[10:47:36] jlcjohn joins the room [10:48:34] softgear joins the room [10:52:15] enrico joins the room [10:53:58] Is there any material ? [10:54:07] skame joins the room [10:55:34] sk joins the room [10:55:52] sk leaves the room [10:56:06] sebastian.kiesel joins the room [10:57:27] skame leaves the room [10:57:45] MartinStiemerling joins the room [10:58:09] NYX joins the room [10:58:34] mlepinski joins the room [11:01:54] Alissa Cooper joins the room [11:02:01] Wolfgang Beck joins the room [11:02:27] HannesTschofenig joins the room [11:03:00] dmeyer joins the room [11:03:44] Jan.Seedorf joins the room [11:03:46] momose joins the room [11:04:02] jon-ietf joins the room [11:04:24] ietfdbh joins the room [11:05:01] meeting material are available on the proceeding page [11:05:07] https://datatracker.ietf.org/meeting/75/materials.html [11:07:07] shamus joins the room [11:07:23] enrico leaves the room [11:07:32] eburger joins the room [11:07:43] Eric, your humble Jabber Scribe has arrived [11:07:47] (and modest, too) [11:08:01] Just finished the introductions, not moving to alto-reqs discusssion [11:08:12] (Sebastian Kiesel) [11:08:21] enrico.marocco joins the room [11:08:49] slide 2 [11:09:13] thanks Eric for being jabber scribe [11:09:13] >>> Anyone participating remotely? [11:09:28] yes! [11:09:39] I'll channel for you, if you want to ask anything. [11:09:41] Slide 3 [11:09:46] thx [11:10:10] I'm also remote [11:10:26] NP Martin. I suppose Stockholm is to close to your home ;-) [11:10:51] yeah, it's only approx. 1 hour flight [11:10:53] we're saving the travel budget for Hiroshima.. [11:10:59] :-) [11:11:22] hyperfeng87745695 joins the room [11:11:31] ..which would be hard to attend remotely, because of the time difference. [11:11:31] Slide 4 :So why all this fuss... [11:12:57] What slide are we looking at? [11:12:58] Slide 5: Host Group Descriptor requirements [11:13:14] (just switched to it) BTW, hope you are feeling better [11:13:31] it was just added to the meeting materials page [11:15:35] Slide 6: Rating Criteria requirements [11:16:31] Slide 7: ALTO Server Discovery requirements [11:16:33] Lars joins the room [11:19:30] Should we document preferences for one of the options? [11:19:32] Ayone? [11:19:36] Leave as open issue? [11:19:45] Anyone awake? [11:19:46] doc in this draft [11:19:50] Has someone asked operators whether they would want to provide that information to end hosts? [11:19:52] rababy joins the room [11:20:07] Rich Alimi at mic [11:20:08] should be in the resource consumer [11:20:09] Simon Perreault joins the room [11:20:33] If the end host learns it right away from the access provider then the solution is more stable. [11:20:41] discovery of ALTO servers in other networks sound like a difficult problem [11:21:26] Finding the access provider based on the IP address of the end host is complicated and error prone (as we know from the VoIP emergency services work). [11:22:02] Hannes: right, but still this seems to be a valid question, as the tracker is usually not in the access net or even close [11:22:40] @Hannes: we (DTAG) are willing to tell peers as much as we are willing to tell trackers [11:22:54] Glenn Parsons joins the room [11:22:58] Roland Hedberg joins the room [11:23:19] alexmcmahon joins the room [11:23:28] ..that is, we don't want to reveal too much about our topology, but are willing to sort candidates [11:24:06] Simon Perreault leaves the room [11:24:16] @Wolfgang: For the discovery you don't have to real a lot about topology. [11:25:31] Rich Woundy at mic [11:25:54] @Wolfgang: For the discovery you don't have to reveal a lot about topology. (fixed a bug) [11:26:38] Roni Evan at mic [11:26:52] @Hannes: yes, discovery is not that critical. But still it requires some effort to find the relevant server, if you have more than one. [11:26:54] Roni: why do we need to document at all? Both work and are fully compatible [11:27:25] Who @ mic? [11:27:43] Albert Chang @ mic [11:27:52] Ren(?) fr. Junpier? at mic [11:27:56] reinaldo [11:28:02] Thanks Lars! [11:29:00] I don't care about documenting it in a requirements document. Nobody is going to read that document anyway after it is finished. Discovery is important if this should have anything todo with talking to the ALTO server close to end host (in the access network). [11:29:01] hyperfeng87745695 leaves the room [11:29:05] Simon Perreault joins the room [11:29:07] David Geison (sp?) [VZ] at mic [11:29:17] hyperfeng87745695 joins the room [11:31:26] make sense hannes? [11:31:32] Rich [11:31:42] Making the decisions is fine but arguing where to capture it isn't [11:31:47] agreed [11:31:54] Lars @ mic [11:32:02] i just think some things get surfaced from this discussion that we might not see otherwise [11:32:03] At the end, I hope that someone will write a solution document (like a new DHCP option) [11:32:05] so it's useful [11:32:27] Lars: back to requirements; this looks a lot like a solution [11:33:00] Will take question to list [11:33:17] Slide 8: ALTO Server Discover requiremetns (39 ^ 40) [11:33:34] This is a process requirement. Is it useful to have in the document? [11:34:19] Slide 9: Security requirements: authentication [11:34:52] Lars @ mic [11:36:15] Ruediger Volk: MUST vs SHOULD [11:36:15] if skype works behind it, then alto must too [11:36:20] heh [11:36:21] Slide 10: Next steps [11:36:48] David Geisen(?) [11:37:02] Dave McDyson [11:37:25] @Lars: you're closer & can read badges; can you do "Who is at the Mic now?" for me? Thx [11:37:30] can do [11:37:44] and otherwise i'll just make up names :-) [11:37:51] [like this one?] [11:37:55] jakub helber he's not wearing a name tag [11:39:31] NEXT UP: ALTO Protocol [11:39:38] enrico.marocco joins the room [11:39:47] @Eric: Thanks for being the jabber scribe :) [11:39:56] Reinaldo Penno speaking [11:40:01] @Martin: NP [11:40:24] do we have a design team in the WG? [11:40:30] I guess so... [11:40:38] I was oblivious, but I'm always oblivious [11:40:49] hmm,wasn't made public, or I have missed this.. [11:41:23] this is not an "official" design team subgroup [11:41:24] we have at least two groups: alto client protocol and alto discovery [11:41:43] Slide 3: Design history [11:41:44] thanks Jon for the clarifciation [11:41:48] it's a group of participants who are co-authoring a doc [11:41:58] if we were going to form an official design team it would be on the list, never fear ;) [11:42:16] yes, I do trust you guys, was just confused by the speaker... [11:42:29] some people are not fully in IETF terminology [11:42:39] Slide 4: Contributing Proposal: P4P [11:43:00] softgear leaves the room [11:43:19] a question about slide 3: can the authors outline (mailling list) what and how they did merge?? [11:43:32] isn't clear to me at all, including possible side effects [11:43:41] give it a minute or two here [11:43:43] Slide 5: Contributing Proposals: InfoExport [11:44:10] ack [11:44:39] Gustavo joins the room [11:44:52] enrico.marocco leaves the room [11:45:15] Slide 6: Contributing Proposals: Proxidor [11:45:59] jon-ietf leaves the room: Replaced by new connection [11:46:06] jon-ietf joins the room [11:46:12] eburger leaves the room: Replaced by new connection [11:46:16] eburger joins the room [11:46:28] enrico.marocco joins the room [11:46:36] Slide 7 (sorry for the disconnect - -a.1x opportunity): Query/Response [11:47:42] what happened to the proxidor approach of sorting ip addresses in the server, abandoned? [11:47:52] many contributing proposals.... [11:47:54] (sorry for asking so many times...) [11:48:57] Slide 8: ATTP [11:49:15] @Martin: do you feel there was an answer w.r.t. Proxidor? It may be coming up later in the discussion... [11:49:27] ... i think yeah, it's coming [11:49:28] nope , no answer at all. [11:49:29] if not i'll ask it [11:49:32] thx [11:49:33] thanks! [11:49:40] Slide 9: DNS based NetLocation Service [11:50:07] Back to Slide 8 [11:50:36] if all that remains of the "sorting oracle" approach is some additonal metrics, it is a strange claim that the idea has been merged with P4P [11:50:41] Slide 10: Outline [11:50:49] hm, what about the semantics? ( [11:50:56] don't care about syntax... [11:50:59] ) [11:51:00] hyperfeng87745695 leaves the room [11:51:35] Slide 11: my-Internet View [11:51:54] Slide 12: Hosting ALTO Server [11:52:45] Slide 13: ALTO Cost Type [11:53:00] hyperfeng87745695 joins the room [11:53:20] Slide 14: Location Grouping [11:53:24] Lars @ mic [11:55:17] (Eric sleeping at the Jabber screen) @ mic [11:55:30] Slide 15: Outline [11:55:37] Ah - Richard at mic, and now giving presentation [11:55:53] does A really have to know about the cost between B and C? [11:55:57] This My Internet View is confusing for me [11:56:13] @Hannes: The basic idea is nice and good [11:56:22] however, the current spec needs some more work and definiton [11:56:22] Dave @ mic [11:58:46] Cost = arbitrary cost, not $$$ cost [11:58:48] costs is just the wrong term [11:58:57] take the term form the requirements [11:59:03] rating critertion [11:59:11] Slide 16: ALTO Protocol Scope [11:59:38] Slide 17: ALTO Service Model [12:00:35] Slide 18: ALTO Query Types [12:01:19] martin joins the room [12:01:39] martin leaves the room [12:02:06] Lars @ mic [12:02:26] who defines properties? do you really want to tell the world about your individual usage caps? [12:02:38] +1 to Lars' comment [12:02:51] @wolfgang: need me to channel? [12:02:56] Or, take to list [12:03:07] Slide 19: ALTO Protocol Design [12:03:12] @eric: no, I think Lars covered it. [12:03:14] [Or, here comes an answer] [12:03:28] Slide 20: Network Map [12:03:43] (we need a second projector displaying the jabber room with all the questions ;) [12:04:04] That would be so Web 2.0 :-> [12:04:18] yeah [12:04:39] Simon Perreault leaves the room [12:04:56] Dave @ mic [12:05:07] on privacy, ALSO asking about usage cap [12:06:26] Slide 21: Network Map Queries [12:06:45] Slide 22: Path Rating [12:07:31] Slide 23: Path Rating Query [12:07:41] any reason why there are numerical and ranking? [12:07:51] isn't numerical enough to rank? [12:08:02] Slide 24: Protocol Message Encoding [12:08:03] KISS should also apply here ( [12:08:09] @Martin, I would think so [12:08:16] (Numerical is enough) [12:08:34] dimitri papamitriou at mike [12:09:02] i don't think telcos will give you that level of detail. Or else we would let you access our internal routing protocols.. [12:09:11] volker hiltl at mike [12:10:06] I hear this very common that we use XML & HTTP and hence we are flexible. [12:10:16] +1 from south of Stockholm to what Volker said [12:10:55] hannes: i on the other hand is happy that we're not seeing another "need to build from scratch using udp" effort [12:11:25] but still this whole is rather unstable with too many options, yet http+xml might be good [12:11:52] http over TCP or over SCTP? [12:11:59] over udp [12:12:02] Chris Griffiths joins the room [12:12:16] How about XML payload in INFO over SCTP over TLS over UDP? [12:12:19] jk [12:12:27] no he is for serious [12:12:29] autsch [12:12:30] Slide 25: [12:12:37] Slide 26: Use Case 1 [12:12:42] Slide 27 - pretty pictures [12:12:43] Simon Perreault joins the room [12:12:50] in san franciso we were arguing about "MUST use tcp", now XML-over-HTTP-over-TCP is widely agreed? [12:12:58] Emil Ivov joins the room [12:12:59] Slide 28 [12:13:16] Slide 29 [12:13:27] peterson looks like the xml parser in his head is malfunctioning :-) [12:13:32] DCCP tunneled in as SRTP payload in UDP [12:13:36] no no i was going to sneeze [12:13:50] that may however be a symptom of excessive xm [12:13:52] Worse yet, it looks like serious XML abuse if you ask me. PLEASE take this to some XML folks before trying to publish [12:13:59] SLIDE 31 [12:14:06] (Sorry I missed 30) [12:14:09] skipped 32 [12:14:11] Slide 33 [12:14:12] Slide 34 [12:14:20] lisa is looking over these [12:14:37] enrico.marocco leaves the room [12:14:42] Lisa looks lost. Sorry kitten lost job... [12:14:44] Slide 36 [12:15:05] slide 38 [12:15:17] Slide 39 [12:15:20] Slide 40 [12:15:52] I don't see this yet ready as WG item [12:16:11] stefano previdi i think [12:16:16] (at the mike) [12:16:21] yup [12:16:23] enrico.marocco joins the room [12:16:23] stefano [12:16:27] needs much more work and some clarifications what happened during the merger with some features [12:16:39] http://www.standardstrack.com/ietf/IMG00068-20090726-1915.jpg [http://www.standardstrack.com/ietf/IMG00068-20090726-1915.jpg] [12:16:57] well, martin, i hear what you're saying [12:17:02] thanks [12:17:06] wtf? [12:17:17] the picture: after reading the xml or before? [12:17:19] ;) [12:17:25] >:-o [12:17:36] but i'd note that the decision to adopt is about the contents of the doc rather than the process by which the doc was assembled [12:17:38] in other words [12:17:47] Gabriel Hege joins the room [12:17:49] i'd ask that we stay focused on what we want to be in the doc rather than how it got there [12:17:51] make sense? [12:17:55] are there any other serious options at this point? [12:18:02] Lisa @ mic [12:18:07] make sense [12:18:17] Lisa: "Best RESTful design in the IETF" [12:18:28] Lisa: but... Fixed URL's kind of s*ck [12:18:35] (which may not be saying much...) [12:18:38] enrico.marocco leaves the room [12:18:46] Lisa: but why do they s_ck? [12:19:05] we can always throw in some NAPTR magic [12:19:51] for the content of the document: I'm fully convinced if the current approach is workable beyond the P4P trial [12:19:56] enrico.marocco joins the room [12:19:59] should read: I'm not fully [12:20:09] [Orange] @ mic [12:20:35] berthrand [12:20:41] Chris Griffiths leaves the room: Replaced by new connection [12:20:51] berthrand mathieau [12:20:52] +1 [12:22:06] Lisa Dusseault joins the room [12:22:47] stefano at mic [12:24:22] alberttian joins the room [12:24:30] Jan at mic [12:26:19] NEXT UP: ALTO-FCP [12:26:41] yushun wang at mike [12:27:24] Chris Griffiths joins the room [12:28:45] Slide 2 [12:30:30] Slide 3: Protocol [12:32:34] hyperfeng87745695 leaves the room [12:32:45] Slide 4 [12:33:06] hyperfeng87745695 joins the room [12:33:52] isn't this pretty far out of scope? i believe the charter restricts alto currently to optimizing the *initial* peer selection [12:34:05] I would call it "interesting" :-) [12:34:16] Ted Hardie joins the room [12:34:33] NEXT UP: ALTO Service Descovery [12:34:44] Ted Hardie leaves the room [12:34:49] Dr. Tomsu ;-) [12:34:50] demil joins the room [12:35:13] everyone gets to be in an 80s synth band this week [12:35:13] hyperfeng87745695 leaves the room [12:35:21] hyperfeng87745695 joins the room [12:35:55] Slide 2: Status [12:36:36] Slide: Mechanism Options [12:36:57] i vot for sip [12:37:00] vote [12:37:03] i vote for dchp [12:37:22] I vote for an XML payload transported by INFO over SIP over SCTP [12:37:32] Slide: Feedback Addressed [12:38:24] Slide: Next Steps [12:38:31] text-to-speec it and send it over rtp [12:38:32] ..and it must be TISPAN SIP to keep the fixed relationship between the access line and the SIP authentication :-/ [12:38:44] Lisa Dusseault leaves the room [12:38:46] how about 3GPP? [12:38:49] Oooh - I like that >:-) [12:38:54] Lisa runs to the mic [12:38:56] even more 3GPP2? [12:39:13] @martin: all the same mess [12:39:19] true :) [12:40:37] Lars @ mic (he cannot say he's @ mic :-) ) [12:40:38] eric: this would have the disadvantage of producing a chicken-and-egg problem, if I want to ask ALTO which P-CSCF to use best [12:42:41] yeah, and i can't remember to say my name either - sorry [12:42:45] you can't choose the P-CSCF! That's the whole point of it. To think that you could contact every server in the Internet is frivolous ! [12:44:08] Steven Wright (AT&T) [12:44:10] @ mic [12:44:41] enrico.marocco leaves the room [12:44:54] we have learned how flexible the alto protocol is, so i expect ALTO to tell me when it's time to cancel my v*d*f*n* subscribption and sign up with T-* :-) [12:45:05] enrico.marocco joins the room [12:45:12] NEXT UP: ALTO SERVICE BASED ON BGP ROUTING INFORMATION [12:45:17] OKAY THANKS ERIC [12:45:20] speaker: Zoran Despotovic [12:45:27] Just call me "LOUD HAL" [12:45:36] or "STUCK CAPS LOCK" [12:45:37] Slide 2 [12:45:38] Slide 3 [12:46:59] Slide 4 [12:47:24] hyperfeng87745695 leaves the room [12:47:44] Slide 5 [12:48:02] @sebastian: the ALTO server will be run by your provider.. [12:48:25] Slide 6 - done [12:48:36] David Dyson @ mic [12:49:05] ietfdbh leaves the room [12:49:49] with this stuff, you can build a screw-the-isp p2p client that probes for and uses the most expensive paths. not sure if that is intended... [12:50:06] Lars - excellent point! [12:50:09] wow [12:50:22] Of course, the ISP can just send the "best" [12:50:47] and ISPs never ever lie to you :-) [12:53:04] martin.thomson joins the room [12:53:20] luigi iannonw at mike [12:53:23] iannone [12:53:26] Any other comments? [12:53:29] Lisa Dusseault joins the room [12:53:34] open mic [12:53:37] NYX leaves the room [12:53:38] mlepinski leaves the room [12:53:46] Gabriel Hege leaves the room [12:53:52] dmeyer leaves the room [12:53:59] BOFs notes [12:54:04] BYE, got to go! [12:54:06] Simon Perreault leaves the room [12:54:08] Lars leaves the room [12:54:15] Jan.Seedorf leaves the room [12:54:19] jon-ietf leaves the room [12:54:27] demil leaves the room [12:54:28] Chris Griffiths leaves the room [12:54:33] Roland Hedberg leaves the room [12:54:33] thanks to the scribs! [12:54:36] scribes! [12:54:36] shamus leaves the room [12:54:39] martin.thomson leaves the room [12:54:39] enrico.marocco leaves the room [12:54:40] MartinStiemerling leaves the room [12:55:05] Emil Ivov leaves the room [12:55:25] Lisa Dusseault leaves the room [12:55:33] rababy leaves the room: Computer went to sleep [12:57:09] enrico.marocco joins the room [12:58:19] momose leaves the room [12:59:02] enrico.marocco leaves the room [12:59:07] Alissa Cooper leaves the room [13:01:10] Alissa Cooper joins the room [13:01:16] sebastian.kiesel leaves the room: Replaced by new connection [13:01:16] sebastian.kiesel joins the room [13:01:30] Lisa Dusseault joins the room [13:03:06] clns joins the room [13:07:26] Wolfgang Beck leaves the room [13:12:37] Simon Perreault joins the room [13:12:45] Simon Perreault leaves the room [13:13:46] eburger leaves the room [13:17:35] alexmcmahon leaves the room [13:18:22] Gustavo leaves the room [13:19:14] Alissa Cooper leaves the room [13:19:22] Alissa Cooper joins the room [13:20:01] sebastian.kiesel leaves the room [13:22:59] Glenn Parsons leaves the room [13:26:45] Glenn Parsons joins the room [13:27:41] alberttian leaves the room [13:31:06] Lisa Dusseault leaves the room [13:31:21] Alissa Cooper leaves the room [13:32:33] Lisa Dusseault joins the room [13:32:50] Lisa Dusseault leaves the room [13:51:39] sebastian.kiesel joins the room [13:53:07] sebastian.kiesel leaves the room [13:57:02] pdf [14:37:49] Lisa Dusseault joins the room [14:40:00] clns leaves the room: offline [14:51:36] jlcjohn leaves the room [14:53:23] ietfdbh joins the room [15:15:19] ietfdbh leaves the room [15:20:24] Lisa Dusseault leaves the room [16:45:21] Mark Teodoro joins the room [16:45:51] Mark Teodoro leaves the room [17:07:47] Glenn Parsons leaves the room [17:19:58] HannesTschofenig leaves the room