[15:11:36] --- juampe has joined
[15:51:41] --- iljitsch has joined
[15:52:25] --- nm has joined
[15:54:57] --- iljitsch2 has joined
[15:57:00] <iljitsch2> /topic shim6
[15:57:17] * iljitsch2 doesn't understand jabber that well
[15:57:54] --- iljitsch2 has left
[16:01:08] --- bnsmith has joined
[16:01:24] --- wouter has joined
[16:01:55] --- becarpenter has joined
[16:01:58] <wouter> Hi iljitsch
[16:02:11] <iljitsch> hi
[16:02:12] --- nm has left: Replaced by new connection
[16:02:12] --- nm has joined
[16:02:13] --- nm has left
[16:02:18] --- nm has joined
[16:02:40] --- nm has left
[16:02:55] <iljitsch> can you see my full jabber name?
[16:02:59] <wouter> Scribe questions. I'm jabber scribe.
[16:03:08] <wouter> No, iljistch
[16:03:16] <wouter> Administrativia
[16:03:27] <iljitsch> we start in ~30 mins, right? (stupid time zone system...)
[16:04:12] <wouter> Right now start.
[16:04:21] <wouter> Chair is reading the agenda.
[16:04:27] <iljitsch> ok better turn on the audio then :-)
[16:04:34] <wouter> No additional changes to agenda.
[16:04:35] <wouter> yes.
[16:04:47] <wouter> Who has read tge draft?
[16:05:10] <wouter> Not much response (as far as your jabber scribe can see)
[16:05:16] --- mo7sen has joined
[16:05:29] <wouter> Brief flavor of the draft contents. Presnetation by Geoff
[16:05:40] --- kurtis has joined
[16:05:47] <wouter> Resolving last call issues
[16:06:04] <kurtis> Iljitshch: WE are starting now
[16:06:05] <wouter> Level of feedback has been low for due diligence
[16:06:46] <wouter> Chair asks for feedback.
[16:07:00] --- gr8k@jabber.org has joined
[16:07:37] <wouter> Sorely need feedback.
[16:07:53] <iljitsch> oh man, the audio is just as bad as monday for v6ops
[16:07:54] <wouter> See the slides elsewhere for slide contents.
[16:08:18] --- peetu has joined
[16:08:28] --- ks has joined
[16:08:32] <kurtis> Iljitsch do you want us to try and change microphone?
[16:08:48] <wouter> Geoff explains multihoming problem, routing scaling.
[16:08:57] <juampe> it's a problem of the audio capture system
[16:09:04] <kurtis> ok
[16:09:05] <iljitsch> if you think it will help... I don't think it's the actual microphone.
[16:09:08] <becarpenter> iljitsch, which channel?
[16:09:13] <iljitsch> 4
[16:09:55] <wouter> Microphone audio stuff.
[16:09:58] <kurtis> Geoff changed micophone
[16:09:59] <iljitsch> ah great this helps
[16:10:03] --- fparent@jabber.org has joined
[16:10:13] <wouter> Can you multihome without stressing the routing system.
[16:10:13] <becarpenter> Ah, that is supposed to ok, channels1 and 7 have problems
[16:10:14] <iljitsch> when they closed the ticket from monday it said: "audio channel should be a bit better... it's never going to be perfect on 1 and 7 sorry."
[16:10:51] <wouter> Shim6 approach.
[16:10:54] <wouter> * Host based
[16:11:08] <wouter> * network layer, per host pair
[16:11:25] <wouter> * discoverable negotiated
[16:11:28] <wouter> * no new identifiers
[16:11:48] --- nm has joined
[16:12:06] <wouter> This is not SCTP , not HIP
[16:12:21] <wouter> You connect normally, then later shim6 happens
[16:13:13] <wouter> Pretty picture from shim6.org
[16:14:20] <wouter> The shim changes upper level identifiers into routing system locator addresses (and back again)
[16:14:35] <wouter> Slide Initial Contact
[16:15:05] --- gr8k@jabber.org has left
[16:15:27] <wouter> Slide Shim Activated
[16:16:21] <wouter> Slide Locator Failure
[16:16:41] <wouter> Olaf volunteers to scribe
[16:17:08] <iljitsch> I have a question for later on. It's pretty long so I'll type it in now. Since the interest in shim6 is somewhat lackluster (see lack of reviews), there is no PI in IPv6, and the IAB workshop results are very likely going to result in something that has an id->loc mapping service, maybe it's a good idea to not go to last call for the shim6 protocols, but rather wait for IAB ws results so we can adapt shim6 if this makes sense?
[16:17:42] <wouter> Yes iljitsch, that question makes sense to me.
[16:17:54] <kurtis> Iljitsch: I can ask the question, and I also know the answer :)
[16:18:15] <kurtis> but let's wait until after the presentation
[16:18:19] <wouter> Slide: Shim6 proto issues.
[16:18:32] <wouter> Everyone is up to speed and discussion starts.
[16:18:36] <iljitsch> I'm eagerly waiting for that. :-) Yes, of course, I just wanted to get it typed in in time.
[16:19:03] <wouter> Marcelo is speaking on proto issues
[16:19:15] <wouter> Shim6 HBA Issues
[16:19:21] <wouter> IPR concerns
[16:19:45] --- richardn has joined
[16:19:58] <wouter> Ericsson has released a better IPR disclaimer, and issue acceptable for people\
[16:20:24] <wouter> Chair asks if there are any IPR issues. No reaction.
[16:20:39] <wouter> Marcelo continues
[16:21:24] <wouter> Chair asks if anyone has a problem with CGA compatible choices. No reaction.
[16:21:52] <wouter> Geoff asks on key exchange issues
[16:22:05] <wouter> and SHA1 dependency issues. Is this all resolved?
[16:22:14] <wouter> Are there any issues on security consideration on HBA?
[16:22:24] <wouter> Olaf goes to mike.
[16:22:38] <wouter> Minutes show silence/
[16:22:48] <wouter> Brian goes to mike.
[16:23:01] <wouter> Does not understand reference to multiple hash cga draft
[16:23:15] <wouter> response to one of the coments.
[16:23:43] <wouter> The dependency on sha1 in cga. Included a paragraph, and another doc to allow hash function agility for all cga-s
[16:23:54] <wouter> Brian: is that a normative reference.
[16:24:00] <wouter> Not sure. But you need to know
[16:24:15] <wouter> Please find out soon, we need to know
[16:24:40] <wouter> Question from. That is a problem anyway, may not need a normative refernce.
[16:24:47] <wouter> The sha1 dependency issue
[16:25:00] <wouter> Not a normative reference.
[16:25:21] <wouter> Jim from HP: more to PS because that causes implementation
[16:25:26] --- arifumi has joined
[16:25:49] <wouter> You will not see wide implementation, and IPR issues, and people do not believe this will work.
[16:26:02] <wouter> Jim wants to get it out there as a PS
[16:26:29] <wouter> But do not get your hopes up. The internet community, ISPS, govts, are not buying into this
[16:26:36] <wouter> Are we chasing this correctly?
[16:26:40] <iljitsch> maybe this is a good moment for my comment that I typed before.
[16:26:47] <iljitsch> can someone relay?
[16:26:48] <wouter> But that has nothing to do with pushing to PS.
[16:26:52] <wouter> Kurt will you relay?
[16:27:19] <wouter> Iljjitsches comment channeled by Kurt
[16:27:41] --- petrescu7 has joined
[16:27:42] <wouter> Yari: personal perspective
[16:27:43] <kurtis> done
[16:27:51] <iljitsch> thnx
[16:27:53] <wouter> We know they are not sufficient.
[16:28:00] <wouter> Tonights plenary
[16:28:15] <iljitsch> (jari arkko speaking right now)
[16:28:15] <wouter> Wants go get existing efforts to go on .
[16:28:27] <wouter> Don't want to stop because of the bigger issues.
[16:28:42] <wouter> Brian carpenter, agrees with yari
[16:29:28] <wouter> Maybe add features that could be of use to shim6t
[16:29:42] <wouter> \Someone agrees with previous speakers.
[16:29:52] <wouter> Kurtis agrees
[16:30:16] <wouter> Wants to get it done so people can decide to use it.
[16:30:29] <wouter> If we change it afterwards this is not the first time this are changed/
[16:30:38] <wouter> Marcelo continues
[16:31:06] <wouter> Protocol document
[16:31:32] <wouter> Interaction with IPsec
[16:32:15] <wouter> Proposed solution is to implement shim6 in the hardware for the BITW stuff
[16:32:44] <wouter> Comments asked by chair, Brian - only logical response.
[16:32:57] <wouter> But we may not see that implemented
[16:33:02] <wouter> Marcelo:
[16:33:33] <wouter> It is possible to put SHIM6 above IPsec.
[16:34:08] <wouter> This is not to protect shim6
[16:34:14] --- arifumi has left
[16:34:18] <wouter> but it would work
[16:34:21] --- arifumi has joined
[16:34:56] <wouter> Geoff(chair): do you think we need to pursue this?
[16:35:00] <wouter> And prepared to do work?
[16:35:18] <wouter> Erik: does not see need to do work
[16:36:34] <wouter> This doesn't preclude doing IPsec tunnel mode on the wire
[16:36:38] <wouter> suggested text for document
[16:37:08] <wouter> Issue: provide shim6 security based on IPsec SAs.
[16:37:17] --- josoinin@jabber.org/Meebo has joined
[16:37:52] <iljitsch> I'm a bit behind on the mailinglist and I'm not sure what's up on the screen now, but I'm pretty sure we concluded at some earlier point that you can't have bump in the wire AND use shim6 implemented in the host at the same time, i.e., you must disable shim6 OR BITW, or if you want to use both, implement shim6 together with BITW
[16:38:23] <wouter> The issue is already done. If you do BITW on both it works.
[16:38:34] <wouter> This is about protecting shim6 with ipsec
[16:38:50] <iljitsch> yes but what if you don't? there is no way to be sure on the protocol level.
[16:39:15] <wouter> Kurtis poses Iljitches question
[16:39:25] <wouter> That is what has been concluded says Marcelo
[16:39:35] <wouter> You have to implement shim6 in the hardware.
[16:39:53] <wouter> Ex mobike cochair. Don't do there, already a protocol that does that. And it works with NATs and so on.
[16:40:08] <wouter> Chari lastcall on shim6? Silence
[16:40:33] <wouter> Support mulitple security mechanisms
[16:41:19] <wouter> Chair: is what we have good enouygh. Silence. So good enough
[16:41:43] <wouter> DoS based on 2^47 tag space
[16:41:56] <wouter> is 4way handshake enough?
[16:42:21] <wouter> What is documents editors viewpoint on this?
[16:42:24] <wouter> Erik to the mike
[16:42:39] <wouter> Analogous to TCPe attacks
[16:43:03] <wouter> Like doing TCP handshakes
[16:43:09] <wouter> We already have this
[16:43:23] <wouter> Another comment. If you have a number you can attack it.
[16:43:27] <wouter> No worse than today
[16:43:48] <wouter> About forking
[16:44:09] <wouter> By req of transport ADs
[16:44:17] <wouter> Geoff explains the background
[16:44:41] <wouter> For thing s like VOIP you want different than TCP
[16:44:56] <wouter> so you have the capability to treat some apps differenly in t he shim6\
[16:45:15] <wouter> The exact way to signal this, was deemed an extension instead of core spec
[16:45:41] <wouter> Reviewer Dave answers; if was kinda weird architecture. but this is sufficient
[16:46:26] <wouter> Keep using renumber ULIDs
[16:46:36] <wouter> security issues raised
[16:46:46] <wouter> proposal to remove recommendation from base spec
[16:47:04] <wouter> Brian: Not sure he understand
[16:47:32] <wouter> At what point do you have to kill the shim6. after deprecation period. make clear in text
[16:48:09] <wouter> Another: Why is this a security risk?
[16:48:51] <wouter> (Stig) kept inside the site
[16:48:56] <wouter> Geoff explains.
[16:49:28] <wouter> If the same host starts to talk to the renumbered number that is reused.
[16:49:53] <wouter> This will create confusion.
[16:50:17] <wouter> Jim: agrees.
[16:50:36] <wouter> Please write this correctly\
[16:50:47] <wouter> Do this garbage collection
[16:50:54] <wouter> Thanks for a good job
[16:51:07] <wouter> CGA key length
[16:51:43] <wouter> Does shim6 need a minimum length, fixed, negotiate
[16:51:58] <wouter> Geoff. Chair. We use CGA not define them.
[16:52:10] <wouter> normative refernce them, let them figure it out.
[16:52:16] <wouter> Why would we in shim6 dive into cga
[16:52:20] <wouter> Jari
[16:52:50] <wouter> What is our x?
[16:53:09] <wouter> Doesn't care too much. More about length of address, than public key
[16:53:32] <wouter> Basically the same as sha1 negotiation
[16:53:46] <wouter> Put number in same spec
[16:53:59] <wouter> Proposes to use same number as in the sen(?) spec
[16:54:30] <wouter> any issues with broken flag? None
[16:54:39] <wouter> Failure detection issues
[16:54:49] <wouter> Mike to Jari
[16:54:59] <wouter> A few reviews to the list
[16:56:36] <wouter> Would be useful comments on state machine picture
[16:57:22] <wouter> Any discussinoo n the points that Jari related
[16:57:32] <wouter> Tmo Hendersson. Regarding substates in state machines.
[16:57:43] <wouter> Do not care in substate introduction. Some word changes.
[16:57:51] <wouter> What was on and off at what time
[16:57:58] <wouter> It could be clarified.
[16:58:16] <wouter> Put in the negatiation to get the time value will provide text to the list
[16:58:32] <wouter> Any other issues we want to take up failure detection? No one\
[16:58:38] <wouter> Geoff returns to the mike
[16:59:16] <wouter> Will do wg LC again. This time ample opportunity to comment with rough consensus they will go to IESG
[16:59:34] <wouter> Next agenda item. Update drafts.
[16:59:55] <wouter> API draft updates, then locator pair sleection draft uipdate. then implemnentation update.
[17:00:05] <wouter> Shinta will show his slides on the API draft
[17:00:34] <wouter> Individual draft
[17:02:23] <wouter> Changes.
[17:05:42] <wouter> Things and next steps
[17:07:02] <wouter> Dave taylor ccomments.
[17:07:36] <wouter> If you support preference on a per send basis, like in sendmsg ancillary, so no problem with different apps.
[17:07:47] <wouter> No these are pre sockets.
[17:08:23] <wouter> Dave there are ones that are per packets. And then you cna translate as if it were done on a perpacket basis. And then no problem. Like in ancilarry data option.
[17:09:00] <wouter> Erik. ancillary data question. One way is that this causes a context to fork
[17:09:24] <wouter> Maybe forking should be explicit
[17:10:06] <wouter> We need to sort out when to fork
[17:10:23] <wouter> If app has different prefs on every sendmsg.
[17:10:50] <wouter> Requires some thought
[17:11:11] <wouter> Dave dangerous to modify anything based on apps/socket behaviour
[17:11:15] <wouter> default is OK.
[17:11:19] <wouter> ancilla
[17:11:27] <wouter> ancillary can define shim6 defaults
[17:11:35] <wouter> Have some common default that never changes
[17:12:25] <wouter> Erik: sockets and preferences and forking
[17:12:40] <wouter> Geoff: you saying the same thing
[17:12:52] <wouter> put this on the mailing list if it has more meaning
[17:13:24] <wouter> Marcelo on the agenda for this locator paire selction draft
[17:15:00] <wouter> Changes
[17:16:04] <wouter> To do
[17:16:04] <wouter> Does any one want sometyhings.
[17:16:04] <wouter> Brian to mike
[17:17:08] <wouter> Problem is that shim6y may work OK for a small client, but may be real prpoblem for alarge server with thousands of clients.
[17:17:08] <wouter> Keeping all that state.
[17:17:08] <wouter> Marcelo: his xml doc has a commented section on that.
[17:17:08] <wouter> Erik: are you asking for numbers? dunno, the problem was raised
[17:18:06] <wouter> Be explicit about role pattersn causing different ratios between applifcation transport and shim6 state
[17:18:06] <wouter> Multiple hort connections from same IP and shim6 state
[17:18:06] <wouter> Could have signification amount of memory in shim6
[17:18:18] <wouter> There are things that change tha mount of memory being consumed
[17:18:59] --- jasso1 has joined
[17:19:36] <wouter> ~Cost of TCP relative to cost of shim6 state
[17:19:36] <wouter> Geoff: are you dictating here? Or are you going to use keyboard yourself
[17:19:36] <wouter> Erik goesz one
[17:19:36] <wouter> goes on with talking
[17:20:05] <wouter> About the nature of the application
[17:20:27] <wouter> App preferer referer, doc is cited, is not it
[17:20:44] <wouter> Geoff and Erik docide on some text to bounce to the mailing list
[17:21:18] --- jasso1 has left
[17:21:36] <wouter> Jim: for deployment , if we can get this working in Wimax or multihoming reployment strategy
[17:21:40] <wouter> code in linux, handhelds
[17:22:12] <wouter> Iljisch, you're on the slide.
[17:22:28] <wouter> default locator pair selection slide.
[17:24:03] <wouter> To do
[17:24:32] <wouter> do you only allow reasonable scoped addressed? (linklocal vs other types)
[17:24:42] <wouter> Anyone opinion? No one
[17:24:50] <wouter> Formula to exprtess the set of rules
[17:25:30] <wouter> Dave again; the rfc has some constraints and prefs, is that not suffficient?
[17:25:45] <wouter> rfc talks about destination not source? No it does talk about both
[17:25:55] <wouter> Yes it talks about src and dest combinations
[17:26:15] <wouter> Why need other set of rules?
[17:26:27] <wouter> Role to integrate shim6 information in to this set of rules
[17:27:16] <wouter> No ther ccomments: silence
[17:27:27] <wouter> Kunwoo on implementation
[17:27:34] <wouter> this is the next agenda item
[17:27:52] --- fparent@jabber.org has left
[17:27:57] <wouter> blue screen slideshow.
[17:28:41] <wouter> Slides are up. kunwoo about implementing shim6 protocol
[17:29:06] <wouter> He follows the slides.
[17:29:28] <wouter> The demonstration does not seem to be working.
[17:29:45] <wouter> no webpages
[17:29:45] --- juampe has left
[17:29:57] <wouter> He lists email addresses in his presentation
[17:29:59] --- keyajima has joined
[17:30:34] <wouter> For linux
[17:31:03] <wouter> Question The listr of feature in pahe one?
[17:31:13] <wouter> phase one ((Dino asks)) no answer that I could tell.
[17:31:32] <wouter> Slide on features
[17:31:46] <wouter> Blue is what they have done, black lines they have not done
[17:34:03] <wouter> Any questions
[17:34:05] <wouter> None
[17:34:24] <wouter> Last point on agenda. Next steps on working group.
[17:34:43] <wouter> Chairs believe that LCs and people are implementing it.
[17:34:49] <wouter> Need experience with real world.
[17:35:04] <wouter> Proposal no wg meeting in prague
[17:35:05] --- keyajima has left
[17:35:21] <wouter> There is nothing to do in the wg now.
[17:35:31] <wouter> Maybe later implementation or discussion.
[17:35:43] <wouter> Anyone have any views on this? Proposed by the chairs.
[17:36:02] <wouter> Jari: Get protocol out there and implemeneted. Additional features are not so useful
[17:36:13] <wouter> See if it is useful, what you need on top of it.
[17:36:18] <wouter> Have something real
[17:36:23] <wouter> Then take a new look at it.
[17:36:40] <wouter> No comments on that? Geoff closes the meeting.
[17:36:58] <wouter> See each other in Chcago or later depending on how busy the implementing has been.
[17:37:32] --- petrescu7 has left
[17:37:38] --- dudi has joined
[17:37:48] <wouter> Session is closed.
[17:38:34] --- ks has left
[17:38:39] --- kurtis has left
[17:38:46] --- richardn has left
[17:39:42] <wouter> Bye
[17:39:46] --- wouter has left
[17:39:57] --- iljitsch has left
[17:40:33] --- dudi has left
[17:44:11] --- nm has left: Replaced by new connection
[17:44:13] --- nm has joined
[17:44:28] --- arifumi has left
[17:46:02] --- josoinin@jabber.org/Meebo has left
[17:54:11] --- peetu has left
[17:57:27] --- becarpenter has left
[18:12:18] --- nm has left
[18:28:57] --- mo7sen has left
[19:06:48] --- becarpenter has joined
[19:07:14] --- isomer has joined
[19:08:00] --- isomer has left
[19:23:26] --- becarpenter has left
[20:05:31] --- mo7sen has joined
[20:05:41] --- mo7sen has left