[04:00:23] calvin joins the room [04:01:54] calvin leaves the room [04:06:47] scott.mansfield joins the room [04:07:36] marshall joins the room [04:08:10] yinxing joins the room [04:08:15] I will be jabber scribing this session [04:08:33] please remember, if you speak, to say your name [04:11:29] Glenn Parsons joins the room [04:13:08] rhe joins the room [04:13:29] lix joins the room [04:13:54] lix leaves the room [04:14:02] foobar joins the room [04:19:49] Sorry - this is [04:19:50] Using mLDP through a Backbone where there is no Route to the Root - [slides TBA] - http://tools.ietf.org/id/draft-wijnands-mpls-mldp-csc-02.txt [http://tools.ietf.org/id/draft-wijnands-mpls-mldp-csc-02.txt] - IJsbrand Wijnands [04:20:04] Daniel King joins the room [04:20:17] What if we temporatry replace the new FEC with a temporary new FEC [04:20:28] where the route is the IP address of the exit router [04:20:39] and the original FEC is encoded and retained [04:20:51] mLDP carriers carrier [04:20:56] (animated slided) [04:21:51] there is one label trough and trhough [04:22:18] when the customer looks at he FEC on the other end, the original FEC is still there [04:22:35] this solution works both in the global table and the VRF table [04:23:00] if it is in the VRF context, we add the RD [04:23:34] the Router Distinguisher is discovered dynamically via an unicast route lookup on [04:23:38] the c-root address [04:23:49] why come to L3VPN ? [04:24:13] the L3VPN part is small [04:24:33] but the authors would like to put it through both MPLS and L3VPN [04:24:45] Is the WG OK with this context [04:25:07] Ron Bonica : The draft is about MPLS but the motivation is about L3VPN [04:25:30] what's concerning me is that it defines new procedures for a carrier's carrier, but [04:25:41] we already have procedures for a carrier;s carrier [04:25:55] Ijsbrand : which procedures [04:26:09] Ron : mcast and mcast-bgp draft [04:26:17] satoru.matsushima joins the room [04:26:50] Thomas Morin : Draft Mcast in the BGP specification explained procedures for during this. The motivation for why another solution is needed would be useful [04:27:15] In what is alrady supported you have the ability to aggregate trees [04:27:42] Ross Callon - As AD it seems to me that this draft has interest in both MPLS and L3VPN [04:27:57] becoming a WG draft gives it some credibility [04:28:31] I think that in this case the question should be put to both WG even though it can only be a draft for one WG [04:28:53] Yiqun Cai : Is this being to L3VPN or MPLS ? [04:29:30] Ross ; My understanding is that it has already been put to MPLS [04:29:44] Loa Anderson : We are comfortable to do this in the MPLS WG [04:30:10] Marshall : What would the author's prefer ? [04:30:37] Ijsbrand : We would prefer the MPLS WG [04:31:05] Loa : THe chairs of the two WG would have to agree to this, biut that shouldn't be difficult [04:31:05] ah@jabber.org joins the room [04:31:24] Next : Thomas Morin : Manditory Features in a multicast VPN [04:32:56] Thomas : This draft was updated one year ago [04:33:03] there was a WG last call last June [04:33:18] revisions and 5 followed to address comments [04:33:32] One section was removed - 3.3.6 [04:34:01] many improvements in Appendix A [04:34:14] improvements for SSM were extended to ASM [04:34:22] hide.zebra joins the room [04:34:48] avoided adding variables for SSM/BGP, using a RR meshing as a solution [04:34:57] naby textual improvements [04:35:08] many textual improvements [04:35:37] added a short section on why PE-CE multicast routing is not considered when comparing PE-PE [04:35:43] C-multicast routing [04:35:57] many clarifcations and style updates [04:36:35] One pending item [04:36:52] comparison of C-multicast routing optins for inter-AS [04:37:18] authors think that WG LC comments were addressed in the current draft [04:38:23] Danny Mcpherson : I have been going back and forth on Thomas and Eric and what remains is the inter-AS stuff and present that to the WG for LC [04:38:50] Marshall : Is a goal of before the next IETF reasonable ? [04:38:55] Danny : Before [04:39:41] Ijsbrand : I don't think that a comparson between BGP and PIM in the core should be in the draft [04:39:57] Thomas : you mentioned protocol improvements [04:40:12] further improvements are completely out of scope [04:40:38] If you think that there are personal opinions in the draft, please bring them up [04:41:09] Yiqun : There was a big debate on whether PC PIM overhead should be considered or not [04:41:26] we have to evaulate the full piece [04:41:43] Thomas : This has been discussed on the list many times [04:42:45] If there are multiple options for PC in PIM, you could convert them [04:43:22] As I said, this explaination is in the draft itself, and feel free to comment on it [04:43:34] Next : Multicast VPN fast fail-over [04:43:42] Thomas again [04:43:47] presented at the last IETF [04:43:52] 2 mechanisms [04:44:28] UMH selection based on the status of P-tunnels - avoid waiting for unicast convergence [04:44:45] - second, standby C-multicast route - [04:44:59] these mechanisms can be used independently or together [04:45:18] Standby support in Inter AS [04:45:38] the Last revision adds Standby C-multicast routes in inter-AS [04:46:23] PE's not in the AS of the sort build a normal and standby routes [04:47:20] in Inter-AS the standby router has a local preference of zero - the normal route always takes precedance [04:47:38] sorry, the above is for INTRA-AS [04:47:58] in INTER AS, that has to be mapped into a value of infinity [04:48:39] we would like for WG adoption as soon as the WG starts accepting new documents [04:50:24] sprry, I have to talk some now [04:51:10] Marshall: Ross, do you have an AD perspective on the l3vpn charter? [04:51:22] thanks ! [04:51:52] Ross : in terms of this WG going forward, I would like to see the 2 multicast drafts to go forward [04:52:09] Ross: In terms of the WG moving forward, the two mcast drafts that have been out forward, is there more work on l3vpn's to be done. Then maybe yes, I am guessing its additional mcast work. [04:53:00] Marshall: Yes, the question is there more work. [04:53:45] Ross: We need to figure out what we want to do, there is some vagueness with the charter, we need to discuss what operators want to do with l3vpn. [04:53:54] Marshall: Yes, lets open it up to the floor. [04:54:02] thanks ! [04:55:09] Yiquin : There is a draft from Eric Rosen and others, including myself, which we believe forms a reasonable basis for moving forward [04:55:25] ah@jabber.org leaves the room [04:55:50] we also want feed back on whether to accept this draft as a one document or many document [04:55:53] we are open [04:56:14] to spliiting it up, or keeping it as one document [04:56:28] there are 6 pieces [04:56:53] finish specification on bi-dir PIM [04:56:53] foobar leaves the room: Replaced by new connection [04:56:54] foobar joins the room [04:57:13] extranet support using PIM [04:57:20] foobar leaves the room [04:57:33] bi-dir [04:57:38] extra-net [04:57:46] enhancements of existing procedures [04:58:06] For example, transport IPv6 is not clearly specified today. [04:59:09] An additional item was: Offering MVPN service with PIM control plane in concert with unicast "Hub and Spoke" VPN service and anycast-source service. The exact date of the email was: 24 Aug 09 from Eric. [04:59:28] Ron Bonica : You might want to poll the operator community [04:59:37] gregdavies joins the room [04:59:37] starting with the operators in this room [04:59:51] Andy Malis : other than multicast we are pretty happy [05:00:07] Thomas Morin : speaking both as an operator and a contributor [05:00:11] there are some items [05:00:14] extranet [05:00:38] the other item is the multicast fast failover document [05:01:23] ah@jabber.org joins the room [05:01:26] it is also very important to have a clear view of how the current documents will progress through the IESG before we continue [05:01:31] there are some areas [05:01:34] extranet is one [05:02:07] Ben Niven-Jenkins : [05:02:11] extranet is one of them [05:02:39] Ron Bonica : We know we have a lot of operational experience with unicast [05:02:58] should we wait for some operational experience [05:03:02] with multicast [05:03:18] Yiquin [05:03:33] Yiquin : We had an operator experience draft [05:03:39] lix joins the room [05:03:42] 6 years of experience isn't a short time [05:06:23] Marshall: Who has experience and is willing to discuss new deployments and use of the technology. [05:06:57] Ben : does the group feel that there is enough experience with multicast that the problems with any solutions are clear ? [05:07:27] Marshall : Who is deploying BGP-multicast L3 VPN [05:07:37] is the question [05:07:53] Ron : THere are 3 operators in the room [05:08:00] not a large number [05:08:22] Here is a solution - don't shut the WG down [05:09:10] lix leaves the room [05:09:14] but go quiet for a year or so to see how things work out [05:09:26] Yiquin : I suggest we go forward with the new items [05:09:49] I don't see any problem with going forward with [05:09:55] new solutions [05:10:13] Thomas Morin : I think that closing the WG would not be a good idea [05:10:38] I hope that clearing the plate would happen a lot faster than 6 months [05:10:51] lix joins the room [05:11:06] lix leaves the room [05:12:02] Ron : THat sounds like a solution right now [05:13:28] Ross ; If the considerations draft is done within the next month or so, there is no reason we couldn't show up in Anaheim with additional work to take on [05:15:24] Marshall: Look at the charter, it’s fairly wide ranging (IP Multicast) [05:15:27] ah@jabber.org leaves the room: Replaced by new connection [05:15:33] Ross: Agree [05:16:10] Marshall: Would it be appropriate to start consider new drafts in January and not wait for Anaheim (IETF 77) [05:16:20] Ross: Yes [05:16:39] tetsuya.innami joins the room [05:17:43] Daniel King : Approximately 2 years ago we had an anonymous VPN survey - it sounds like something similar is probably useful now [05:18:28] make it anonymous [05:18:32] gregdavies leaves the room [05:18:59] Ross : If we have a suspicion that L3 VPN BGP multicast is being deployed [05:19:05] gregdavies joins the room [05:19:23] we can use an anonymizer [05:21:02] Ron : I like the idea of an anonymizer [05:21:37] The chairs could talk about it in Nanog or Apricot or similar meetings [05:21:48] Yiquin - I am willing to share experiences [05:22:05] Thomas Morin : I am a bit skeptical [05:22:23] Before, we were trying to get the numbers [05:22:47] here, I am not sure I have identified the driver - I don't see this as a requirement for taking on new work [05:23:13] Ben : We are not suggesting linking the two things [05:25:08] Thomas : When vendors or operators suggest new items. this is feedback [05:25:17] there are forums for getting feedback [05:25:44] satoru.matsushima leaves the room [05:25:46] Daniel King leaves the room [05:25:47] Yiquin - I agree [05:26:16] we are done ! [05:26:17] hide.zebra leaves the room [05:26:37] rhe leaves the room [05:32:45] marshall leaves the room [05:48:40] yinxing leaves the room: offline [05:50:48] gregdavies leaves the room [05:58:26] tetsuya.innami leaves the room [06:06:53] Glenn Parsons leaves the room [06:17:42] Glenn Parsons joins the room [06:28:33] clns joins the room [06:43:29] clns leaves the room: offline [07:01:41] clns joins the room [07:03:46] clns leaves the room: offline [07:11:33] scott.mansfield leaves the room [08:36:11] Glenn Parsons leaves the room [08:40:56] Glenn Parsons joins the room [08:41:08] Glenn Parsons leaves the room [09:00:54] clns joins the room [09:01:06] clns leaves the room: offline