[00:04:15] Barry Leiba joins the room [00:05:11] sftcd joins the room [00:05:15] bortzmeyer joins the room [00:05:18] =JeffH joins the room [00:05:36] haibin joins the room [00:05:38] enrico joins the room [00:05:56] hi [00:06:19] rich going through the agenda [00:06:21] HI Haibin. Audio is working well, looks like everything is working great for remote participants... [00:06:36] roni_even joins the room [00:07:03] For those who uploaded slides more than say 10-15 minutes ago, note that they uploaded 2 or 3 more talks a few minutes ago [00:07:37] SoftgearKo joins the room [00:08:20] there will be an additional strawman proposal presentation if time permits [00:08:40] Remote participants: if you'd like me to channel a comment to the microphone, please prefix your message with "MIC: ". Thanks. [00:09:54] lampson0505 joins the room [00:09:56] FSP-AKW308A joins the room [00:10:52] first talk -- Richard Alimi [00:11:00] "DECADE Problem Statement" [00:11:13] slide 2 -- P2P Content Distribution Paradigm [00:12:15] slide 3 -- Contributes Significant Traffic [00:12:30] slide 4 -- Stress on Infrastructure [00:12:48] slide 5 -- In-Network Storage [00:13:17] slide 6 -- Problem 1: Weaknesses of Existing P2P Caches [00:14:09] slide 7 -- Problem 2: Weak/No Integration with Applications [00:15:29] slide 8 -- DECADE Overview [00:15:51] Glenn Parsons joins the room [00:16:23] slide 9 -- Example Operation [00:17:37] slide 10 -- Use Case 1: P2P Users Sharing Content [00:17:37] momose joins the room [00:18:24] slide 11 -- Use Case 2: Content Publishers Sharing Content [00:19:11] One of the co-chairs was just speaking [00:19:15] slide 12 -- Working Group Goal [00:19:22] Slide 13 -- Key Benefits [00:20:27] Slide 14 -- Scope [00:20:40] shamus joins the room [00:21:13] Slide 15 -- Comments and questions? [00:21:33] Ed Juskevicius [00:22:46] Leif Johansson [00:23:01] Leif: how is this different from ALTO [00:24:00] co-chair [00:24:32] momose leaves the room [00:24:32] FSP-AKW308A leaves the room [00:24:32] Glenn Parsons leaves the room [00:24:32] shamus leaves the room [00:24:32] roni_even leaves the room [00:24:32] haibin leaves the room [00:24:32] sftcd leaves the room [00:25:56] mic: Unless I am missing something, one comment/clarification that I think is not quite obvious here is slide 14 states this is for P2P, but technically speaking, given an open API, what you are really creating here really is a form of cloud storage (minus location, as stated) P2P seems the most likely customer, but there may be others (I could write an offsite backup app for my computer or even a dropbox like app, both using the decade storage I "lease" from my ISP, for example). (note that I think that is a nice side benefit, not a bad thing, BTW) [00:26:40] Dy Kim: [00:27:16] david: i'm in the queue [00:27:18] Dy: Are you going to have mechanism for controlling data ownership [00:27:36] barry: thx [00:28:07] Speaker: Yes, there will be an authorization mechanism [00:28:13] Stuart Cheshire [00:28:33] Stuart: What will be the incentives to make people use this? [00:29:12] Stuart: What will prevent folks from viewing it as free storage? [00:29:35] barry: you can drop my question, Stuart is going down this line... [00:29:47] Speaker: Provider could put limits on storage offered to a user [00:29:54] (admiditily with a negative spin, where I see it as positive ;) ) [00:30:03] Enrico Marocco [00:30:18] David: dropped. Thanks. [00:31:30] Enrico: clarified difference between ALTO and DECADE [00:31:34] Omer Luzzatti [00:31:42] enrico leaves the room [00:33:29] next presentation: "DECADE Requirements" -- same speaker [00:33:42] This is Richard Alimi speaking [00:33:52] slide 2 -- Overview [00:34:05] slide 3 -- Data Access [00:34:32] slide 4 -- Resource Control [00:34:33] enrico joins the room [00:35:27] slide 5 -- Authorization [00:35:51] slide 6 -- Data Management [00:36:13] slide 7 -- Error Conditions [00:36:20] sorry went back to slide 6 [00:36:30] now slide 7 [00:36:51] mic: One issue that I think also needs to be discussed is if this group will actually make a protocol that handles the actual transfer (fragmentation, error checking, etc.) or we "wrap" some existing IETF protocol that actually does the transfer of data (we have done that work before, of course) [00:37:08] (sorry Barry, may be making you stand up a bit today! ;) ) [00:37:29] slide 8 -- Comments and questions? [00:37:36] Barry Leiba [00:37:51] What about functions like list, move, etc. [00:38:54] Barry: what about atomic move of data from one cache to another? [00:39:20] Ed Juskevicius [00:39:44] Ed: for this to be most useful, this should be near to me as a client. Is there anything in the protocol that indicates how close the storage is to me? [00:40:25] Speaker: If one storage location available, then discovery; if multiple, then ALTO would be better [00:40:30] Stuart Cheshire [00:40:40] FSP-AKW308A joins the room [00:40:42] Stuart: Don't these services already exist? [00:40:56] Lisa (?): [00:41:10] Lisa: would like to see a gap analysis [00:41:21] inami.masaaki joins the room [00:42:19] myjoni joins the room [00:42:38] Lisa: There's a very close match with existing system (Webxx?). [00:42:55] Joe Hildebrand joins the room [00:43:05] Lisa: What's stopping client applications from using this? [00:43:18] This is Lisa Dusseault [00:43:48] atarashi joins the room [00:43:52] shamus joins the room [00:44:16] David Bryan: [00:44:23] (Brian?) [00:44:56] atarashi leaves the room [00:45:13] (general comment and agreement to use existing transport protocols) [00:45:22] roni_even joins the room [00:45:26] next talk : DECADE Survey -- Ning Zong [00:45:47] slide 2 -- Survey Overview [00:46:23] slide 3 -- In-network Storage System Components [00:46:54] haibin joins the room [00:47:21] slide 4 -- Transparent P2P Cache [00:48:39] Enrico Marocco: [00:48:49] Joe Hildebrand leaves the room: Disconnected. [00:48:51] clarifying that this is what we should NOT do, right [00:49:24] slide 5 -- Web Cache [00:50:54] slide 6 -- CDNs [00:51:19] slide 7 - NFS [00:51:30] slide 8 -- Comments and questions? [00:51:39] HUI joins the room [00:52:01] other slides have info on other systems [00:52:11] co-chair speaking [00:52:29] So the working group would do a survey [00:53:37] lampson0505 leaves the room [00:53:39] Rich Alimi speaking: [00:54:29] resuming use of Chair slides -- slide 7 [00:54:55] Rich Woundy: [00:55:06] "The Elephant in the Room -- Illegal Content" [00:55:55] sftcd joins the room [00:56:19] Atarashi Yoshifumi joins the room [00:58:15] Barry Leiba: [00:58:35] Rich again: [00:58:40] so since jabber was out for a bit, I wanted to say +1 to wondering whether webdav could do all the presented requirements [00:58:55] presnick joins the room [00:59:07] slide 8 -- more on the elephant [01:03:15] Dave Crocker [01:04:43] momose joins the room [01:05:48] Pete Resnick [01:06:08] DCN-0D4B06D5DF0 joins the room [01:07:19] Leif Johansson [01:07:24] Rich Alimi [01:08:17] Rich Woundy [01:10:47] Leif Johansson [01:12:03] mic: If no one else says it, there is an advantage to the implementors in that content will be available more widely (potentially when peers are offline) and from better connections. There are incentives for implentors. [01:12:22] to sftcd: it seems that webdav can be a good starting point, but not fully match the requirements. For example, I checked RFC 4331 (LIsa is a co-author), it specifies only available bytes in the storage. For DECADE, an important resources/quota control may be proportional bandwidth sharing, for example. [01:12:52] In addition to pplive, also thunder and flashget, major p2p systems in china [01:13:20] @FSP: be good to see that analysis in the survey I guess [01:13:25] ahout half in the room have read the charter -- there are perhaps 50-70 people in the room [01:14:34] Glenn Parsons joins the room [01:15:10] http://trac.tools.ietf.org/bof/trac/wiki/decade [01:16:37] Lisa Dusseault [01:20:02] Richard Alimi [01:22:07] inami.masaaki leaves the room [01:25:20] next talk -- strawman proposal 1 (Richard Alimi) [01:26:31] with cooperation with PPLive and China Telecom [01:27:32] system architecture [01:28:34] Lars Eggert : looks like from the Network File System working group [01:28:49] slide 5 -- System Architecture [01:29:24] slide 6 -- Storage Server Resource Model [01:31:35] slide 7 -- Access Protocol [01:35:35] slide 11 -- Access Protocol Data Interface (cont'd) [01:35:44] FSP-AKW308A leaves the room [01:36:19] haibin leaves the room [01:37:44] FSP-AKW308A joins the room [01:39:39] next talk: Network of Information [01:39:51] This isn't Borje -- but I didn't catch the name [01:40:20] slide 3 -- Problems resulting from a host-centric view [01:40:33] Martin [01:40:39] DCN-0D4B06D5DF0 leaves the room [01:40:50] slide 4 -- NetInf Scenarios [01:41:16] slide 5 -- API for accessing any type of object, regardless of location [01:41:50] sal joins the room [01:42:12] slide 6 -- Organize Information -- IO, DO and BO definitions [01:42:43] FSP-AKW308A leaves the room [01:43:14] slide 7 -- NetInf Naming [01:43:35] Martin Johnsson [01:44:09] slide 8 -- NetInf Architecture Overview [01:45:46] Lisa Dusseault [01:48:47] The last speaker was Craig White [01:49:05] (by speaker, I meant commenter at the mic) [01:49:16] Lars Eggert [01:50:20] Leif Johansson [01:50:28] Joe Hildebrand joins the room [01:51:07] myjoni leaves the room: Replaced by new connection [01:51:09] myjoni joins the room [01:51:14] Rich Alimi [01:51:26] Lars Eggert [01:52:22] Masato Yamanishi [01:53:39] That was Haibin Song [01:53:48] This is Richard Woundy [01:54:36] FSP-AKW308A joins the room [01:55:34] Lisa Dusseault [01:57:20] Alexey Melnikov [01:57:44] Chair slides -- slide 11 [01:58:28] lampson0505 joins the room [01:58:31] Question appears to be to cover multiple topics. So hum was mixed [01:59:09] 1. Is the problem statement clear? [01:59:19] not unanimous but 70:30 that it's clear [01:59:49] 2. Is the problem useful to solve [02:00:08] hum was 80:20 [02:00:41] 3. Is the problem well-scoped? [02:01:59] myjoni leaves the room: Replaced by new connection [02:02:02] myjoni joins the room [02:02:06] can't see on the audio stream. [02:02:14] thx [02:02:34] The Working Group (WG) will design and specify a standard protocol for P2P applications to utilize in-network storage. [02:03:02] The protocol will include operations for a P2P applications system to (1) store and retrieve data [02:03:23] (2) specify access control on who can store into and retrieve from its allocated storage, and [02:03:32] (3) indicate resource sharing policies. [02:03:55] is that tim speaking? [02:04:03] yes [02:04:03] Yup [02:04:06] This WG will focus mainly on the protocol between the in-network storage and P2P applications. [02:04:25] The protocol will be focussed on P2P applications, though it may also be used in other types of applications.. [02:04:39] That's it for the problem statement [02:04:44] Lars Eggert [02:05:01] rsync, etc. [02:05:13] (sorry if that was brought up earlier) [02:05:37] Alexey [02:05:45] Atarashi Yoshifumi leaves the room [02:06:29] Was that Stuart Cheshire? [02:06:37] Sorry, that was me. [02:06:44] I froggot to identify. [02:06:56] But Stuart is now heading to the mic. [02:07:03] so I'm now confused about whether we're humming for Lisa's (no requirements thing) or the text on screen [02:07:29] This is Stuart Cheshire [02:08:17] In theory we're about to home on whether the problems statement is well-scoped. [02:08:22] Lisa Dusseault [02:09:02] MIC: Amazon DOES do that today, and it's cheap. [02:09:16] (no, i'm not in the room this time :) ) [02:09:28] @karen: there's an I-D called problem statement and there's the text on screen being called that just to add to the fun [02:09:46] thanks [02:10:15] So re: which are folks supposed to be humming? [02:10:50] the mic's over the room from me [02:10:55] Lars Eggert [02:11:06] np. it's not relevant enough to interrupt. [02:11:17] can't hear lisa! [02:11:39] Lisa just said that there are too many different protocols [02:12:44] but if you use Amazon S3, you've got exactly what we're talking about, in a protocol-inspecific manner. [02:12:45] One of the claims for DECADE is that service providers would have to only implement one. [02:12:49] Barry: [02:13:31] so let's add *another* protocol? [02:13:47] Or pick one that's close, and extend it. [02:14:01] Craig White speaking. [02:14:04] Craig White [02:14:39] zhouzp joins the room [02:15:33] I didn't realize, but S3 already has bittorrent interfaces. [02:15:35] done. [02:16:20] Rich Woundy: [02:17:05] HUI leaves the room [02:17:19] Lucy Yong [02:17:19] Lucy Yong [02:17:23] Joe H: Well, if we allow that as a reason not to build something, the IETF is picking winners. "Bit Torrent and Amazon do it, customers can just use them, so there is no need for a standard" Bit Torent isn't the only P2P system, and amazon isn't the only person wanting to provide storage (or prvoide storage because they want to reduce usage in an ISPs case...) [02:17:35] zhouzp leaves the room: Logged out [02:18:32] Alexey [02:18:37] David: solid point [02:18:56] Lisa: [02:19:39] i'm in the camp that says that if people want to work on something with a good amount of energy, the IETF should let them. [02:19:47] Lars: [02:20:55] Alexey [02:20:56] roni_even leaves the room [02:21:20] isn't that better as hands? [02:21:21] Hum: re: working group for gap analysis [02:21:47] unanimous hum [02:21:50] Barry: [02:21:57] thanks, barry. [02:22:16] [02:22:27] [02:22:27] zhouzp joins the room [02:22:47] quantify decent, please? [02:22:53] 0-ish [02:22:55] About 20 raised hands [02:22:56] haibin joins the room [02:22:57] 20-ish [02:23:09] 0-ish wouldn't be enough... :) [02:23:45] [02:23:59] oh, wait, now he is talking about specific drafts...darn lag [02:24:33] sftcd leaves the room [02:24:41] momose leaves the room [02:24:43] zhouzp leaves the room [02:24:45] shamus leaves the room [02:24:48] karen.s.seo leaves the room [02:24:50] thanks barry and karen! [02:24:54] presnick leaves the room [02:24:56] enrico leaves the room [02:25:04] Thanks everyone who relayed for me [02:25:09] FSP-AKW308A leaves the room [02:25:15] bortzmeyer leaves the room [02:25:18] Barry Leiba leaves the room [02:25:24] David Bryan leaves the room [02:25:59] SoftgearKo leaves the room [02:27:08] lampson0505 leaves the room [02:27:22] Joe Hildebrand leaves the room [02:29:10] =JeffH leaves the room [02:38:49] haibin leaves the room: Computer went to sleep [02:41:48] sal leaves the room [02:57:12] Glenn Parsons leaves the room [03:06:46] Atarashi Yoshifumi joins the room [03:10:12] Atarashi Yoshifumi leaves the room [03:13:58] Ruri Hiromi joins the room [03:14:29] Ruri Hiromi leaves the room [03:42:33] bortzmeyer joins the room [03:43:57] bortzmeyer leaves the room [04:09:22] myjoni leaves the room