[12:00:04] --- Rik Wade has joined [12:11:29] --- Ilya Varlashkin has joined [12:12:17] --- Chris L has joined [12:17:24] I wonder if videolan channel for this session is also swapped to ch5 instead of annoucned ch4 (was this morning) [12:20:15] --- Chris L has left: Replaced by new connection [12:25:00] --- cdl has joined [12:27:21] --- yjs has joined [12:27:38] Danny going over status [12:28:08] work is in progress on new charter [12:30:18] new charter will be publicized in a week or two [12:31:00] Praveen Muley Praveen Muley coming up to speak on PW Redundancy [12:31:27] both docs are WG docs [12:31:36] it's quite hard to hear the mic [12:31:44] how is it now ? [12:31:53] better :-) [12:32:19] --- twalsh has joined [12:32:33] framework and reqs draft - new authors, reqs and terminology [12:33:40] --- howard.green has joined [12:33:59] new terminology - primary and secondary PWs, revertive, etc. [12:34:47] reqs - N:1 is optional (1:1 mandatory, 1+1 not addressed) [12:35:17] revertive behavior is optional [12:35:56] now discussing the status bit def draft [12:36:25] new appendix on redundancy procedures [12:37:04] manual lockout case needs to be covered [12:38:23] comments from floor - need to harmonize terminology with ITU (SG15/Q9) [12:39:15] Dinesh can't present as his flight was cancelled [12:40:01] Fred (FT) on AII reachability [12:41:05] (I just realized that I am co-author of this draft ...) [12:41:40] the idea is to populate AII routing tables by LDP [12:42:05] I think that jabber dropped again [12:42:10] Jabber is fine here. [12:42:13] did it? [12:42:17] back up ... [12:43:19] in access network can not use standard autodiscovery mechanisms since no BGP or IGP [12:45:17] idea - new LDP addressing family [12:46:01] feedback ? do we need to extend to AGI too ? [12:46:19] LDP is rather simple proto, adding new functionality is adding complexity [12:46:29] wouldn't be easier to deploy IGP and don't touch LDP? [12:47:02] well no. The operator is not going to deploy a routing protocol just for autodiscoery [12:47:30] Luca - is there an application for AGI autodiscovery ? [12:48:43] Hamid - what is meant by interworking with other autodiscovery mechanisms ? [12:49:06] Fred - just relay info acquired by routing protocol [12:49:21] jabber down [12:49:24] will BGP/IS-IS/OSPF need to be aware of the new AF? [12:50:18] Nabil is going to give Dinesh's presentation [12:51:19] Ethernet OAM interworking (native Eth OAM and PW OAM) [12:52:07] map defect states between 1ag and PW OAM (extends OAM message mapping draft) [12:54:35] describing the defects of AC and PSN in both forward and reverse directions [12:55:52] talking about options (too many?) for detecting and entering defect state [12:56:31] should this be a separate doc, or merged with OAM msg mapping ? [12:59:06] Yaakov and Ali - need to cover 3ah OAM as well [13:00:10] Luca - I have asked years ago for this to be in OAM msg mapping ... (chair bashing) [13:02:53] Danny - need to keep terminology consistent [13:06:05] discussing if OAM msg mapping is ready for WG LC (this will be an informative reference) [13:06:45] George - the Ethernet work is the most important [13:07:14] vach on hash label [13:08:10] the question is how to exploit multiple PSN paths for PW load balancing [13:08:23] similar to fat PW draft [13:09:47] a number of reasons to exploit multiple paths (insufficient BW in available pipes) [13:10:01] at PW ingress there are L2 fields that can be used [13:10:46] but LSRs don't know (not even if IP or PW) other than "expensive" hacks (e.g. 1st nibble) [13:11:21] but we always have the label stack - how can we use ? [13:14:22] solution: 2 LDP session endpoints exchange hash labels (popped at endpoint) that differentiate in the middle of nework [13:17:19] --- yjs has left: Replaced by new connection. [13:17:35] --- yjs has joined [13:17:52] network fell - I am back now [13:18:00] extensible to PSNs [13:18:50] for each intermediate hop endpoints exchange hash labels [13:25:45] --- twalsh has left [13:27:43] questions at mike ... (too many to list, but general feeling is positive) [13:30:40] Kireeti suggests a DT for where the hash labels should go [13:30:52] Don - is this only for BE ? [13:31:40] --- Will Ivancic has joined [13:32:39] ali and sasha ask about impact of this on OAM [13:34:14] need requirements work [13:34:33] if OAM doesn't follow same path then need fix [13:35:00] George - most of industry only hashes on bottom-most label [13:35:18] also if general mechanism then should be in MPLS WG [13:42:43] Mark and Dave Allen - IP already has load balancing so pressing need is only for PWs [13:43:53] --- howard.green has left [13:44:26] Stewart rehashing the same lecture as this AM at MPLS [13:45:43] --- amalis has joined [13:46:37] history - IAB DT came up with 61 comments on ITU on OAM for T-MPLS docs [13:47:25] went to SG13 meeting in Seoul - considerable work on reqs, which became a non-normative "supplement" [13:48:00] this supplement has lots of useful info [13:48:28] didn't get to protocol spec at the SG13 meeting, so was it withdrawn from approval process [13:49:32] next stop - SG15. a joint working group (JWT) was formed with Dave Ward and Malcolm Betts as chairs [13:49:58] two options - option 1 (IETF standards process ) and option 2 (separate T-MPLS from MPLS) [13:50:19] JWT needs to recommend to ITU management on which option to choose by the end of April [13:51:21] IAB DT has completed its work, but there are 2 other teams now [13:59:54] --- kenichi has joined [14:00:16] Malcolm Betts is talking about ITU viewpoint - stresses present teams will not change documents - that needs to be done using regular processes [14:00:45] --- kenichi has left [14:00:50] next I am up - sorry can't jabber it [14:03:00] --- DD has joined [14:03:42] --- DD has left [14:20:40] --- Will Ivancic has left: Computer went to sleep [14:22:48] --- Rik Wade has left [14:23:03] --- amalis has left [14:23:22] --- yjs has left [15:35:58] --- amalis has joined [15:36:10] --- amalis has left [16:15:37] --- Stewart Bryant has joined [16:16:11] --- Stewart Bryant has left