[05:03:37] jlcjohn joins the room [06:52:36] danwing joins the room [06:54:42] James Galvin joins the room [06:56:21] smx joins the room [06:56:33] smx is now known as sm [06:56:33] sm is now known as smx [06:56:33] smx is now known as sm [06:56:46] sm leaves the room [06:56:46] smx joins the room [06:57:03] smx is now known as sm [06:57:03] sm is now known as smx [06:57:25] smx leaves the room [06:58:19] sm joins the room [06:59:50] Joe Hildebrand joins the room [07:00:23] Chris Newman joins the room [07:02:30] Is anybody getting the audio stream (not muzak, that is)? [07:04:09] Glenn Parsons joins the room [07:04:52] sal joins the room [07:05:02] cyrus joins the room [07:05:11] is the audio stream worked out? [07:05:13] stpeter joins the room [07:05:35] stpeter leaves the room [07:05:51] Randall Gellens joins the room [07:05:59] eburger joins the room [07:06:11] Let's get this party started! [07:06:35] stpeter joins the room [07:06:36] They say 2mins. [07:06:46] stpeter shall be scribing [07:06:48] (on the audio stream) [07:06:50] Lisa Dusseault joins the room [07:07:39] ylafon joins the room [07:07:47] Well, the muzak stopped, anyway... [07:08:03] no Muzak™??? [07:08:12] Atarashi Yoshifumi joins the room [07:08:13] or maybe Muzak® [07:08:26] resnick joins the room [07:08:37] ah, resnick is here, the meeting can start [07:08:43] heh [07:08:54] we are starting [07:08:54] OMG! This room is not anonymous! [07:08:59] How will I live? [07:09:02] agenda bashing? [07:09:26] no bashing [07:09:47] tonyhansen joins the room [07:09:54] Alexey reports on NomCom [07:10:14] tonyhansen has set the subject to: AppArea General Meeting, IETF 75 [07:10:20] stpeter leaves the room: Disconnected: connection closed [07:10:57] stpeter joins the room [07:11:01] tlyu joins the room [07:11:04] Alexey says "Get off your butt and talk to the nomcomm to comment *and* volunteer for AD!!!!" [07:11:10] hmm, my pre-alpha client crashed [07:11:46] stpeter finds a different client [07:12:10] FTP64 presentation by Iljitsch van Beijnum [07:12:53] FTP through v4-v6 translators [07:13:28] client is on v6, server on v4 [07:13:55] RFC 959 assumes v4 address+port [07:15:35] if all entities use EPSV FTP it works fine [07:15:48] Dan Wing did some testing... [07:15:54] 65% working EPSV [07:16:03] 28% 50x commant not implemented [07:16:09] 6% timeout [07:16:16] 1% 425 error [07:16:43] also some cases where passive IPv4 doesn't work [07:17:30] not a lot of client testing [07:17:45] only client that doesn't use EPSV is Windows command line client [07:17:53] possible solutions... [07:17:58] Just shocked the Windows' client doesn't work... [07:18:07] 1. updates servers to suppose EPSV [07:18:22] 2. Update clients to fall back on PASV (hack it into IPv6) [07:18:33] 3. Implement app layer gateway [07:18:47] wolfgang.beck01 joins the room [07:19:09] suggestion: mandate (1) and (2), describe (3) [07:19:40] psavola joins the room [07:20:02] haa joins the room [07:20:46] [see slides] [07:21:34] why didn't they bring us beer? Dublin rocked. [07:21:41] haa leaves the room [07:22:12] heh [07:22:28] hmm, FTP three-way? [07:22:37] shirleyhm joins the room [07:22:38] kinky [07:22:57] are the suggestions in this presentation reasonable? [07:23:19] haa joins the room [07:23:27] Barry Leiba at the mic [07:24:01] BL: I like how this has worked out, ALG is problematic but described as a last resort so OK with me [07:24:02] dmeyer joins the room [07:24:15] Chris Newman leaves the room [07:24:28] CNNIC-AA5684158 joins the room [07:24:49] Q: will this introduce changes to the FTP spec? [07:25:07] A: this would provide updates to 959 etc. [07:25:14] the ALG stuff would be informational [07:25:20] [07:25:21] CNNIC-AA5684158 leaves the room [07:25:28] next... [07:25:29] CNNIC-AA5684158 joins the room [07:25:38] dmeyer leaves the room [07:26:15] Michael Tüxen -- SCTP for the application developer [07:26:23] why use SCTP? [07:26:31] current choices are TCP or UDP [07:26:42] CNNIC-AA5684158 leaves the room [07:26:52] either totally unordered and unreliable, or totally ordered and reliable [07:26:56] stark choice [07:27:00] psavola leaves the room: Replaced by new connection [07:27:01] psavola joins the room [07:27:02] Alissa Cooper joins the room [07:27:28] SCTP gives you an alternative in the middle [07:27:41] SCTP might give you better performance [07:27:48] dmeyer joins the room [07:27:50] also, some services provided only by SCTP [07:27:59] SCTP features.... [07:28:06] - connection oriented [07:28:15] - message oriented, not byte oriented [07:28:34] - supports fragmentation and reassembly of large messages [07:28:39] - congestion control [07:28:42] - flow control [07:28:55] - parameters are widely configurable per association [07:29:01] - runs on v4 and v6 [07:29:09] - support for multihoming [07:29:24] (endpoint not identified by one IP+port) [07:29:32] can bind both v4 and v6 at the same time [07:29:47] CNNIC-AA5684158 joins the room [07:29:49] addresses are negotiated during setup of association [07:30:10] addresses can be changed dynamically during the lifetime of an association [07:30:16] (SCTP extension defines this) [07:30:29] can use multiple addresses for redundancy [07:30:39] load sharing is possible but not yet standardized [07:31:04] originally developed for sigtran [07:31:14] message ordering... [07:31:24] SCTP has a number of streams in each direction [07:31:38] e.g., can have 10 streams in one direction and 50 in the other or whatever [07:31:49] messages are ordered only within a particular stream [07:32:01] streams are unidirectional message channels [07:32:16] sender can request unordered sending as well [07:32:30] intDance joins the room [07:32:43] message reliability... [07:33:05] SCTP has extension for partial reliability [07:33:14] if not used, messages sent reliably by default [07:34:08] can use this PR-SCTP extension to limit retransmission etc. [07:35:23] SCTP built in to FreeBSD, Linux, Solaris [07:35:35] loadable kernel extension for Mac OS X [07:35:45] userland stack that runs on Windows [07:36:22] kernel impls use a socket API [07:36:41] see draft-ietf-tsvwg-sctpsocket [07:36:53] easy to port TCP and UDP apps to SCTP [07:37:04] L41587M21F40C20 joins the room [07:37:38] I wonder if these entities like L41587M21F40C20 are anonymous users... [07:37:39] haa leaves the room [07:38:14] Next up: SCTP CHARGEN [07:38:44] sorry, I dont know why it displays that [07:39:03] no worries [07:39:59] Michael shows some code examples of changing a UDP application or TCP application to SCTP [07:40:49] deployment considerations... [07:41:04] TLS defined in draft-ietf-tsvwg-dtls-for-sctp [07:41:14] NAT traversal in two I-Ds [07:41:26] (UDP tunnelling and SCTP-aware NAT) [07:41:41] example: HTTP over SCTP [07:42:05] no change to HTTP protocol itself [07:42:13] HTTP has alway maintained it is not tied to the transport. [07:42:14] use multiple streams for multiple requests [07:42:27] serverr sends response on corresponding stream [07:42:44] patched Apache and Firefox [07:43:18] demo [07:43:23] (running code!) [07:45:08] remote audio seems to be functioning (with LOTS of echo) [07:45:30] jcjohn: we had that problem last time, too [07:45:31] wow! the echo stopped! [07:45:41] HTTP/SCTP -- define SRV record? [07:45:45] new URI? [07:46:03] Barry Leiba at the mic [07:46:18] BL: article forthcoming in IEEE Communications [07:46:18] fear the combinatorial explosion of uri schemes :) [07:46:28] Dave Crocker at the mic [07:46:44] DC: question about deployment and availability [07:48:00] Ted Hardie: do we need to do any hacks to make this work on existing networks? [07:48:40] Cyrus Daboo at the mic [07:48:53] L41587M21F40C20 leaves the room [07:48:59] CD: I see this as more beneficial for server-to-server connectivity (XMPP?) [07:49:28] next presentation [07:49:31] Answer to Ted: you must run over udp in order to get through existing nats [07:49:45] L41587M21F40C20 joins the room [07:50:38] I missed the presentation title and presenter [07:51:17] So did I [07:51:24] Andrew Yourtchenko [07:51:32] Successful Introduction of New Technology to HTTP [07:51:39] Lisa: thanks [07:53:35] for v4 and v6, set a preference and increment or decrement depending on which fails or succeeds [07:53:45] similar methodology for SCTP vs. TCP [07:54:33] L41587M21F40C20 leaves the room [07:54:45] tested with HTTP [07:54:56] L41587M21F40C20 joins the room [07:55:01] perhaps useful for multipath TCP, IM [07:55:18] no questions at the mic [07:56:00] next presentation [07:56:16] Alexey Melnikov on Vendor Name Registry [07:56:56] existing registry established by RFC 2244 for ACAP [07:57:24] used by RFCs 5257, 5258, 5464 (IMAP extensions) [07:57:37] could also be reused by vCard update work [07:57:59] psavola leaves the room: Replaced by new connection [07:57:59] psavola joins the room [07:58:04] draft-cridland-acap-vendor-registry decouples the registry from ACAP [07:58:37] open issues.... [07:58:49] OK to restrict this to US-ASCII subset? [07:58:56] disallow SP? [07:59:36] some at the mic whose name I didn't catch [08:00:08] psavola leaves the room: Replaced by new connection [08:00:08] psavola joins the room [08:00:14] why do we need another registry instead of private entity registry? [08:01:07] Alan at the mic (last name?) [08:01:26] psavola leaves the room: Replaced by new connection [08:01:26] psavola joins the room [08:02:09] Alan: differentiate between vendor name and private enterprise number [08:02:37] rababy joins the room [08:02:38] point from the mic: vendor name changes [08:02:51] Cyrus Daboo: might need long form name and short form name [08:03:31] point at the mic: trademark dispute policy? [08:03:42] Chris Newman to reply [08:04:24] CN: two-fold answer... (1) first come, first served -- if courts tell IANA to follow some policy, it will [08:04:56] (2) grandfather existing names (?) [08:05:06] fallback: if we get squatters, use numbers [08:05:11] Larry Masinter at the mic [08:05:50] LM: inherent problem when registry entities live longer than the things being registered [08:07:36] next presentation [08:07:46] Cyrus Daboo on time zones [08:08:02] this is a follow-up to discussion at SF meeting [08:08:17] key issues... [08:08:29] - we want a secure reliable way to get at time zone data [08:08:48] - Olson database + zoneinfo the main current solution [08:08:58] - driver is calendaring and scheduling work [08:09:34] (see SF presentation for details about the problem statement) [08:11:39] bhoeneis joins the room [08:12:11] proposed solutions include... [08:12:39] IANA registry [08:12:50] TZ service [08:13:00] plan of succession for Olson data [08:13:12] digital signatures for TZ data [08:13:19] apolitical TZ data [08:13:20] psavola leaves the room: Replaced by new connection [08:13:21] psavola joins the room [08:14:35] Pete Resnick at the mic [08:15:16] PR: would we want to take this up with the ITU given the political aspects etc.? [08:16:05] psavola leaves the room: Replaced by new connection [08:16:05] psavola joins the room [08:16:18] Cyrus: the registry would be a registry of publishers, not the data itself [08:16:22] John Klensin at the mic [08:16:36] JK: potential for turf battles in the UN as well [08:17:07] Huh. Could we put all of this in the DNS? [08:17:14] Ted Hardie: disjoint here? multiple namespaces and namespace roots? [08:17:33] psavola leaves the room [08:17:54] TH: why not ask Olson to choose a successor? [08:18:12] Joe: sure, you could put this in the DNS. That's just a publication mechanism, thoguh [08:18:44] That doesn't solve all of the problems this involves [08:19:19] use UUIDs? [08:19:24] intDance: it does solve the disjoint namespace problem, though, with explicit delegations. [08:19:45] If you are willing to declare a single root for this, you can solve this in a wide variety of ways. [08:20:36] But the "well-known names" but multiple publishers implies that getting agreement on the root of that namespace might be difficult [08:20:56] another alternative: define URN scheme? [08:21:04] (er, not "scheme") [08:21:05] Unless Olson annoints someone, and that is good enough [08:21:17] Cyrus: we do want a registry of publishers [08:21:44] Cyrus: promote common APIs and libraries [08:21:51] Internet-wide service would be helpful [08:21:55] Barry Leiba at the mic [08:22:13] yao joins the room [08:22:15] BL: there are problems with naming here that will never be solved [08:23:08] Cyrus: those problems are typically implementation matters [08:23:40] intDance leaves the room [08:23:42] Dave Crocker at the mic [08:24:37] yao leaves the room [08:24:55] DC: if Chicago moves to Eastern Time, that's a change of TZ location -- if they change the label for Central Time to Chicago Time then that is a different problem [08:25:20] Alissa Cooper leaves the room [08:25:22] Ted Hardie at the mic [08:25:28] we could just let the registrar for Tanzania decide on authoritative delegations. [08:25:30] TH: why not solve this with the DNS? [08:25:44] (TH channeling question from the Jabber room) [08:27:14] Cyrus: we looked at the DNS for this, but something like that would be appropriate for a service not the registry of publishers [08:27:21] Alissa Cooper joins the room [08:27:50] Cyrus: where do we need to go with this work? [08:28:27] intDance joins the room [08:28:32] e.g. est.tz [08:29:20] "Link header" and "Host Metadata for the Web" drafts [08:29:31] link header in Last Call (is that WGLC?) [08:30:01] Mark will be here starting Tuesday night [08:31:41] stpeter talking about oauth [08:31:55] see the list for the breakfast bof location [08:32:27] Barry Leiba on OGPX [08:32:46] OGPX = open grid protocol [08:32:52] BOF at SF was too broad [08:33:14] BL: intro and requirements I-D [08:33:25] (look in datatracker with "ogp" string) [08:33:33] L41587M21F40C20 leaves the room [08:33:34] second BoF at Stockholm, smaller scope [08:33:42] L41587M21F40C20 joins the room [08:33:47] 6lowap [08:33:53] richard.barnes joins the room [08:34:09] look at application protocols for small nodes -- embedded nodes etc. [08:34:17] L41587M21F40C20 leaves the room [08:34:24] packet size max 128 bytes [08:34:26] L41587M21F40C20 joins the room [08:34:30] smart grids, home networks, etc. [08:34:41] which preso are we on? [08:34:42] 6LowPAN WG [08:34:46] ROLL WG [08:34:52] (INT and RTG) [08:34:57] time for apparea WGs? [08:35:21] appropriate to use HTTP, XMPP, or some other application protocols? or new work? [08:35:47] Bar BoF on Tuesday evening @ 18:30 in Room 202 [08:35:58] Larry Masinter on IRI [08:36:17] Bar BoF on Thursday evening in same room at same time [08:36:22] ray joins the room [08:37:37] correction: starts at 19:30 on Thursday [08:37:59] Alexey points to http://trac.tools.ietf.org/area/app/trac/wiki/BarBofs/IETF75 [08:39:23] L41587M21F40C20 leaves the room [08:39:32] CNNIC-AA5684158 leaves the room [08:39:38] topic: move SPF / SenderID to Historic? [08:39:54] Dave Crocker at the mic [08:40:04] DC: these have substantial deployment [08:40:27] Pete Resnick: asked "why not"? [08:40:27] L41587M21F40C20 joins the room [08:40:41] PR: being in use is different from "being in use usefully" [08:41:01] PR: are they in fact working? [08:41:24] Barry Leiba: they are being used usefully [08:41:43] Xiaodong@CNNIC joins the room [08:41:48] BL: no harm in leaving them as Experimental [08:42:11] John Klensin: RFC 2026 defines Historic as "things we're not going to standardize" [08:42:34] L41587M21F40C20 leaves the room [08:42:37] L41587M21F40C20 joins the room [08:42:58] Ted Hardie at the mic [08:43:15] JK: we could use an applicability statement as individual draft [08:43:20] SNOWBALL602C9192 joins the room [08:43:22] TH: move to Informational? [08:44:24] TH: trying to move this Historic will lead to a rat hole of unusual size [08:45:00] Lisa: no consensus to move this to Historic [08:45:10] DC: agree with move to Informational [08:45:50] richard.barnes leaves the room [08:46:37] Joe Hildebrand leaves the room: Replaced by new connection. [08:46:37] Joe Hildebrand joins the room [08:47:01] moving right along... [08:47:22] Abuse Report Format - Murray Kucherawy [08:47:28] (ARF) [08:47:32] richard.barnes joins the room [08:47:43] DSN-like format for abuse reports [08:47:55] draft-shafranovich-feedback-report [08:48:38] work originated in MAAWG (operational group) [08:48:46] L41587M21F40C20 leaves the room [08:49:03] format has undergone considerable evolution [08:49:12] now relatively stable [08:49:18] used at various ISPs [08:49:20] ARF ARF [08:49:25] Woof! Woof! [08:49:31] implemented in open-source code [08:49:34] e.g., mail filters [08:49:47] Joe Hildebrand leaves the room: Replaced by new connection. [08:49:49] Joe Hildebrand joins the room [08:50:17] community interest in sponsored implementations (open source tools) [08:50:29] purpose of an ARF WG would be to refine the spec [08:50:34] make necessary changes [08:50:47] try to ensure backward compatibility [08:51:01] consider extensions [08:51:29] is INCH appropriate? [08:51:42] integrate ARF into DKIM? [08:51:52] L41587M21F40C20 joins the room [08:52:38] wolfgang.beck01 leaves the room [08:52:45] Lisa: IESG selected Barry Leiba as a liaison to MAAWG [08:52:56] Dave Crocker at the mic [08:53:15] DC: well-established, rock solid [08:53:26] DC: not a lot of pressure to make changes [08:53:44] Roland Hedberg joins the room [08:54:05] DC: do we really need a WG given that there is no strong need to do technical work [08:57:14] Aaron Falk at the mic [08:57:47] spencerdawkins joins the room [08:58:05] AF: is there a disconnect here to do something email-centric vs. something more broad? [08:58:11] Roland Hedberg leaves the room [08:58:16] sorry - what topic are we on? [08:58:27] still talking about ARF [08:58:31] thanks! [08:58:32] Abuse Report Format [08:58:54] L41587M21F40C20 leaves the room [09:00:41] BL on Message Recall [09:01:51] tlyu leaves the room [09:01:53] often desirable to recall a message that you sent [09:02:00] lots of complications [09:02:10] SNOWBALL602C9192 leaves the room [09:02:29] - authorization across domains, what if recipient has seen it, multiple recipients, etc. [09:02:34] - are there attacks here? [09:03:30] possible approach: each message has a unique ID or secret code [09:03:51] include hash in message [09:03:58] recall request includes original code [09:04:13] two-stage hold/recall? [09:05:27] The lying attack can be solved by just hashing twice in the first place. send out hash(hash(secret)) to start. start the process by sending out hash(secret), and finish by sending out secret. [09:05:38] I wonder if anyone currently sends "blahblah wishes to recall message foo" just to get people to read the original message :-) [09:06:29] MORG WG discussion [09:07:16] Tony Hansen at the mic [09:07:32] doable as extension to message tracking protocol? [09:07:47] yao joins the room [09:07:58] Chris Newman at the mic [09:08:11] yao leaves the room [09:08:22] CN: the Internet protocol suite is weak here, this is a hole we need to fill [09:08:54] Randal Gellens at the mic [09:09:09] Alissa Cooper leaves the room [09:10:57] Andrew Yourtchenko at the mic [09:10:59] coopdanger joins the room [09:12:42] Dave Crocker at the mic [09:13:11] dmeyer leaves the room: Replaced by new connection [09:13:24] DC: Barry's motivation for doing this with SMTP = control plane instead of user messaging plane [09:13:51] BL on DRM Proxy Architecture [09:13:57] spencerdawkins leaves the room [09:13:59] spencerdawkins joins the room [09:14:16] Digital Rights Management exists, service providers need to deal with it in a flexible way [09:14:40] parties: content provider, service provider, service consumer, content consumer [09:15:20] service consumer is something like a licensing server, things on the home lan, TiVo / DVR, iTunes, etc. [09:15:43] content consumer: user-controlled device [09:16:26] architecture might involve a DRM proxy between requester and DRM server [09:17:24] questions... [09:17:32] is it interesting to standardize this? [09:17:42] better than having non-standard stuff? [09:17:46] future BoF? [09:18:21] no strong interest expressed, it seems [09:18:35] Alexey: new AppArea website at http://www.apps.ietf.org/ [09:18:51] any volunteers to help review / edit content? [09:19:47] (please contact Alexey and Lisa) [09:19:53] coopdanger leaves the room [09:20:18] Alexey: would it be helpful to have WG status reports in the AppArea meeting? [09:21:20] intDance leaves the room [09:21:34] Barry Leiba: having people report only because they need to report is not all that helpful [09:21:44] Dave Crocker: perhaps report at least once a year [09:21:55] I'm not sure it's useful or helpful [09:22:05] to force people to report on their progress [09:22:11] in general or here [09:22:11] wolfgang.beck01 joins the room [09:22:32] Barry: perhaps require reports from WGs that have met? [09:22:50] Randall Gellens leaves the room [09:22:51] Aaron Falk at the mic [09:23:01] (I think, the room is dark and it's hard to see :) [09:23:56] Alexey: do we want an AppsArea WG? [09:24:04] Barry: this might take some load off the ADs [09:25:51] richard.barnes leaves the room [09:26:14] Joe Hildebrand: might be helpful to report on blocking issues [09:26:27] cyrus leaves the room [09:26:39] Atarashi Yoshifumi leaves the room [09:26:44] spencerdawkins leaves the room [09:26:46] Lisa Dusseault leaves the room [09:26:47] meeting adjourned [09:26:56] shirleyhm leaves the room [09:27:01] Joe Hildebrand leaves the room: Disconnected. [09:27:03] ylafon leaves the room [09:27:08] eburger leaves the room [09:27:41] bhoeneis leaves the room [09:27:57] jlcjohn leaves the room [09:29:37] ray leaves the room: Replaced by new connection [09:30:06] stpeter leaves the room [09:30:21] sm leaves the room [09:31:33] wolfgang.beck01 leaves the room [09:31:36] rababy leaves the room: Computer went to sleep [09:32:05] danwing leaves the room [09:33:07] resnick leaves the room [09:33:21] tonyhansen leaves the room [09:34:43] Xiaodong@CNNIC leaves the room [09:41:23] James Galvin leaves the room [09:43:10] sal leaves the room [09:58:07] Glenn Parsons leaves the room [10:22:12] spencerdawkins joins the room [10:26:50] spencerdawkins leaves the room [10:51:56] intDance joins the room [10:52:06] intDance leaves the room [10:54:14] Roland Hedberg joins the room [10:55:46] Roland Hedberg leaves the room [10:57:50] richard.barnes joins the room [10:59:02] Joe Hildebrand joins the room [11:02:43] Lisa Dusseault joins the room [11:38:12] Lisa Dusseault leaves the room [12:10:09] Joe Hildebrand leaves the room: Disconnected. [12:20:40] Joe Hildebrand joins the room [12:33:36] Joe Hildebrand leaves the room: Disconnected. [12:46:28] Glenn Parsons joins the room [13:03:59] richard.barnes leaves the room [13:35:23] Joe Hildebrand joins the room [13:35:33] Joe Hildebrand leaves the room [13:41:25] richard.barnes joins the room [13:42:26] richard.barnes leaves the room [13:42:55] richard.barnes joins the room [13:43:01] richard.barnes leaves the room [13:46:50] Glenn Parsons leaves the room