[17:44:02] --- venaas has joined
[18:09:30] --- frodek has joined
[19:24:13] --- frodek has left: Disconnected
[19:39:40] --- abierman has joined
[19:41:19] --- Roy Brabson has joined
[19:42:12] --- sleinen has joined
[19:42:41] <abierman> Hi Simon
[19:44:04] --- ray_atarashi has joined
[19:46:21] --- loughney has joined
[19:46:49] <loughney> aron falk asked about bashing agenda, but not bashed due to presenter's needs
[19:47:01] --- kivinen has joined
[19:47:19] <loughney> David talked about proto team work - ops area is using the wg chair sheparding stuff.
[19:47:32] <loughney> DNSext WG chair selection change
[19:47:47] <loughney> Asked for candidates & interviewed candidates
[19:47:57] --- raeburn has joined
[19:47:58] <loughney> Used this to select the new chair, had input from the WG.
[19:48:19] <loughney> Will do this for new WG chairs or new WGs
[19:48:47] <loughney> Agenda is:
[19:48:48] <loughney> Draft Agenda (v1) for the Operations & Management Open Area Meeting IETF64 When: 1740-1950, MONDAY, November 7, 2005 Where: Salon 2/3 A. Administrative stuff - appointment of scribe - blue sheets - agenda bashing (David Kessens, Bert Wijnen) B. Proposed working groups - AAA Maintenance (John Loughney, speaker not yet confirmed) - Diffserv Control Plane Elements (DCPEl) (Kathleen Nichols, Scott Bradner) mailing list: dcpel@ietf.org archive: https://www1.ietf.org/mailman/listinfo/dcpel - MAVS BOF proposal (Martin Halstead) C. SNMP/ISMS issues: - Overlap between SNMP+ISMS/NETCONF (Simon Leinen & Andy Bierman - speakers not yet confirmed) Some people see potential in evolving NETCONF - which currently focuses on configuration management - into a more general network management protocol, which would subsume much of what SNMP was designed for. At the same time, the ISMS (Integrated Security Mechanisms for SNMP) WG is converging towards the same connection-oriented underlying transport (SSH) as NETCONF. Now seems to be a good time to think about the overlap between SNMP+ISMS and NETCONF, and whether an effort should be made in the IETF at unifying the two. ... Y. Open Mike Z. AOB ---
[19:49:09] <loughney> Bert has decided to step down as OPS AD
[19:49:15] <loughney> Bert encourages new blood.
[19:49:39] <loughney> applause for Bert.
[19:49:50] <loughney> SNMP/ISMS discussion
[19:51:33] --- admcd has joined
[19:54:49] --- venaas has left: Logged out
[19:54:53] --- loughney has left: Replaced by new connection
[19:55:43] --- loughney has joined
[19:56:04] <loughney> Overview of Operations protocols - slides do better justice than what I can do here.
[19:56:57] --- frodek has joined
[19:57:10] <loughney> AAA - RADIUS & TACACS+ most common protocols
[19:57:49] --- bert has joined
[19:59:00] <loughney> Data modeling slide giving a good overview.
[19:59:15] <loughney> Comment from the audience, but I missed it - anyone else catch it?
[19:59:51] <loughney> Data modeling languages SMING vs. XML.
[20:00:07] <loughney> Possible Convergence work
[20:00:33] <bert> comment was that syslog cannot be transported over NetCOnf at the moment, or at least it is not specified how
[20:00:41] <loughney> Multiple approaches to discuss.
[20:00:57] <loughney> should we use same secure transport?
[20:01:10] <loughney> develop common NM auth?
[20:01:34] <loughney> have netconf actually do snmp & replace snmp ultimately ...
[20:01:53] <loughney> develope extend ops for accessing SMP data to supplement snmp.
[20:02:05] <loughney> Any questions or comments?
[20:02:31] <loughney> Chris Lonqvick ? syslog chair
[20:02:41] <loughney> says they have syslog over BEEP.
[20:02:56] <loughney> Syslog wg meeting next session.
[20:03:19] <loughney> Bert - is syslog making a mandatory protocol?
[20:03:45] <loughney> chris - did make a madatory to implement for transport
[20:04:10] <loughney> Chris - udp is the only mandatory to implement, as it is already supported.
[20:05:33] <loughney> Eliot Lear - great chart!
[20:05:57] --- jschoenwae@jabber.org has joined
[20:05:58] <loughney> Eliot - interested in a common substrait. get a shared understanding and experience with it.
[20:06:32] <loughney> Eliot - get common authorization and authentication mechansism. also get to have multiple sessions over same transport.
[20:07:02] <loughney> Eliot - you get to have seperate functions over a common substrait.
[20:07:42] <loughney> Sharon Chism (?) - supports the idea of getting convergence - saves on draft space (ha!) and simplifies consumption.
[20:07:49] --- david has joined
[20:08:19] <loughney> Sharon - we got some things right in the past, some things not so good. You do get some improvements over time.
[20:09:03] <loughney> Wes Har(something or other) - we are no longer picking the transport because it is what we need, but we choose it because the security infra we have deployed.
[20:09:17] <loughney> Wes - eg - we
[20:09:52] <loughney> chose tcp because of the security, not because we wanted tcp
[20:10:36] <loughney> (someone) - trying to have a common provisioning mechanism for sip devices behind nats. and want to manage them behind home nats (or was it nets)?
[20:11:27] <loughney> Sharon - follow-up take SMNP mibs and xml-ize them - this worries her.
[20:12:37] <loughney> Russ Monday - as we look at the transport level, it might be good to develop something like an API that other pieces could reuse them.
[20:13:23] <loughney> David responds - as we put SNMP over SSH, SNMP does auth based upon the user, so when we do it at a lower layer, it makes things more difficult.
[20:13:40] <loughney> Jean-francois - Cable Labs (maybe) - agrees with Sharon
[20:13:45] --- Yoshifumi Atarashi has joined
[20:14:14] <loughney> David Black - in the enterprise, the network isn't the only thing to be managed.
[20:14:29] <loughney> David Black - this adds more things to manage.
[20:14:43] <loughney> Bert - mentioned that he / they are aware of this
[20:17:04] <loughney> Discussion of DMTF and some other stuff ...
[20:17:28] <abierman> NETCONF transport of SNMP PDUs is not a really hard problem. If it's so important, there should be lots of proprietary solutions to look at any potentially standardize. Where are these implementations?
[20:17:41] <loughney> Simon Leinen - he brought this up first.
[20:18:21] <loughney> Simon - maybe this is irtf type work?
[20:19:31] <loughney> Bert has a question - on the slide, these are the WGs they thought of.
[20:20:03] <loughney> Bert - if there are other Working Groups that are affected by this, please let the OPS ads know.
[20:20:25] <loughney> There ipfix & patch-xml bof that both might be related.
[20:20:53] <loughney> J-F - in sipping, it was decided to add some security model about something or other ...
[20:21:08] --- bert has left
[20:21:08] <loughney> J-F - its related, he thinks ...
[20:21:20] <loughney> Bert - can you check with that WG and let us know.
[20:22:15] --- venaas has joined
[20:23:28] <loughney> Somene - talking about Policy model ...
[20:24:18] <loughney> Eliot - responding to Bert - it is possible to use netconf to transport the data representing a schema for sim, so it shouldn't be too difficult.
[20:24:55] <loughney> Eliot - about ipfix -.they use SCTP, and it relates to Wes's concern about selection of transport protocol.
[20:25:17] <loughney> Eliot - we might run into the installed base issue ...
[20:25:49] --- jschoenwae@jabber.org has left: Replaced by new connection
[20:27:23] <loughney> Next topic is Provisioning in Internet-wide VPN Services
[20:27:52] <loughney> Christian Jacquenet - discussion of providing internet wide vpn services
[20:28:08] <loughney> 3 topics: Conext & modivation; issues & requirements; next steps
[20:28:42] <loughney> Emerging 'triple-play' services, some that require traffic isolation
[20:28:57] <loughney> Wanting to be able to automate provisioning
[20:30:37] <loughney> [specifics are on the slides, so I am not repeating them all]
[20:34:01] --- bert has joined
[20:34:33] <loughney> discussed in mavs@ietf.org - multi-as vpn services mailing list.
[20:34:46] <loughney> Want to have a bof at the next ietf
[20:35:18] <loughney> David - this was proposed as a bof, but was rejected for admin reasons (no mailing list at the time).
[20:36:31] <loughney> Russ Callon (sp?) - a number of issues here. Are we interested in a multi-service proider VPN scenario? He thinks yes.
[20:36:56] <loughney> Russ - There's a number of topics here, but where do to the work.
[20:38:32] <loughney> Open issues about QoS - inter AS QoS.
[20:40:07] --- raeburn has left: Disconnected
[20:42:06] --- ray_atarashi has left: Replaced by new connection
[20:42:34] --- ray_atarashi has joined
[20:43:02] --- ray_atarashi has left
[20:43:36] --- ray_atarashi has joined
[20:47:22] <loughney> - was at the mic, so this discussion go lost -
[20:47:56] <loughney> somone - interprovider QoS is an open issue - monitoring SLAs is also an other area.
[20:49:18] <loughney> diffserv Control Plane Elements - Kathleen Kichols
[20:49:57] --- ray_atarashi has left: Disconnected
[20:51:39] <loughney> Do other people see a problem here?
[20:52:01] <loughney> What to solve a problem end-to-end, differentiated services.
[20:52:31] <loughney> Had some question on what comes first.
[20:52:54] <loughney> Maybe first see what you can do within a domain, then look at the inter domain later.
[20:53:41] <loughney> Scott - the idea is not new. After diffserv specs completed, Scott asked the chairs if they could start this work.
[20:53:53] <loughney> Some overlap with NSIS and other WGs.
[20:54:10] --- raeburn has joined
[20:55:14] --- admcd has left: Replaced by new connection
[20:55:24] --- ray_atarashi has joined
[20:55:27] <loughney> Scott - diffserv is only covering operations within an operator, not how you tell a router what to do.
[20:56:38] <loughney> Scott - not talking about signaling, but how you make the decision to do something.
[20:57:45] <loughney> David K - 2 more presentations on this, let's have the presentations first.
[20:58:14] <loughney> Kathy - there is a bar-bof at lunch tomorrow at the MacKenzie room
[20:58:24] --- admcd has joined
[20:58:58] <loughney> Bob Braden - no doubt that the IETF should work on this. Is always thinking that there is a 'then some magic happens' with diffserv.
[20:59:36] <loughney> Bob -.RSVP started out initially to be QoS only. NSIS was signaling is signaling.
[21:01:58] <loughney> Hannes - NSIS in an off-path world - the control plane
[21:07:16] <loughney> NSIS can be used in an off-path mode, as described, but the verticle signaling is open, depending upon the need.
[21:13:25] <loughney> Someone from Lucent - about the NGN/ITU-T - purpose of resource admission control is to provide real-time configuration of routers along the path
[21:14:06] <loughney> - how you mark packets based upon a certain criteria. Question is the diffserv marking real-time or not?
[21:14:35] <loughney> Scott says yes, let's discuss.
[21:14:35] <loughney> Jeff Pulliam - Domain Managed QoS
[21:15:01] <loughney> descriped in draft-nichols-dspel-strawman-arch-00
[21:17:21] --- Bill has joined
[21:19:39] <loughney> Need standards for diffserv comtrol plan arch.
[21:22:39] <loughney> Hannes - read the draft and has lots of questions. Doesn't know what problem there is.
[21:23:55] --- admcd has left
[21:24:39] <loughney> Scott - how to think about configuring a diffserv environment
[21:25:01] <loughney> Allison - I don't understand the question about how to think about how to think about ...
[21:26:24] <loughney> Scott - the document didn't have the problem that he had in mind.
[21:28:02] <loughney> Scott admits he has articulate very well.
[21:28:31] <loughney> Kathy - draft doesn't say we should go & prototype it - but this is the general model and report what they have done.
[21:33:03] <Roy Brabson> Diff
[21:33:10] --- Roy Brabson has left
[21:33:23] <loughney> David Kessens - what a sense of the room.
[21:35:57] <loughney> Scott - we have not been clear enough about what we are asking - will produce more material.
[21:36:25] <loughney> Bert - don't have a clear problem statement & will wait to see if there interest on the mailing.
[21:37:37] --- admcd has joined
[21:37:55] --- ray_atarashi has left
[21:43:13] --- Yoshifumi Atarashi has left
[21:43:41] <loughney> Open mic - no takers.
[21:43:47] --- venaas has left: Logged out
[21:43:47] <loughney> AOB?
[21:43:49] <loughney> Beer!
[21:43:51] --- loughney has left
[21:43:59] --- raeburn has left: Disconnected
[21:44:08] --- frodek has left
[21:44:21] --- Bill has left: Logged out
[21:44:26] --- kivinen has left
[21:45:50] --- admcd has left
[21:46:04] --- david has left
[21:46:20] --- abierman has left
[21:56:03] --- sleinen has left: Disconnected
[22:05:59] --- bert has left