12:20:28] @Dave: there are these cases now, though. [12:21:10] In doing database consulting, for instance, it's happened to me that I had more than one VPN open to different customers while working on the same problem for different people at the same time. [12:21:30] You still have one DNS server per interface then. [12:21:44] slide 5 [12:21:59] yes, but the host is usually stupid and doesn't have any clue which DNS server to ask [12:22:09] so you get NXDOMAIN because it just picks one [12:22:41] the DHCP option here gives you a clue: "no, ask _him_ for example.com and _her_ for otherexample.com" [12:22:46] Ted on mic [12:24:21] (my case: I had to go to db1.company1.example and db1.company2.example. If I ask my own DNS server, they're both Name Error, if I ask company1's server for db1.company2.example, I still get Name Error, and so on.) [12:24:25] @Andrew: you don't need IPs in the suffix list to do that. [12:24:44] Margaret on RtgDir review [12:24:46] you have a suffix list per interface, and a DNS server list per interface [12:24:46] Ah, I get that point. Yeah. [12:25:20] Margaret will communicate offline and get back to the group [12:25:40] Dapeng on API extensions [12:26:51] no page numbers [12:27:08] API definition slide [12:27:33] Address selection API slide [12:28:12] danny joins the room [12:28:33] danny leaves the room [12:29:23] route selection [12:29:30] interface selection [12:29:49] RFC5014 does not solve his problem of address selection API ? [12:30:00] what's the difference between the setsockopt() option for source address vs. bind() ? [12:30:15] I agree, didn't make sense to me either [12:30:35] interface information [12:30:53] DNS selection [12:30:54] and gethostbyname() is deprecated [12:30:54] I think this is one of the "big" questions. IN an age when we're trying to wean apps away from knowing too much about topology-dependent info (i.e. addresses) do we really want to encourage this? [12:31:02] i understand there is a need for choosing address by its nature/type. it can be achievable by rfc5014. [12:31:43] notification [12:34:48] Dave on abstract vs concrete API [12:35:26] Margaret: will do abstract API [12:36:27] rvdp leaves the room [12:39:46] X on mic: overlap with multi path TCP WG mptcp [12:40:44] Alan Roke [12:41:24] Jouni on RA offload draft [12:41:35] background slide [12:42:04] so we have two ways to specify source address: bind() and IPV6_PKTINFO [RFC3542] (plus the non-standard IP_PKTINFO for IPv4) [12:49:08] Gang on happy eyeballs extension [12:49:33] overview slide [12:50:35] Algorithm slide [12:51:09] pselkirk@jabber.isc.org leaves the room [12:52:06] example slide [12:52:52] additional considerations [12:56:10] behcet.sarikaya leaves the room [12:59:03] swb leaves the room [12:59:39] Andrew leaves the room [13:02:34] Simon Perreault leaves the room [13:03:49] simon.perreault joins the room [13:04:19] t.k leaves the room [13:04:38] t.k joins the room [13:04:59] shengjiang leaves the room [13:05:47] paventhan leaves the room [13:05:55] t.k leaves the room [13:06:43] g.e.montenegro leaves the room [13:08:21] simon.perreault leaves the room [13:08:28] simon.perreault joins the room [13:09:50] Tomasz Mrugalski leaves the room [13:10:52] t.k joins the room [13:10:52] t.k leaves the room [13:11:24] Tomasz Mrugalski joins the room [13:12:23] simon.perreault leaves the room [13:15:41] simon.perreault joins the room [13:16:22] simon.perreault leaves the room [13:16:42] Dave Thaler leaves the room [13:17:13] Dave Thaler joins the room [13:17:15] g.e.montenegro joins the room [13:20:26] g.e.montenegro leaves the room [13:24:07] Dave Thaler leaves the room [13:27:46] arifumi leaves the room [13:38:04] Tomasz Mrugalski leaves the room [13:39:21] Bing.Liu leaves the room [13:50:11] Tomasz Mrugalski joins the room [13:50:18] Tomasz Mrugalski leaves the room [14:39:39] malc leaves the room [18:24:29] james woodyatt joins the room [18:24:34] james woodyatt leaves the room