[14:59:04] lllmartini joins the room [15:03:23] one new rfc - [15:03:28] Mark Townsley joins the room [15:03:32] req for ms-pw has been issued [15:04:21] fc-encap draft went almost to completion [15:04:39] mattew and stewart got concerned about the structure of the draft [15:04:51] My notes for draft-ietf-pwe3-cep-mib-12.txt say that I am waiting on RFC Editor text from the authors [15:04:52] want to split into two componets [15:05:07] to address Dan R's concerns [15:05:12] fc encaps component , and fc reliability component [15:05:25] steve ulrich joins the room [15:05:56] Yakov: saw comment on list , and believes it is confusing to this here [15:06:36] David black : the thing to undestand , is that it is specific to pw in a number of ways [15:06:44] it is not done as general protocol [15:07:00] he does not want to see that protocol taken over in transport [15:07:16] Authors agreed to split [15:07:26] ( stewart said ) [15:08:52] no a single comment on ms-pw arch [15:08:57] we need to have it reviewd [15:09:03] cpignata joins the room [15:09:19] Peny joins the room [15:09:31] apparently since lot's of people are sitting in the first row , it must mean they all read the drafts ... [15:09:57] haence please send comments to the list on the ms-pw arch draft [15:10:37] a number of drafts - rfc have been out there for a while . [15:14:53] can anybody pick up the jabber when I talk ?? [15:15:17] p-to-mp , there have been 2 drafts subminnetd [15:16:04] please read these and prepare to discuss at the next IETF. [15:16:07] yjs joins the room [15:17:22] need a BPC on partitioning between NSP , and PW [15:17:55] stewart : we need to do some work , and also how we use the signalling protocols [15:18:06] we need guidelines on what is reasonable [15:18:12] to put i nthe signalling [15:18:38] First up: Ben , on BFD draft [15:21:02] ben basically read the slides , nothing more [15:21:11] next is Peter [15:21:43] pw OAM message mapping draft [15:22:08] thanks to Mustapha for editing the last version [15:23:06] bnsmith joins the room [15:24:05] re-organized draft - like all atm related information is now together in one ection [15:24:12] removed large tutorial [15:24:16] moved to sec 7 [15:24:25] great improovement in our view [15:27:33] I sent email to pwe3 about the MIB documents... Looks like on all 3 of those I am waiting on the authors to resolve issues identified during IESG review. [15:29:03] Satoru: [15:29:14] I think this draft has made good progress [15:29:32] i have comments on sinlgle loop emulated mode [15:29:47] i think it makes sense to use the pw status , and ac status [15:30:04] second comment: especially in the ATM pw case [15:30:22] there is nothing describing specifically segment oam , aor end-to end oam [15:30:26] please improove. [15:30:45] Peter Agrees , we need to improove this . [15:31:11] Mark Townsley leaves the room: Replaced by new connection [15:31:11] Mark Townsley joins the room [15:31:15] Satoru's point is that the Ac could be owned by the customer of the sp [15:31:52] so middle SP of ms-pw would not see atm AIS alarm if we do not sent an ac status message . [15:32:24] Peter believes that we can deal with this at the native service level. [15:32:36] peter propose , to discuss this with co-authors [15:32:52] yjs leaves the room [15:33:45] mathis joins the room [15:34:39] Luca question send AC down status as well as OAM? Peter yes could do that but not defined in the specs three options-- [15:35:19] Yaakov: we could do the other mode , and it should be the same [15:35:52] peter why does sp in the middle need to know [15:36:21] yaaakov , but what is sp in the middle get a call asking for an explenation about AIS ... [15:36:45] Italo: if you generate ais you have server layer alarm [15:37:10] we need to understand the boundaries behind this model [15:37:16] no more coments [15:37:38] Just to summarize : [15:37:42] ( peter ) [15:37:53] we need to consider Satoru's prpoposal [15:38:01] finish ATm part , and then request LC [15:38:15] Matthew : next on agenda , ethernet OAM [15:38:43] Nabil : is setting up to peak [15:38:48] speak [15:39:55] Lot's of fiddling with the MAC .... [15:40:39] ok giving up : we move to Yaakov [15:40:55] we'll get back to nabil later , once the MAC works ;-) [15:41:18] Yaakov : on PW bonding [15:43:35] yaakov is reading the slides [15:44:33] yakkov wants to request this should be WG draft [15:44:59] Stewart polling the room : [15:45:28] no over whelming support to do anything on the subject [15:45:39] we need to have more opinions. [15:46:24] next Ming, form VZ [15:51:04] missing slide : [15:51:17] what we want to do is get more input from SP/vendors [15:51:26] what are the requirements on this topic [15:51:35] which wg does this draft beling to [15:52:42] If you are going to give the same presentation to 3 groups, please tailor it for each... e.g., what does *this* have to do with PWE3? [15:52:45] I don't see it. [15:52:51] right .. [15:52:55] Looks like general MPLS stuff. [15:52:57] also this is a solution slide [15:53:01] (or ccamp) [15:53:06] "virtual interface " [15:53:14] that is a solution .... [15:54:04] thanks to fred , adrian, and Ron [15:54:55] Yakkov : hasing is a problem because we need to break the flow into idividual mico flows [15:55:03] this draft is compliment to hasing [15:55:19] always 2 trains moving foraws , te based , and non-te based [15:55:26] thsi is compliment to hasing [15:55:45] satoru joins the room [15:56:06] env is l2/l3vpn traffic into the nework [15:56:10] lion joins the room [15:56:15] traffic flow varies greatly [15:56:36] dwf joins the room [15:56:39] yaakov : how do you do traffic measuremt [15:57:05] answer : traffic measurement is done on input ctg virtual int [15:57:38] when we need to move traffic we take a small hit [15:58:01] we can make adjustment to parameters to move ... [15:58:27] Stewart : tictoc charis asking how do we make timing traffic cope with these moves. [15:58:38] asnwer both questions relate to solutions ... [15:58:58] giles.heron joins the room [15:59:54] stewart: there are some traffic of traffic thatcannot be moved [16:00:08] we need text in draft to make this clear . [16:00:33] yaakov : where this should be done [16:00:45] not here since this is an ip/mpls problem [16:01:03] George : Ad will need to figure out where this belongs to in ietf [16:01:16] we certainly want to help with a solution [16:01:48] Mark : since you are o nthe schedule , please tailorthe ppt to the particular WG [16:01:55] when you presetn i nthat WG [16:02:23] mark: answer why is this is important for this WG [16:02:48] Andy Malis: [16:03:01] out intention to bring to PWE3 was just for information . [16:03:11] ( andy is co-autothor ) [16:03:23] mathis leaves the room [16:03:45] NExt Nabil : with Fixed MAC. [16:04:17] Mattew : finally we have the pw ethernet oam . WG :-) [16:04:40] define procedures between eth PW and eth OAM [16:04:44] reading slides now . [16:07:16] Florin : [16:07:34] the way we try to use AIS , is in line on how we did it for ATM PW. [16:08:34] bringing it up i ncase somebody from ITU-T ha comments [16:08:44] audience : [16:09:01] the way this works is that we have to define two methods ... [16:09:19] ali : the way it is defined , there are multiple ways [16:09:39] notivication is consistent to y.1731 [16:10:04] Nabil : you need a map defined on Ac where the PW start [16:10:37] Italo : I think that reading the draft , my understanding this draft is addressing coupled model [16:11:13] we need to think about the cases , and arch defined in main oam msg mapping draft [16:12:03] Congestion control : David black [16:12:15] i'm from the transport area , and i will help you . [16:13:01] will try to give you a transport area view of this problem [16:17:35] giles.heron leaves the room [16:17:49] giles.heron joins the room [16:17:52] next step : design team [16:18:17] An excavator is in scope [16:18:36] ( must be a trend for all PWE3 participatns to buy excavators ;-) ) [16:19:08] Yakkov , I have a problem with your use of the word policing [16:19:25] leaving this aside i would live to see this work go forward [16:19:46] Dave Mcdysan [16:19:53] is been following this closely [16:20:11] is is difficult to get direction from the draft [16:20:37] I support trying to separate this transport from tcp [16:21:48] question : if we have PW for commercial custoemrs , and we havea failure that create congestion on the internet we should reduce pw traffic to accompodate the I ? [16:22:15] asnwer ( david ) : a carrier will have some policy , on most important traffic [16:22:31] mechanisms will be important to deal with failures [16:22:39] it's a carrier decision [16:22:45] on what to do [16:23:05] we need to define mechanisms to enable the carrier [16:23:51] stewart : a transpot network is what we are defining in mpls-tp [16:25:08] thomasdnadeau joins the room [16:25:17] david black : OAm mechanisms are not goint to be defined here . [16:26:25] yaakov : in tranport network , g805 defn ... [16:26:44] yous definition is different from g08 transport network [16:27:05] stewart , we are talking about restoring network to health [16:27:35] mscharf joins the room [16:28:27] Dave : pw comes form commercial /govn't customers [16:28:42] how many of these PWs can really be pre-emptable ? [16:29:04] if somebody is payiing a premium , it is difficult to [16:29:34] to pre-empt these [16:30:30] David , this is about spec of mechanisms not about specificating policies [16:30:35] Bruce Davie : [16:31:02] for a long time we said this pw is not really necessary , but it is necessary [16:31:12] if the have a problem of congestion of the network [16:31:30] policy question dave raised does come into the requirements [16:31:57] Dave : again the business we are here for is providing tools , not spec of policy [16:33:12] all design teams members that accepted are stewart , yaakov, luca, Bruce, [16:34:01] ( luca comment , Ben also accepted to be involved ) [16:34:52] end of agenda [16:35:07] first time ever that we finish 1 hour earlyer ! [16:35:13] end of meeting [16:35:19] lllmartini leaves the room [16:35:26] steve ulrich leaves the room [16:35:34] donley.chris joins the room [16:36:32] donley.chris leaves the room [16:37:52] satoru leaves the room [16:39:22] lion leaves the room: Computer went to sleep [16:40:03] lion joins the room [16:42:27] mscharf leaves the room [16:42:45] giles.heron leaves the room [16:45:09] Mark Townsley leaves the room [16:45:15] lion leaves the room [16:45:17] Mark Townsley joins the room [16:55:10] cpignata leaves the room [16:55:10] cpignata joins the room [16:58:13] dwf leaves the room [17:04:47] Mark Townsley leaves the room [17:28:03] Peny leaves the room [17:32:44] Mark Townsley joins the room [17:33:06] Mark Townsley leaves the room [17:35:38] Mark Townsley joins the room [17:38:53] Mark Townsley leaves the room: Replaced by new connection [17:38:54] Mark Townsley joins the room [18:08:31] thomasdnadeau leaves the room [19:01:01] Mark Townsley leaves the room [19:02:45] satoru joins the room [19:02:55] satoru leaves the room [19:08:28] Mark Townsley joins the room [19:08:50] Mark Townsley leaves the room [19:16:30] Mark Townsley joins the room [19:19:53] Mark Townsley leaves the room: Replaced by new connection [19:19:53] Mark Townsley joins the room [19:24:50] Mark Townsley leaves the room [21:02:57] cpignata leaves the room