IETF
pwe3
pwe3@jabber.ietf.org
Tuesday, March 12, 2013< ^ >
Room Configuration
Room Occupants

GMT+0
[12:44:05] Andy Malis joins the room
[12:44:27] <Andy Malis> Testing 1 2 3
[12:58:48] david.sinicrope joins the room
[12:59:09] <Andy Malis> Hi Dave
[12:59:27] <david.sinicrope> hi, I'm using Messages for this.... so far not as good as Adium
[12:59:34] Y(J)S joins the room
[13:00:47] <Y(J)S> bashing agenda
[13:01:03] <Y(J)S> last minute addition - Curtis will present the MPLS forwarding draft
[13:01:13] <Y(J)S> milestones (out of date)
[13:01:27] <Y(J)S> 1 new RFC - 6870 (PW Redundancy Bit), none in RFC Ed Queue
[13:01:53] <Y(J)S> Ethernet OAM Interworking approved, but has outstanding comments that need to be addressed
[13:02:03] <Y(J)S> VCCV Implementation Survey Results dead due to lack of action
[13:02:11] <Y(J)S> Andy volunteered to continue
[13:02:19] <Y(J)S> (new version soon)
[13:02:32] <Y(J)S> P2MP PW Requirements sent back to WG, expired, there are volunteers to revive
[13:02:44] <Y(J)S> ICCP - not on agenda, ready for LC?
[13:03:17] <Y(J)S> only 2 people have read a recent version
[13:03:29] <Y(J)S> Status reduction - refreshed
[13:04:33] <Y(J)S> congcons - Y(J)S at mike - need respin and then LC
[13:04:41] <Y(J)S> OAM config - expired
[13:04:47] <Y(J)S> MPLS-TP MS-PW - ongoing
[13:05:25] <Y(J)S> VCCV for GAL - expired, Tom at mike - didn't get to around to updating
[13:05:35] <Y(J)S> 4447bis - on agenda
[13:05:47] <Y(J)S> MPLS-TP over bidirectional LSP - ongoing, recently became WG draft
[13:05:59] <Y(J)S> SPE redundancy  ongoing
[13:06:10] <Y(J)S> Dynamic MS-PW and MS-PW ER awaiting writeup (by Giles since chairs are authors)
[13:07:20] <Y(J)S> Y(J)S at mike - VCCV2 : no progress
[13:08:09] <Y(J)S> Load at mike - mpls-tp draft moved to PWE3 (sorry I missed the exact draft name)
[13:08:20] <Y(J)S> Pseudowire Setup and Maintenance using the Label Distribution Protocol
[13:08:24] <Y(J)S> Giles w/o slides
[13:08:53] <Y(J)S> background : 4447 has errata and want to get it right this time
[13:09:22] <Y(J)S> Luca did a 01 before the meeting
[13:10:21] <Y(J)S> problems - PW label no longer at BoS, nodes that don't support status messaging, wildcard withdrawal
[13:10:35] <Y(J)S> Luca asks for people to read and comment
[13:10:50] <Y(J)S> Stewart - need security section fixes (only has MD5)
[13:10:59] <Y(J)S> should point to other LDP documents
[13:14:04] <Y(J)S> discussion at Mike with Stewart, Loa and Y(J)S on pointer to MPLS security draft
[13:14:25] <Y(J)S> PW Endpoint Fast Failure Protection - Yimin (with Rahul and Wim)
[13:14:48] <Y(J)S> draft-shen-pwe3-endpoint-fast-protection-03
[13:15:16] <Y(J)S> single major update - IGP advertisement has high-level guidelines
[13:15:55] <Y(J)S> reason - generic to allow present and future approaches
[13:16:06] <Y(J)S> important to have FRR type mechanism for egress AC and PE, and S-PE
[13:16:52] <Y(J)S> e2e restoration within 10s of ms
[13:17:21] <Y(J)S> discussing highlights
[13:19:39] <david.sinicrope> who else is in the room besides Andy, Yaakov and I
[13:23:13] <Y(J)S> someone from Huawei at mike :  how do you coordinate different layers of protection ?
[13:24:01] <Y(J)S> Yimin: the draft gives guidelines, transport tunnel could be RSVP/LDP/IP-tunnel,
[13:24:48] <Y(J)S> draft gives details on how compute backup path
[13:26:23] <Y(J)S> David - there are only the 3 of us !
[13:26:55] <david.sinicrope> I think the comment is a case of "it hurts when you do this" and the answer is "don't do that"
[13:27:17] <Y(J)S> Lucy clarifying the question - is there coordination between protection levels? can operatror choose ?
[13:27:22] <Y(J)S> answer - yes
[13:27:56] <Y(J)S> Curtis at mike :  transport tunnel LSRs may not know about PWs
[13:28:28] <Y(J)S> Yimin: here LSR DO know, since context ID address is known by LSR
[13:28:50] <Y(J)S> Curtis - but what about backwards compatibility, ie. LSRs that don't support this mechanism
[13:28:59] <Y(J)S> Yimin - explained in draft
[13:30:23] <Y(J)S> Andy - when I read the draft I was looking for a section on backwards compatibility - can you include in the next version ?
[13:32:33] <Y(J)S> Matthew : who has read ?  7 people
[13:32:42] <Y(J)S> Please read and send comments to list
[13:33:00] <Y(J)S> MS-PW bandwidth availability - Min
[13:33:46] <Y(J)S> draft-long-pwe3-ms-pw-availability-01
[13:34:33] <Y(J)S> tunnel may go thru MS-PW segments with different BWs (microwave, DSL)
[13:35:23] <Y(J)S> with bad weather BW decreases and low priority service will be dropped
[13:35:44] <Y(J)S> problems with MS-PW scenarios
[13:36:16] <Y(J)S> S-PE won't know if tunnel can satisfy BW req
[13:36:35] <Y(J)S> solution - carry availability req in LDP message
[13:37:28] <Y(J)S> when S-PE can not satisfy sends status message "unavailable"
[13:37:43] <Y(J)S> T-PE withdraws in the opposite direction
[13:40:45] <Y(J)S> Y(J)S at mike - how do you measure availability?
[13:41:02] <Y(J)S> Mn - based on long term past average
[13:41:17] <Y(J)S> Y(J)S then should be in management, not in LDP
[13:41:36] <Y(J)S> Greg - LDP really the wrong place, should be in RSVP-TE
[13:41:42] <Y(J)S> Himanshu - should be in CCAMP
[13:42:04] <Y(J)S> Min - have submitted to CCAMP
[13:42:29] <Y(J)S> TRILL Over Pseudo Wires - Lucy (with Donald Eastlake, Sam Aldrin, Jon Hudson)
[13:42:58] <Y(J)S> draft-yong-pwe3-trill-o-pw-00
[13:43:04] <Y(J)S> TRILL WG in IETF (RFC 6325)
[13:43:06] <Y(J)S> TRILL switches need to connect over MPLS
[13:43:32] <Y(J)S> PW can be used to connect TRILL switches at different sites very far apart
[13:44:07] <Y(J)S> TRILL can run over arbitrary link protocol, so don't need new PW type - can use PPP PW (4618) or Ethernet PW (4448)
[13:44:37] <Y(J)S> no need to change TRILL or PW protocols
[13:44:59] <Y(J)S> draft describes procedures at TRILL switch, autoconfig, security
[13:47:50] <Y(J)S> Y(J)S at mike - is this informational?
[13:47:59] <Y(J)S> Lucy - no standards track
[13:48:11] <Y(J)S> Y(J)S - isn't this an NSP?
[13:48:19] <Y(J)S> Matthew and Stewart agree
[13:48:33] <Y(J)S> in that case not for PWE3, so perhaps should be in TRILL
[13:48:43] <Y(J)S> Donald - would need a TRILL charter update
[13:48:51] <Y(J)S> but one is in process anyway
[13:50:01] <Y(J)S> Lucy - so will bring to TRILL
[13:50:24] <Y(J)S> Rajiv - do endpoints need standardization of anything?
[13:50:34] <Y(J)S> Lucy - yes there are procedures and encaps
[13:52:25] <Y(J)S> Curtis redoing forwarding draft presented yesterday at MPLS WG
[13:52:30] <david.sinicrope> YJS:  Is this an informational draft?
Lucy: Standards track, autoconfig and procedures are included
YJS: but procedures are for trill gateway so why not in TRILL
Lucy: TRILL switchs starts and ends PW
Matthew: is TRILL the native service or PW.  We don't typically have a native node starting the PW.  Is TRILL the service?
Lucy: TRILL is using the PWs.
YJS: not a new PW type, using PPP or Eth PW.  It's an NSP, because it is showing how to put TRILL on the PW.
Stewart: NSP is outside the scope of the PW.  This can be done in the TRILL WG.  No changes to PW needed so not PWE3 issue
Donald Eastlake: Not in the TRILL charter.
Stewart: so need a TRILL charter update
Donald: TRILL charter update in progress so can pursue this there and is a good time to do this.
Stewart: in the same way that L2VPN uses PWs, then if TRILL wnats to use PWs as an application, then this work is outside the scope of this WG
Rajiv Assari?:  can't find anything about procedures that needs to be standardized between two endpoints.
Lucy:  specify process and procedure at each endpoint. Also specify two encapsulation schemes so trying to standardize.
YJS: why PPP, and not just Ethernet?
Lucy: in cases you have PPP and no Ethernet, so we have PPP
YJS: but you are adding function to the TRILL switch so why not add Ethernet and reduce options?
Donald: 6361 TRILL over PPP, and Ethernet eats 16bytes per frame
[13:53:37] <Y(J)S> comments from the gallery - yes, and BW is so expensive nowadays :)
[13:54:23] <Y(J)S> want to cover basic PW forwarding in this draft
[13:54:43] <david.sinicrope> especially between data centers
[13:55:46] <david.sinicrope> @Yaakov: you mean I have to replace my FR link with Ethernet?
[13:58:59] <Y(J)S> why would I do that ?  FR can have unlimted frame size, while Ethernet is limited to jumbo frames
[14:01:40] <Y(J)S> YJS at mike - do you get into SN processing?
[14:01:52] <Y(J)S> Curtis - no, just general wording on whether drop
[14:02:15] <Y(J)S> YJS - Luca and I wrote generic text on SN processing, offers to sell it
[14:02:29] <Y(J)S> Stewart - concerned about expansion of scope of draft
[14:02:46] <Y(J)S> Curtis - all we say is CW may be a MUST for some providers
[14:03:27] <Y(J)S> points to PWE documents
[14:04:55] <Y(J)S> Matthew - anything else ?
[14:04:59] <Y(J)S> done !!!!!
[14:08:05] Andy Malis leaves the room
[14:13:12] Y(J)S leaves the room
[14:21:30] david.sinicrope leaves the room
[14:41:08] You-GT joins the room
[15:01:53] Michael Barnes joins the room
[15:41:44] Michael Barnes leaves the room
[16:31:36] You-GT leaves the room
[17:24:08] david.sinicrope joins the room
[17:24:14] david.sinicrope leaves the room