ppvpn@conference.ietf.jabber.com - 2002/11/20
[06:34] %% amalis has arrived.
[07:18] %% mrose has arrived.
[07:28] %% paul.knight has arrived.
[07:35] %% paul.knight has left.
[07:35] %% paul.knight has arrived.
[07:37] <paul.knight> PPVPN GENERIC REQUIREMENTS - Ananth Nagarajan
[07:39] %% swb has arrived.
[07:40] <swb> Lots of activity eh?
[07:40] <paul.knight> Loa Andersson - Terminology draft
[07:40] <amalis> Everyone was trying to keep up with Marco's firehose.
[07:41] <swb> My intention is to subvert the physical meeting and have the real meeting here
[07:42] <paul.knight> Issue - this draft might end up as a bottleneck
[07:43] <amalis> Paul - do you mean the requrements or the terminology?
[07:43] %% paul.knight has left.
[07:43] %% paul.knight has arrived.
[07:44] <paul.knight> The terminology draft might be a bottleneck if it is needed as a normative reference.
[07:45] <paul.knight> Ron Bonica - CE-CE authentication
[07:46] <paul.knight> Will apply for UDP port number
[07:47] <paul.knight> Vach Kompella - does PPVPN do protocols?
[07:47] <amalis> It's a polite fiction that we're not doing protocols!
[07:48] <paul.knight> Marco: will need to coordinate...
[07:48] <paul.knight> CE autoconfiguration - Cheng-Yin Lee
[07:52] <paul.knight> push, pull, poll, trigger, deafening silence ?
[07:53] <paul.knight> IPsec Protected Virtual Links for PPVPNs - Mark Duffy
[07:55] <paul.knight> CE-CE, PE-CE, PE/VR-PE/VR, CE-PE/VRF, PE/VR-PE/VRF applicability
[07:56] <amalis> There's actually someone at the mike!
[07:57] <paul.knight> Eric Rosen - separate SAs per-VPN probably not needed
[07:57] <paul.knight> Mark: Right, will address in a few minutes
[08:00] <paul.knight> transport mode with other tunneling (IP-in-IP, GRE, etc.) looks like appropriate approach
[08:01] <paul.knight> Eric Rosen - how does this relate to [my] existing 2547-in-IPsec draft?
[08:01] <paul.knight> Mark: This is separate.
[08:03] <paul.knight> Q: Yu Shun - how does this differ from draft-Touch and draft-Knight?
[08:04] <paul.knight> Mark: This addresses multiple VPNs in one tunnel, and identifying them.
[08:06] <paul.knight> Multiple Instances of OSPF for the PE/CE protocol in BGP/MPLS - 10 min - Kunihiro Ishiguro
Multiple Instances of OSPF for the PE/CE protocol in BGP/MPLS - 10 min - Kunihiro Ishiguro
Multiple Instances of OSPF for the PE/CE protocol in BGP/MPLS - Kunihiro Ishiguro
[08:07] <paul.knight> (Jabber panel wrapped on the paste operation, sorry!)
[08:10] <amalis> Maybe you need the Exodus client? :-)
[08:11] <swb> I'm not impressed with the Exodus client, but I haven't tried any others. I'm spoiled by mooing in emacs.
[08:12] <paul.knight> Actually, Exodus is what I have... not sure exactly the issue... it hangs sometimes
[08:13] <paul.knight> L2 DT report - L. Andersson
[08:16] <paul.knight> L2 requirements [Augustyn] - Yetik Serbest
[08:23] <paul.knight> VPLS Architectural Model - Eric Rosen
[08:26] <paul.knight> separation of IEEE bridge, VPLS forwarder function
[08:27] <paul.knight> In implementation, they share a MAC table
[08:29] <paul.knight> Issues with spanning tree... how to handle changes if not participating in Spanning Tree.. snoop?
[08:31] <paul.knight> M. Suzuki - Issues of VPLS
[08:33] %% bensons has arrived.
[08:36] <paul.knight> MAC routing !!!?
[08:38] <paul.knight> cooperation with IEEE 802 - N. Finn (? missed intro)
[08:38] <paul.knight> Hi Benson!
[08:39] %% jaltman has arrived.
[08:39] <swb> Norm Finn
[08:39] <bensons> Hi, Paul.
[08:40] <bensons> I'm unable to attend in person, and in absense of mcast this seemed like the next best thing.
[08:40] <bensons> thanks for the text updates
[08:41] <paul.knight> Glad to know it is useful for remote participation.. I'm trying to promote it!
[08:41] <swb> I still want to push http://cvw.sourceforge.net
[08:42] <swb> A real sharing tool, for permanent (not just during meetings) objects, for use by the WGs
[08:42] <bensons> I don't have experience with CVW, but I agree completely that something of this nature would be useful ongoing.
[08:43] <mrose> swb: persistent rooms isn't a problem, the problem is when you start raising the bar for multimedia, etc.
[08:43] <paul.knight> Maybe evolution into VAXnotes ?? :^)
[08:43] <swb> (successfully sucked Marshall in ... first objective achieved)
[08:43] <swb> marshall: a good tool wouldn't require multimedia, just allow it in the future.
[08:44] <paul.knight> Vach Kompella - good to have IEEE support for Q-in-Q (Q-tagging stacked)
[08:45] <amalis> It would help a great deal if all of the presentations were available online prior to the meetings, for both local and remote participants.
[08:45] <swb> :nods
[08:45] <bensons> Being a remote participant, I agree fully. ;)
[08:45] <swb> In my case I was modifying mine based on what previous speakers had to say ... but we do have wireless, so we could do that just-in-time
[08:46] <paul.knight> Yes, even last-minute presos can be uploaded
[08:46] <mrose> first, is ppvpn still meeting? if so, i don't want to step on this room with this discussion. if they're done meeting fine...
[08:46] <paul.knight> CE-based VPLS - Cheng-Yin Lee
[08:46] <swb> still meeting
[08:47] <paul.knight> [Yes, still meeting, but not much meeting-based traffic except my notes.]
[08:47] <swb> Marshall, we'll have fun later
[08:48] <mrose> swb: okie-dokie. there's always the plenary chatroom
[08:52] <paul.knight> LDP-based Signaling for L2VPNs - Eric Rosen
[08:52] <paul.knight> Signaling: Identifying PW Endpoints
[08:53] <bensons> is pwe3 update?
[08:54] <paul.knight> It would definitely be good for anyone remote to have the presentations already online
[08:55] <paul.knight> Not quite sure what you're asking on PWE3...
[08:55] <bensons> :agree
[08:55] <swb> bensons: he says PW will own the signaling but MPLS will have to use it, so MPLS should pay attention
[08:56] <bensons> ok
[08:56] <paul.knight> problems: VCid too short, especially interprovider. Also, VPN-IDs not quite right
[08:57] <bensons> PPVPN->l2vpn = PWE3 + MPLS ...
[08:57] <paul.knight> Identifier should include VPN-ID, local CEnumber, remote CE number
[08:57] <bensons> VPN-IDs per RFC vs. new GID format should be discussed someday
[08:57] <bensons> soon
[09:00] <paul.knight> Proposal: generalized identifier field
[09:00] <paul.knight> Comment: this is needed
[09:01] <paul.knight> Marco: we need to figure out what is needed here in PPVPN.
[09:02] <paul.knight> Yakov: Q: VPN-ID - can it use RT extended community already defined?
[09:02] <paul.knight> Eric: not restricted to 8 octets..
[09:03] <paul.knight> Hamid OUld-Brahim: VPN-ID can still be supported?
[09:03] <paul.knight> Eric: probably...
[09:04] <paul.knight> Virtual Hierarchical LAN Services - Arnold Sodder
[09:05] <paul.knight> mac-in-mac frame encapsulation...
[09:05] <paul.knight> VPN-Identifier and Control Word
[09:06] <paul.knight> <YAVPNID - yet another....VPN-ID>
[09:06] <amalis> YAVPNID :-) :-)
[09:07] <paul.knight> Q: what is motivation for this draft?
[09:08] <paul.knight> A: Some people (Service Providers) want it...
[09:09] <paul.knight> GVPLS/LPE - Generic VPLS Solution based on LPE Framework - Dinesh Mohan
[09:12] <paul.knight> amalis as co-author....
[09:13] <paul.knight> GVPLS - interoperability with other VPLS, optimizing replication, multicast
[09:15] <amalis> It attempts to generalize lasserre-vkompella.
[09:16] <paul.knight> supports multipoint-to-point PseudoWires (wants them for PE-PE)
[09:16] <amalis> You also want mp-to-p for scaling in the core.
[09:19] <paul.knight> Looks like good approach, convergence.
[09:20] <paul.knight> VPLS based on IP Multicast - Ali Sajassi
[09:22] <amalis> I think so too (your remark on the GVPLS draft)
[09:27] <paul.knight> Q: PPVPN req drafts for scalability says network needs to suppor up to 10,000 VPNs
[09:28] <paul.knight> This implies that up to 1 million unique IP addresses would be needed for this approach
[09:28] <paul.knight> Also the routes for these addresses would need to be learned
[09:29] <paul.knight> Answer:
[09:29] <paul.knight> Use private IP address; use in hierarchical way, divide into different spaces
[09:31] <paul.knight> Answer 2: (brief, on Marco's urging) detailed discussion on list ( no fireworks right now.. have to settle for coffee)
[09:32] <paul.knight> Eric Rosen: packet with bad source address could pull traffic around, MAC address learning makes this a fragile approach.
[09:33] %% paul.knight has left.
[09:33] %% paul.knight has arrived.
[09:34] <paul.knight> Juha H: I need inter-AS support (7 AS in my net) This will have trouble for my net.
[09:34] <paul.knight> L2VPN Interworking - Ali Sajassi (next presentation)
[09:35] <paul.knight> [Exodus apparently bounced me out, had to reconnect]
[09:37] <paul.knight> Interworking between disparate native services or disparate Attachment Circuits? Focus is on the native services.
[09:38] <paul.knight> conclusion - need PWs for IP and Multi-protocol
[09:39] <paul.knight> Scott Bradner: when sub-IP ws set up, Fred Baker specifically ruled interworking out-of-scope
[09:40] %% mrose has left.
[09:41] %% paul.knight has left.
[09:41] %% paul.knight has arrived.
[09:42] <paul.knight> Answer: this is not defining interworking, just listing the existing interworking method to use in various cases.
[09:42] <paul.knight> Scott: This is okay.
[09:42] %% swb has left.
[09:43] <paul.knight> Marco: Next work steps on L2 (slide)
[09:44] <paul.knight> functional decompositon; limited set of solutions, no protocol work
[09:45] <paul.knight> discovery, signaling, PW tunneling (encaps), Informational model
[09:46] <paul.knight> structure of functional documents: requirements, options specification, interactons with other functions
[09:47] <paul.knight> N docs per functional block, or one doc per functinal block (n Option sections) ?
[09:47] <paul.knight> See the preso for the real details, I'm missing much...
[09:48] <paul.knight> L2 Design Team work plan - meeting in January - TBD
[09:49] %% jaltman has left.
[09:51] <paul.knight> Meeting closed
[09:51] %% paul.knight has left.
[10:01] %% bensons has left.
[10:09] %% amalis has left.
[11:12] %% amalis has arrived.
[11:12] %% amalis has left.