IETF
v6ops
v6ops@jabber.ietf.org
Wednesday, 16 November 2011< ^ >
Patrik Halfar has set the subject to: v6ops -- IETF80
Room Configuration

GMT+0
[00:04:00] Bjoern A. Zeeb leaves the room
[00:44:07] yaovct joins the room
[00:44:42] furry joins the room
[00:53:23] Joel Jaeggli joins the room
[00:54:58] yaovct leaves the room
[00:57:36] <Joel Jaeggli> v6ops will begin at 0900
[01:00:00] naito.kengo@jabber.org joins the room
[01:00:06] ggm joins the room
[01:00:10] Colman Ho joins the room
[01:00:38] TACHIBANA toshio joins the room
[01:01:15] <Joel Jaeggli> good morninging
[01:01:21] <Joel Jaeggli> meeting is starting
[01:01:27] <Joel Jaeggli> agenda we have 10 documents
[01:01:45] <Joel Jaeggli> 3 have been accepted as wg documents
[01:02:01] <Joel Jaeggli> the rest of these (7) are new proposals
[01:02:42] <Joel Jaeggli> agenda bashing?
[01:02:58] <Joel Jaeggli> lee howard says the shorter agenda is greatly appreciated
[01:03:00] Lee Howard joins the room
[01:03:05] Extensa joins the room
[01:03:36] <Joel Jaeggli> slides rfc 6204bis draft
[01:03:58] <Joel Jaeggli> discvoered a dhcpv6 deplyment problem that is impacting this document
[01:04:16] <Joel Jaeggli> presented by ole troan
[01:04:26] <Lee Howard> presented by Chris Donley.
[01:04:32] <Joel Jaeggli> oops
[01:04:42] <Joel Jaeggli> yes
[01:04:49] EuiJong Hwang joins the room
[01:05:07] <Joel Jaeggli> dhcpv6 deployment problem
[01:06:07] <Joel Jaeggli> dhcpv6 has a retry cycle of every 6.5 minutes forever if no address are available.
[01:06:18] <Joel Jaeggli> currently no way to disable dhcpv6
[01:07:45] <Joel Jaeggli> added langauge to not require dhcpv6 if m = 0
[01:08:55] Atarashi Yoshifumi joins the room
[01:10:27] <Joel Jaeggli> not included int this draft is the ongoing lan work (homenet) and dhcpv6 server storm (dhc wg)
[01:10:32] tsavo_work@jabber.org/Meebo joins the room
[01:11:11] kawashimam joins the room
[01:11:34] <Joel Jaeggli> fred - question - the chairs of pcp are going to take the critical draft forward friday does that fit?
[01:11:35] mcharlesr joins the room
[01:12:01] <Joel Jaeggli> chris d - yeah by early decement
[01:13:02] <Joel Jaeggli> chris d - open issue should we explicity follow the hint in the RA and disable dhcpv6 if m=o=0
[01:13:27] <Joel Jaeggli> ralph d - trying to think of the rgith word? culprit?
[01:14:11] <Joel Jaeggli> the arch detail to be careful about the mo bits describe behavior for all devices on the link
[01:14:43] <Joel Jaeggli> it is to be used for controlling individual device on the link
[01:15:02] Bjoern A. Zeeb joins the room
[01:15:02] <Joel Jaeggli> advise caution for using it for diffrential control on the link
[01:15:21] dudisaki joins the room
[01:15:21] <Joel Jaeggli> dhcp is an individual device configuration protocol
[01:15:52] <Joel Jaeggli> work going on on a return status code that says go away and stop bothering me
[01:16:24] <Joel Jaeggli> fred - dhc is looking at the problem
[01:16:46] <Joel Jaeggli> worried that this would be different than other users of the m and o bits
[01:17:02] <mcharlesr> http://junk.sandelman.ca/junk/ietf82 for PDF of all PPT/PPTX.
[01:17:15] <mcharlesr> (for this WG)
[01:18:07] <Joel Jaeggli> raplh d I can't predict how these to options will progress through the ietf
[01:18:22] <Joel Jaeggli> fred - get some guidance from dhc chairs?
[01:18:46] <Joel Jaeggli> john browzowski - ralphs assessment is fair and accurate
[01:19:13] pselkirk joins the room
[01:19:25] <Joel Jaeggli> fred - I don't want to make a specification that's different that someone's elses.
[01:20:11] <Joel Jaeggli> ted l - we already have the draft unless there is object in dhc I think this will be straight forward to refer to one or the other
[01:20:45] Suz joins the room
[01:20:46] tatsuji joins the room
[01:20:49] <Joel Jaeggli> stuart chesire - for how long do we turn it off?
[01:21:12] satoru.matsushima joins the room
[01:21:14] Jiang Dong joins the room
[01:22:22] <Joel Jaeggli> raplh maximum solicit retry value new option could allow the provider to tell the client to go away for an hour 10 days whever is appropiate
[01:22:48] arifumi joins the room
[01:23:30] <Joel Jaeggli> ? the dhc option sounds like a better option that figuring out new semantics for m and o bits
[01:23:42] vikas.sarawat joins the room
[01:24:04] <pselkirk> could chris or fred update the slides on the datatracker? the slides that are there don't have the dhcp issue.
[01:24:12] Bjoern A. Zeeb leaves the room
[01:24:51] <Joel Jaeggli> yeah I see that
[01:25:16] <Joel Jaeggli> chris d - we've had the cableabs erouter spec out since 2006
[01:25:28] <Joel Jaeggli> weve been testing those for years now.
[01:25:39] satoru.matsushima leaves the room
[01:26:11] <Joel Jaeggli> we need a vehicle to get the transtion technologies plus refinements to 6204
[01:27:25] Bjoern A. Zeeb joins the room
[01:29:17] <Joel Jaeggli> john b - I have a hard time sticking with the current 6204 document
[01:29:51] <Lee Howard> Mark Townsley talking, unidentified
[01:29:59] <Joel Jaeggli> mark t - let me start with dhcpv6
[01:30:50] <Joel Jaeggli> on 6204bis in general I think 6rd sunset requirements need to be incorporated
[01:31:29] <Joel Jaeggli> dslite discussion needs to cover what happens when you turn of ipv4, (we don't discuss that here)
[01:31:35] <Joel Jaeggli> more work to be done
[01:31:52] <Joel Jaeggli> don't rush the bis so that we get even more mistakes
[01:32:18] <Joel Jaeggli> tell cpe manufacturers
[01:32:38] <Joel Jaeggli> fred - where will sunset be discussed.
[01:33:00] <Joel Jaeggli> mark t - here or softwire
[01:33:44] <Joel Jaeggli> alain chair of softwire - it would be ok to ship 6rd but we need a document covering using it.
[01:33:58] <Joel Jaeggli> fred - we'll add this to the agenda for tomorrow...
[01:34:44] <Joel Jaeggli> lorenzo - suffering from split brain - dhcpv6 go away until I tell you seems like a hack on a hack
[01:35:05] evyncke joins the room
[01:35:22] Satoru Kanno joins the room
[01:35:47] <evyncke> Just wondering whether an ACL blocking DHCP packets coming from some CMTS/interfaces would not be easier...
[01:36:10] <Joel Jaeggli> who is speaking
[01:36:26] <evyncke> D)link?
[01:36:43] <Joel Jaeggli> if 6204 is "ok" then maybe we should take our time
[01:37:12] chiyuan joins the room
[01:37:30] <Joel Jaeggli> john b - this is going to be incremental in nature
[01:38:33] <Lee Howard> That was Hans Liu from D-Link, yes
[01:39:17] <Joel Jaeggli> ted l - time option squaelch thing you could include in the squelch message a nonce so that you could later send a reconfigure instruction
[01:40:16] Bing Liu joins the room
[01:40:16] <Joel Jaeggli> multiple routers draft due to disconncect with how we use dhcpv6
[01:40:37] <Lee Howard> Erik - I think the problem with an ACL is that Brzozowski wants to enable DHCPv6 for some customers on a CMTS, but not others, and writing an ACL that explicitly denys or allows DHCPv6 packets (from a device that doesn't yet have an address) has scaling problems.
[01:41:08] <Joel Jaeggli> roger - regarding the problem - it's unwanted traffic
[01:41:15] <Joel Jaeggli> they get filtered by acls
[01:41:17] <evyncke> Lee: ok in this case but this is a weird deployment plan (or is it only in the pilot phase)
[01:41:40] <Joel Jaeggli> this is dhcp for dhcp
[01:42:10] <Joel Jaeggli> this type of coupling will result in a clever scheme
[01:42:24] <Lee Howard> the Mark Townsley who must not be named
[01:42:49] <Joel Jaeggli> mark t - we have a showstopper that others will run into as well
[01:42:50] <ggm> that is not the mark townsley you're looking for
[01:42:54] igor.gashinsky joins the room
[01:43:03] <Joel Jaeggli> it's very important we hve a path for that
[01:43:15] <Joel Jaeggli> we're expanding the scope of the draft again
[01:43:26] <Joel Jaeggli> we will make mistakes again
[01:43:57] <Joel Jaeggli> target the showstoppers through surgical rfc's, keep the expanded scope and let it go at it's own pace.
[01:45:30] <Joel Jaeggli> ralph d - differental interprestation of m and o bits would imply that devices built in different eras with follow variations and have different behavior
[01:45:54] <Joel Jaeggli> needs more discussion we can address things that require quick action in other groups
[01:46:14] Dave Thaler joins the room
[01:46:27] <Joel Jaeggli> reoberta madgilone - we need dslite in this rfc
[01:46:46] <Joel Jaeggli> fred - take the timing discussion to the list.
[01:47:01] <Joel Jaeggli> marks discussion tomorrow.
[01:48:13] <mcharlesr> v6ops-6.* I think.
[01:56:02] Joel Jaeggli leaves the room
[01:56:45] Lee Howard leaves the room
[01:58:09] igor.gashinsky leaves the room
[01:58:29] pselkirk leaves the room
[01:59:03] Joel Jaeggli joins the room
[01:59:09] igor.gashinsky joins the room
[01:59:10] <Joel Jaeggli> approximately 30 participacnts have read draft
[01:59:36] <Joel Jaeggli> Stateless Source Address Mapping for ICMPv6 Packets <http://tools.ietf.org/agenda/82/slides/v6ops-3.ppt>
[01:59:48] <mcharlesr> v6ops-3.*
[02:01:09] <mcharlesr> I wonder if this IPv4 /24 prefix, could come from class D space :-)
[02:01:35] <Dave Thaler> class D is multicast, this is unicast
[02:01:42] <Joel Jaeggli> E
[02:01:43] <mcharlesr> sorry, I meant class E.
[02:02:30] <ggm> oh god, please not that again
[02:02:31] satoru.matsushima joins the room
[02:02:37] <Joel Jaeggli> lol
[02:03:01] gih joins the room
[02:03:11] <evyncke> Class D would be fun though
[02:03:20] Warren Harrop joins the room
[02:04:16] fdupont joins the room
[02:06:52] <Joel Jaeggli> concerns ?
[02:06:56] <Joel Jaeggli> joel -ddos
[02:07:10] <Joel Jaeggli> other than traceroute does this fix anything else?
[02:07:57] <Joel Jaeggli> ron bonica - hvae you considered looking at rfc 5387?
[02:09:27] <Joel Jaeggli> hiroaki hazeyama - summary expirements
[02:14:47] EuiJong Hwang leaves the room
[02:17:38] EuiJong Hwang joins the room
[02:20:55] <Joel Jaeggli> who was speaking?
[02:21:20] <Joel Jaeggli> fred - have you been in contact with the vendors associated with the problems?
[02:21:54] <Joel Jaeggli> hiroaki - yeah in contact with iij, about the AAAA response
[02:22:47] Atarashi Yoshifumi leaves the room
[02:22:47] <Joel Jaeggli> hisaki-san - first time we've had this scale at the camp. iij has their own implemtnation teamp. we continue to run this kind of testbed operation.
[02:23:00] <Joel Jaeggli> fred -other comments?
[02:24:16] <Joel Jaeggli> victor k - wireline incremental
[02:24:27] <Joel Jaeggli> lee howard has offered to help me moving forward
[02:25:17] <arifumi> esaki-san he is. :)
[02:25:42] <Joel Jaeggli> took a position on how an operator can use a phased approach
[02:25:44] <Joel Jaeggli> thanks
[02:26:34] <mcharlesr> where are we now?
[02:26:41] <mcharlesr> (damn telephone)
[02:27:45] <Joel Jaeggli> http://tools.ietf.org/agenda/82/slides/v6ops-4.pptx
[02:29:58] <Joel Jaeggli> hisaki = flames right?
[02:30:27] Lee Howard joins the room
[02:30:44] <Joel Jaeggli> transition phases
[02:31:25] <Joel Jaeggli> reality is operators will run out of v4 at some point during transition
[02:32:38] <Joel Jaeggli> do we put this into a wg document?
[02:32:45] <Joel Jaeggli> fred comments?
[02:32:46] vikas.sarawat leaves the room
[02:32:58] <Joel Jaeggli> brian c - I like the idea,
[02:33:18] <Joel Jaeggli> it's not ready for last call yet. could be very useful
[02:33:26] <Joel Jaeggli> john b - I tihink it's a good idea
[02:33:53] <Joel Jaeggli> this is probably a more scalable mechanism of sharing information.
[02:34:16] igor.gashinsky leaves the room
[02:34:31] <Joel Jaeggli> lee howard - small isps come to me, we think it has value?
[02:35:15] <Joel Jaeggli> fred 20ish people have read it.
[02:35:26] <Joel Jaeggli> hum in favor as awg document?
[02:35:33] <Joel Jaeggli> some in favor,
[02:35:39] <Joel Jaeggli> none opposed
[02:36:47] <Joel Jaeggli> fred - in part because of my companies marketing . the objective is to get the other side of the bridge.
[02:37:28] <Joel Jaeggli> brian c - question of cgns
[02:37:57] <Joel Jaeggli> fred those people who are going to do it are going to if they have to
[02:38:23] <Joel Jaeggli> ole t - we 19 active drafts in softwire that propose how to do this.
[02:38:35] <Joel Jaeggli> it's very much of a moving target
[02:39:26] evyncke leaves the room
[02:39:28] <Joel Jaeggli> victors understand there's multiple approaches.
[02:39:42] <Joel Jaeggli> here are the things you need to think about.
[02:40:28] <Joel Jaeggli> ole would you be fine if I proposed a section on a+p...
[02:40:41] <Joel Jaeggli> lee hoawrd I would love a handing guide book.
[02:41:49] <Joel Jaeggli> operational guidance shouldn't cover things that aren't in draft form.
[02:42:30] evyncke joins the room
[02:43:23] <Joel Jaeggli> joel j as a consumer I'm not going to use something that isn't commercial available.
[02:44:02] <Joel Jaeggli> chris d - reasoanble to put a pointer somewhere to the cgn work
[02:44:43] <Joel Jaeggli> renee depress - we have a an defintional challange around cgn / nat44 vs dslite
[02:45:02] <Joel Jaeggli> lee howard (do we need a taxonomy document?)
[02:45:47] Bjoern A. Zeeb leaves the room
[02:45:48] <Joel Jaeggli> christian france telecom - cgn are no solutions and there are lesser evils.
[02:46:02] <Joel Jaeggli> we will deploy them whether we like them or not.
[02:46:30] Bjoern A. Zeeb joins the room
[02:46:47] Warren Harrop leaves the room
[02:46:59] <Joel Jaeggli> marc blanceht - phase 3 tunneling approach or translation e.g. nat64
[02:47:43] <Joel Jaeggli> victor - not objecting to discusion, but would I do it in a wireline network? no
[02:48:04] <Joel Jaeggli> marc b - not saying should or showld not but it's a kind of generic thing
[02:48:45] <Joel Jaeggli> hui deng -
[02:49:05] <Joel Jaeggli> lorenzo − nat64 by itself is not a solution.
[02:49:27] <Joel Jaeggli> things have to work in the presence of v6only first
[02:49:28] gih leaves the room
[02:50:17] <Joel Jaeggli> i'm saying I disagree v6only doesn't work weel anywhere.
[02:50:34] <Joel Jaeggli> neorjn - fix the applications
[02:51:46] <evyncke> Interesting twist: Lorenzo acting as a vendor :-) (and his point was right BTW)
[02:53:41] <Joel Jaeggli> they bought motorola
[02:53:49] <Joel Jaeggli> brain carpenter
[02:53:50] <Joel Jaeggli> IPv6 Flow Label for Server Load Balancing <http://tools.ietf.org/agenda/82/slides/v6ops-9.pdf>
[02:53:57] pselkirk joins the room
[02:54:07] <mcharlesr> thanks.
[02:54:14] <Joel Jaeggli> Brian
[02:54:21] <Joel Jaeggli> brain is somebody else
[02:55:03] <mcharlesr> I dunno... there is a resemblance, particularly if Brain finally perfected the make himself taller hack.
[02:56:37] Extensa leaves the room
[02:58:00] Victor Yang joins the room
[02:58:21] <Joel Jaeggli> joel - l3 ecmp
[02:59:04] <Joel Jaeggli> lorenzo - note you have to do something to address icmp packet to big going to the wrong proxy
[02:59:23] Victor Yang leaves the room
[02:59:46] yaovct joins the room
[02:59:52] <mcharlesr> Joel, google finds many references to L3 ECMP, but no definitive reference.
[03:00:29] frnkblk joins the room
[03:01:00] <mcharlesr> http://en.wikipedia.org/wiki/Equal-cost_multi-path_routing
[03:01:36] gih joins the room
[03:02:16] <Joel Jaeggli> I've implmented it
[03:02:29] <Joel Jaeggli> but yeah
[03:03:17] <frnkblk> I'm coming in late -- how did the rfc6204bis discussion go?
[03:03:25] <Joel Jaeggli> brian - this method for identify a flow is useful or not?
[03:03:44] <Lee Howard> frank. . . It was extended.
[03:03:48] <frnkblk> Great.
[03:04:12] <frnkblk> Clarify: bis was adopted, or the discussion was extended?
[03:04:27] <Joel Jaeggli> ron bonica - for people doing true load balancers this might not be very helpful.
[03:04:54] <Joel Jaeggli> I don't think you can build a stateless load balancer and still support pmtud
[03:05:08] <Joel Jaeggli> (lorenzo)
[03:05:19] <Joel Jaeggli> I tried it and it didn't work
[03:05:21] <mcharlesr> one way is for all ICMP messages to basically be sprayed at all servers....
[03:05:30] <furry> ok, this ICMP <-> balancer mapper was always a problem for v4 as well..even for balancers pretending to be 'stateful' ;(
[03:05:35] <Joel Jaeggli> brian _ I would like feedback
[03:05:42] <Lee Howard> there was extensive debate (including among the authors and major contributors) about whether it should move ahead quickly so implementers could reference it, or whether it needed more time so we could make sure we get it right and don't need rfc6204ter (or a 6204WG)
[03:06:03] <Joel Jaeggli> yes it is
[03:06:12] <Joel Jaeggli> -furry
[03:06:17] Bjoern A. Zeeb leaves the room
[03:06:37] <frnkblk> @Lee: was a decision/consensus reached, or more tomorrow?
[03:06:51] <Lee Howard> anyone else able to characterize the result?
[03:07:00] Bjoern A. Zeeb joins the room
[03:07:03] gih leaves the room
[03:07:07] kawashimam leaves the room
[03:07:24] chiyuan leaves the room
[03:07:29] Bing Liu leaves the room
[03:07:46] satoru.matsushima leaves the room
[03:08:17] EuiJong Hwang leaves the room
[03:08:28] <Lee Howard> authors in front of me now debating what happened
[03:08:47] TACHIBANA toshio leaves the room
[03:08:57] EuiJong Hwang joins the room
[03:09:07] <frnkblk> @Lee: OK.
[03:09:35] <frnkblk> @Lee: have them stand in front of the mike and discuss. =)
[03:10:11] <Lee Howard> meeting broke up
[03:10:40] <frnkblk> OK
[03:10:42] <Lee Howard> there was not consensus on how/whether to advance.
[03:10:48] Colman Ho leaves the room
[03:10:54] <Lee Howard> I imagine the list will be lively
[03:11:02] ggm leaves the room
[03:11:24] <Joel Jaeggli> complete for the day
[03:11:32] furry leaves the room
[03:11:52] <frnkblk> Was there concern that there might not be enough momentum to complete a ter? or concern about vendor frustration?
[03:12:05] naito.kengo@jabber.org leaves the room
[03:12:07] Suz leaves the room
[03:12:08] fdupont leaves the room: Computer went to sleep
[03:12:29] <Lee Howard> What I heard from vendors:
[03:12:47] Satoru Kanno leaves the room
[03:12:55] <Lee Howard> Hans from D-Link said to stop changing things, we were burning out his R&D team, and they would soon dig in their heels.
[03:13:35] tatsuji leaves the room
[03:13:45] <Lee Howard> Townsley from Cisco/Linksys said that we should look at his 6rd-sunsetting draft (posted yesterday) and think about IPv4 sunsetting under DS-Lite, and make sure all of that got included, rather than doing 6204bis and then a 6204ter.
[03:13:45] <Joel Jaeggli> I think we can soak up some open mic time on thursday
[03:14:14] evyncke leaves the room
[03:14:16] Joel Jaeggli leaves the room
[03:14:19] <Lee Howard> I have to run to another meeting; I'm sure a post to v6ops asking what happened will receive lively response
[03:14:30] <frnkblk> Thanks, Lee.
[03:14:31] Lee Howard leaves the room
[03:14:52] <frnkblk> Thursday 1 pm is my time Wednesday 11 pm, so I hope to listen in tomorrow.
[03:15:52] <frnkblk> I'd rather put that sunsetting in a ter. There's going to be enough stuff coming from homenet that rfc6204 will need to be revisited anyways.
[03:17:00] <frnkblk> more precisely: ..."from homenet that rfc6204 will likely need to be..."
[03:17:56] frnkblk leaves the room
[03:19:16] EuiJong Hwang leaves the room
[03:19:25] Joel Jaeggli joins the room
[03:23:47] Jiang Dong leaves the room
[03:25:17] dudisaki leaves the room
[03:26:17] Dave Thaler leaves the room
[03:26:47] arifumi leaves the room
[03:27:17] Bjoern A. Zeeb leaves the room
[03:31:00] Bjoern A. Zeeb joins the room
[03:44:00] pselkirk leaves the room
[03:56:48] tsavo_work@jabber.org/Meebo leaves the room
[04:02:17] Bjoern A. Zeeb leaves the room
[04:15:30] Bjoern A. Zeeb joins the room
[04:32:57] Bing Liu joins the room
[04:46:18] Bjoern A. Zeeb leaves the room
[04:47:00] Bjoern A. Zeeb joins the room
[04:48:54] Joel Jaeggli leaves the room
[04:51:16] Bjoern A. Zeeb leaves the room
[04:54:05] Joel Jaeggli joins the room
[04:56:29] arifumi joins the room
[05:06:23] Satoru Kanno joins the room
[05:08:09] arifumi leaves the room
[05:08:43] pselkirk joins the room
[05:12:12] Satoru Kanno leaves the room
[05:25:42] gih joins the room
[05:25:50] gih leaves the room
[05:26:23] ggm joins the room
[05:26:31] ggm leaves the room
[05:26:31] ggm joins the room
[05:27:09] pselkirk leaves the room
[05:44:51] ggm leaves the room
[05:54:34] yaovct leaves the room
[06:05:15] ggm joins the room
[06:32:16] TACHIBANA toshio joins the room
[06:32:31] TACHIBANA toshio leaves the room
[07:06:01] Bing Liu leaves the room
[07:06:01] Bing Liu joins the room
[07:14:44] Bing Liu leaves the room
[07:16:32] ggm leaves the room
[07:17:20] Bing Liu joins the room
[08:12:49] Bing Liu leaves the room
[08:13:58] Joel Jaeggli leaves the room
[08:31:02] Colman Ho joins the room
[08:32:40] Colman Ho leaves the room
[08:42:58] Joel Jaeggli joins the room
[09:02:47] Bing Liu joins the room
[10:25:28] Joel Jaeggli leaves the room
[10:34:48] Joel Jaeggli joins the room
[10:36:35] Joel Jaeggli leaves the room
[12:21:16] mcharlesr leaves the room
[14:06:06] Bing Liu leaves the room
[14:06:08] Bing Liu joins the room
[15:34:54] Joel Jaeggli joins the room
[15:35:19] Joel Jaeggli leaves the room
[15:36:13] Joel Jaeggli joins the room
[17:37:53] Joel Jaeggli leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!