IETF
idr
idr@jabber.ietf.org
Tuesday, July 22, 2014< ^ >
Room Configuration
Room Occupants

GMT+0
[12:46:53] Pratima Kini joins the room
[12:47:20] Pratima Kini leaves the room
[12:47:37] Jeffrey Haas joins the room
[12:48:13] room streamer joins the room
[12:49:02] Slide streamer joins the room
[12:49:28] Slide streamer leaves the room
[12:50:05] stream slides joins the room
[12:52:35] Simon Romano joins the room
[12:53:22] Paolo Saviano joins the room
[12:57:18] Deniz Bahadir joins the room
[12:59:30] Deniz Bahadir leaves the room
[13:00:30] Pratima Kini joins the room
[13:00:40] Robert Raszuk joins the room
[13:00:43] <Jeffrey Haas> I'll be handling much of the jabber scribe duties for this IDR session.
[13:00:55] <Jeffrey Haas> If you need your comments channeled to the mic, please let me know.
[13:00:55] Deniz Bahadir joins the room
[13:04:38] atanu joins the room
[13:05:02] <Jeffrey Haas> Session begins.
[13:05:13] <Jeffrey Haas> Note Well.  That includes you on jabber.
[13:05:37] Wes George joins the room
[13:06:39] <Jeffrey Haas> RFCs finished.
[13:07:02] <Jeffrey Haas> Route-refresh and aigp in rfc editor queue.
[13:07:09] Weiler joins the room
[13:07:13] <Jeffrey Haas> WGLC coming
[13:07:26] <Jeffrey Haas> redirect-rt-bis, gr-notification, error-handling, as-migration.
[13:07:30] <Weiler> I'm finding the stream audio level very low; any others having problems?
[13:07:48] <Jeffrey Haas> Is it better now? John has moved closer to mic
[13:08:03] <Weiler> no
[13:08:05] <Jeffrey Haas> redirect-rt-bis. please review.
[13:08:21] Paolo Saviano leaves the room
[13:08:27] <Pratima Kini> the volume has to be kept at max to hear clearly
[13:08:44] <Jeffrey Haas> we're notifying the secretariat to look at audio streaming issus.
[13:08:55] <Jeffrey Haas> gr-notification - light response. please review.
[13:09:24] <Jeffrey Haas> Note to draft authors: 2 implementations - go to wgl and iesg.
[13:09:32] room streamer leaves the room
[13:09:32] Robert Raszuk leaves the room
[13:09:32] Deniz Bahadir leaves the room
[13:09:33] stream slides leaves the room
[13:09:35] Simon Romano leaves the room
[13:09:36] <Jeffrey Haas> if not 2 implementation, can request if ok by wg.
[13:09:42] <Jeffrey Haas> if 3 years and no implementations, sue will be checking.
[13:09:59] <Jeffrey Haas> The new "waiting for implementation" state can be used to park the draft.
[13:10:02] Paolo Saviano joins the room
[13:10:28] mikemlb joins the room
[13:10:38] <Jeffrey Haas> Looking for document shepherds.  Please talk to the chairs.
[13:11:38] <Jeffrey Haas> First presentation: SDIR and bgpsec.
[13:12:03] <Jeffrey Haas> BGPSEC and SIDR
[13:12:37] <Jeffrey Haas> Remote listeners, is Sandy audible better than the chairs?
[13:12:46] <Pratima Kini> yes
[13:12:55] <Jeffrey Haas> I'll let the chairs know.
[13:13:07] <Weiler> yes, though marginally?
[13:13:15] <Jeffrey Haas> IDR attention requested.
[13:14:05] Robert Raszuk joins the room
[13:14:30] <Jeffrey Haas> John: when is bgpsec covered during the friday sidr meeting?
[13:14:35] <Jeffrey Haas> Sandy: First item in the agenda.
[13:15:25] <Jeffrey Haas> John: Where in the publication process is bgpsec?  Ready for wglc soon?
[13:16:09] <Jeffrey Haas> Sandy: Finishing up.  Open issues addressed.  Long pause while threats and req's went through.  Expect more feedback but think we're close to done and hope wglc can happen soon.
[13:17:06] <Jeffrey Haas> Unknown speaker.
[13:17:08] rjs joins the room
[13:17:33] <Jeffrey Haas> Presentation: Error handling.
[13:17:53] <Jeffrey Haas> Slide: draft status
[13:18:51] <Jeffrey Haas> Slide: What's a valid ip host address
[13:18:51] slideer a joins the room
[13:19:22] streamroom a joins the room
[13:20:15] <Jeffrey Haas> Use the iana special address registry?  Good idea?
[13:20:34] <Jeffrey Haas> Slide: A twisty maze of passages
[13:21:02] <Jeffrey Haas> Slide: … all different
[13:21:18] <Jeffrey Haas> (Robert's fault!) :-)
[13:21:25] Simon Romano joins the room
[13:22:03] Paolo Saviano leaves the room
[13:22:31] <Jeffrey Haas> analysis as venial, mortal or trivial.
[13:22:45] Pratima Kini leaves the room
[13:22:46] <Jeffrey Haas> Slide: Are all next-hops errors "venial"
[13:22:46] Pratima Kini joins the room
[13:23:34] <Jeffrey Haas> Slide: When in doubt, punt
[13:24:01] <Robert Raszuk> ;-) Btw on Meetecho the ECHO is terrible ....
[13:24:29] atanu leaves the room
[13:24:32] <Jeffrey Haas> Will notify them on pause.  John is using the walking mic next to the chairs mic.
[13:24:36] <Jeffrey Haas> Slide: Proposal
[13:25:05] <Robert Raszuk> Thx Jeff !
[13:25:42] <Jeffrey Haas> better?
[13:25:53] <Robert Raszuk> Not at all
[13:26:00] <Jeffrey Haas> sorry.
[13:26:13] <Jeffrey Haas> mic: Eric
[13:26:23] akatlas joins the room
[13:26:26] <Jeffrey Haas> eric: Declare victory. 25 years of experience.
[13:26:43] <Jeffrey Haas> "Make it someone else's problem"
[13:26:50] <Jeffrey Haas> Slide: TE Path Attribute
[13:28:39] fenner joins the room
[13:28:52] <Jeffrey Haas> Jeff: Heckling the problem.
[13:29:09] <Jeffrey Haas> Slide: Next steps.
[13:29:22] <Jeffrey Haas> New WGLC. Asymptotically converge on done.
[13:29:27] <Jeffrey Haas> Questions?
[13:29:36] <Jeffrey Haas> Presentation: BGP Remote Next-hop.
[13:29:57] Paolo Saviano joins the room
[13:29:58] <Jeffrey Haas> Presenter G. Van de Velde.
[13:30:07] <fenner> noc says the audio stream sounds great to them, are there still audio problems?
[13:30:18] <Jeffrey Haas> Slide: What is bgp remote next hop?
[13:31:07] <Jeffrey Haas> Slide: Applicable use cases
[13:31:53] <Jeffrey Haas> Slide: Brief history
[13:32:20] <Jeffrey Haas> New use cases. new sub-tlvs. more feedback.
[13:32:22] <Robert Raszuk> Hi Bill ... I do .. I think it is like summary of mic signal with local PC mic - and only affects meetecho.
[13:32:35] <Jeffrey Haas> Slide: rnh highlights.
[13:33:03] <fenner> oh. Try the regular stream then - http://ietf90streaming.dnsalias.net/ietf/ietf908.m3u
[13:33:44] <Jeffrey Haas> Slide: TLV Format
[13:33:56] <Jeffrey Haas> Slide:Benfits
[13:34:07] Robert Raszuk leaves the room
[13:34:20] <Weiler> fenner: no data.  I disconnected.
[13:34:44] <Jeffrey Haas> Slide: Next steps
[13:34:57] Robert  Raszuk joins the room
[13:35:04] <Jeffrey Haas> Ready for wg adoption?
[13:35:08] <Jeffrey Haas> Questions
[13:35:21] <fenner> weiler: my understanding is that meetecho quality is low and http://ietf90streaming.dnsalias.net/ietf/ietf908.m3u quality is normal
[13:35:22] <Jeffrey Haas> John: ready for new version. Is it worth people's time to read current version.
[13:35:34] <Jeffrey Haas> G: Sub-tlvs to be added. Can review now.
[13:35:44] <Jeffrey Haas> John: Will move adoption call to list.
[13:35:46] <Weiler> fenner: normal stream level was definitely low.  no data of how it is now.
[13:35:53] <fenner> ok
[13:35:57] <Jeffrey Haas> Sandy: Uninformed question.
[13:35:58] <Weiler> level, not quality
[13:36:16] <Jeffrey Haas> S: … optional transitive. similar to as4-path problem.
[13:36:53] <Jeffrey Haas> … issue of optional transitive issues similar to what we've seen. have you addressed error handling issues and there's format issues?
[13:37:05] <Jeffrey Haas> … what's the error-handling semantics.  didn't find it.
[13:37:35] <Jeffrey Haas> G: draft saying what happens if it doesn't understand not in there, but this specific problem addressed.  will add a few lines
[13:38:04] <Jeffrey Haas> Shuwan Zhang: Can this be used for inter-as?
[13:38:15] <Jeffrey Haas> G: Yes.
[13:39:02] <Jeffrey Haas> Weiguo from Huawei:
[13:39:26] <Jeffrey Haas> GRE tunnel. ipfix advertisement? Relationship with evpn draft?
[13:40:08] <Jeffrey Haas> Saikot Cisco: Mainly used for l2vpn.  This is mostly for all address families
[13:40:24] <Jeffrey Haas> DJ Rao Cisco: This is reusable.
[13:40:40] <Jeffrey Haas> Some discussion about re-using it for evpn is an open question.
[13:41:39] <Jeffrey Haas> Eric Rosen: No new address family is required.  We already have RFC 5512 but required new safi.  Is it a good idea to avoid new safis?  Maybe good idea to split it into diff address families
[13:41:54] <Jeffrey Haas> Some of the use cases are not carefully enough described, in particular ipv6 multihoming.
[13:42:42] <Jeffrey Haas> Presentation: Nexthop path record attribute, presenter: Robin Zhang
[13:42:58] <Jeffrey Haas> Slide: Nexthop path record attribute
[13:43:42] <Jeffrey Haas> Slide: bgp nexthop path record attribute description
[13:44:44] <Jeffrey Haas> Slide: Deployment considerations
[13:45:18] <Jeffrey Haas> Slide: Next steps
[13:45:33] Paolo Saviano leaves the room
[13:45:54] <Jeffrey Haas> Questions?
[13:46:07] <Jeffrey Haas> Eric Rosen: What is this for?
[13:46:18] <Jeffrey Haas> Vector routing hides information on purpose.
[13:46:42] <Robert  Raszuk> He said for best path selection ...
[13:46:45] <Jeffrey Haas> Maybe useful for SDN?
[13:46:58] <Jeffrey Haas> Need me to channel that Robert?
[13:47:13] <Robert  Raszuk> Nope ...
[13:47:16] <Robert  Raszuk> he said it
[13:47:19] <Jeffrey Haas> Sue: One of the seamless mpls use cases.
[13:47:23] <Robert  Raszuk> It's like rsvp RRO :)
[13:47:44] <Jeffrey Haas> Also could be used in mobile topologies in TE.
[13:48:28] <Jeffrey Haas> Eric: Is use case in there?
[13:48:43] <Robert  Raszuk> yep ... it does sound like it may have some valid uses
[13:48:45] <Jeffrey Haas> (back and forth) unclear if it's in there. can discuss
[13:49:42] <Jeffrey Haas> Saikat: covered by as-path?
[13:49:47] <Jeffrey Haas> Robin: Not really.
[13:50:01] <Jeffrey Haas> S: In route reflector, we have clusterid/list
[13:50:40] <Jeffrey Haas> Sue: If you go up and down the cluster, how do you know where you've gone?
[13:51:22] <Jeffrey Haas> S: Similar to adding as-path discipline.
[13:51:32] <Jeffrey Haas> Sue: There are cases where we can't just use cluster-id
[13:51:48] <Jeffrey Haas> Presentation: Extensions to rt-constrain for hierarchical RR scenario. presented by Jie Dong.
[13:52:06] <Jeffrey Haas> Slide: Problem statement
[13:52:33] <Jeffrey Haas> Slide: Typical scenario
[13:53:16] <Jeffrey Haas> Slide: Typical scenario (2)
[13:53:23] <Robert  Raszuk> For one it may simplify EPE on controller .. even if only single last next hop of your peering AS is carried.
[13:53:49] <Jeffrey Haas> Slide: Proposed solution.
[13:54:52] <Jeffrey Haas> Slide: Received comments
[13:55:49] <Jeffrey Haas> Slide: Next steps
[13:56:28] <Jeffrey Haas> Saikat: There's a diff between rt-c and normal routes.
[13:56:51] <Jeffrey Haas> we should treat it as if we were always originating.
[13:57:23] <Jeffrey Haas> original rfc noted that you could do that, but go back to original alternate mechanism
[13:58:00] rjs leaves the room
[13:58:28] <Jeffrey Haas> Jeff and saikot: talking past each other.
[13:58:46] <Jeffrey Haas> Presentation: Route target contrained distribution of routes without route-targets
[13:59:17] <Jeffrey Haas> audio check from remote listeners please?
[13:59:47] <Pratima Kini> barely audible
[14:00:25] <Jeffrey Haas> Slide: Corner case
[14:00:32] <Jeffrey Haas> better audio?
[14:00:39] <Pratima Kini> yes… thanks
[14:02:56] <Jeffrey Haas> Slide: Solution
[14:03:19] rjs joins the room
[14:05:46] <Jeffrey Haas> Jeff: time for bis?
[14:06:26] <Jeffrey Haas> Saikot: how to deal with l2vpn case?
[14:07:19] <Jeffrey Haas> should be able to signal this case?
[14:07:25] <Jeffrey Haas> which AFs its for?
[14:08:00] <Jeffrey Haas> Sue: Have you considered flags?
[14:08:16] <Jeffrey Haas> Eric: No. Maybe we do small things or as a bis. :-)
[14:08:21] <Jeffrey Haas> Sue: that's a political answer. :-)
[14:08:27] <Jeffrey Haas> John: Short answer is no. :-)
[14:09:07] John Doe joins the room
[14:09:09] <Jeffrey Haas> Eric: These slides are simple case for no-RTs.  Whether we want to have a knob for specific af's and signal it is a different issue.
[14:09:25] <Jeffrey Haas> Presentation: draft-litkowski-idr-rtc-interase presented by Stephane.
[14:09:29] Passov ssd joins the room
[14:09:40] <Jeffrey Haas> Slide: Problem statement
[14:10:13] John Doe leaves the room
[14:10:35] <Jeffrey Haas> Slide: Problem statement (2)
[14:10:51] <Jeffrey Haas> Slide: Problem statement (3)
[14:10:58] A A joins the room
[14:11:18] <Jeffrey Haas> Slide: Proposal
[14:11:27] John Doe joins the room
[14:11:36] A A leaves the room
[14:11:39] John Doe leaves the room
[14:11:53] <Jeffrey Haas> Slide: Proposal (2)
[14:12:23] <Jeffrey Haas> Slide: Conclusion and next steps
[14:13:52] <Passov ssd> Sorry, Meetecho audio sounds better, now?
[14:14:39] <Robert  Raszuk> I am using stream instead ,,,, do not want to break it now by checking. Maybe after IDR :))
[14:15:17] <Jeffrey Haas> Jeff: Should these exception cases for rt-c and error handling be tracked in registry.
[14:15:23] <Jeffrey Haas> Sue: We may want to signal.
[14:15:42] <Jeffrey Haas> John: Base spec is broken for disjoint case.  When written, no one would have thought about CE doing RT-C.
[14:16:06] <Jeffrey Haas> Sue: When specced your use case was considered broken.
[14:16:13] <Jeffrey Haas> But we must deal with pragmatism.
[14:16:27] <Jeffrey Haas> Saikot: We're trying to keep base behavior
[14:16:28] Passov ssd leaves the room
[14:16:41] <Jeffrey Haas> Stephane: Issue is multi-homed scenarios, you'll have less efficient distribution
[14:17:00] <Jeffrey Haas> No routing loops, just more paths.
[14:17:25] <Jeffrey Haas> Stephane: Knows Cisco doesn't do path pruning.
[14:17:38] <Jeffrey Haas> Presentation: Semgnet routing bgpl egress peer engineering ext
[14:18:00] <Jeffrey Haas> Slide: Motivations
[14:18:29] <Robert  Raszuk> It's not that Cisco does not do path pruning .. effectively they did not implement the limitation of originator rewrite to only local RTC routes
[14:18:45] <Robert  Raszuk> And that may be right fix after all
[14:19:02] <Jeffrey Haas> Robert, please post that comment to mailing list to feed discussion
[14:19:19] <Robert  Raszuk> ok
[14:19:25] <Jeffrey Haas> Thanks!
[14:19:36] <Jeffrey Haas> Slide: Reference Diagram
[14:19:43] <Jeffrey Haas> Slide: Objective: Centralized egress peer engineering
[14:20:29] <Jeffrey Haas> slide: ebgp peering topology
[14:21:20] <Jeffrey Haas> Slide: Automated bgp peering sid allocation
[14:22:11] dredgie@wtfismyip.com joins the room
[14:22:20] akatlas leaves the room
[14:23:09] <Jeffrey Haas> Slide: BGP EPE routes
[14:23:41] <Jeffrey Haas> Slide: Controller - decision
[14:23:50] Weiler leaves the room
[14:24:22] <Jeffrey Haas> Slide: Peer NLRI type
[14:25:39] <Jeffrey Haas> Slide: (2)
[14:26:01] <Jeffrey Haas> Sldie: Peer Attributes
[14:26:26] <Jeffrey Haas> Slide: Peering Segments
[14:27:08] <Jeffrey Haas> Question: Bhump(?) controller?
[14:27:56] <Jeffrey Haas> Are you going to share details about implementation for controller?
[14:28:01] <Jeffrey Haas> Stefano: Out of scope.
[14:28:37] <Jeffrey Haas> End to end centralized or domain specific?
[14:28:41] <Jeffrey Haas> Stefano: Up to you.
[14:29:27] <Jeffrey Haas> Hannes Gredler: Thanks for proposing bgp-ls.  Couldn't understand why need a dedicated nlri type for this?
[14:29:39] Paolo Saviano joins the room
[14:29:47] <Jeffrey Haas> Spec has protocol id of direct to describe links to lan where no igp is run.
[14:30:01] <Jeffrey Haas> Have a way already to express link-state edges between ases.
[14:30:17] <Jeffrey Haas> Can already model your three cases.
[14:30:57] <Jeffrey Haas> Stefano: would let saikot answer.  briefly, didnt like idea of using protocol id. source of information. this is topology info, not the source.  sure there are multiple ways to model this.
[14:31:20] <Jeffrey Haas> Hannes: Thats' fine. we have links, nodes and prefixes. you're describing links.
[14:31:20] mikemlb leaves the room
[14:31:35] <Jeffrey Haas> Stefano: yes and no, may want to express multi-object connectivity to ASes.
[14:31:46] fenner leaves the room
[14:31:47] <Jeffrey Haas> Slide: Objective: centralized egress peer engineering
[14:32:19] <Jeffrey Haas> Hannes: Lets do this like the igps.
[14:32:54] <Jeffrey Haas> Stefano: Found my solution cleaner.
[14:32:58] <Jeffrey Haas> John: Discuss on the list.
[14:33:32] <Jeffrey Haas> Saikot: have trouble distinguishing bgp from igp types?
[14:34:29] <Jeffrey Haas> Robin - Huawei: implementation specific, could be enhanced for route reflector.
[14:34:38] <Jeffrey Haas> Sue: Need to cut comments.
[14:35:11] <Jeffrey Haas> John: is it the case that E and F in AS3 are capable to forward exact same prefixes.
[14:35:29] rjs leaves the room
[14:36:04] <Jeffrey Haas> Stefano: Use existing bgp scheme, don't make the controller break bgp behaviors.
[14:36:26] rjs joins the room
[14:36:31] <Jeffrey Haas> Presentation: BGP extensions for inter-as te link distribution.  Jie Dong
[14:36:38] <Jeffrey Haas> Slide: Background
[14:36:58] fenner joins the room
[14:37:34] <Jeffrey Haas> Slide: Typical scenario
[14:38:12] <Jeffrey Haas> Slide: Proposed Solution
[14:38:52] <Jeffrey Haas> Slide: BGP extensions
[14:39:53] <Jeffrey Haas> Slide: Next steps
[14:39:57] <Jeffrey Haas> questions?
[14:40:08] <Jeffrey Haas> Hannes: Same comment. It all can be done using existing stuff.
[14:40:15] <Jeffrey Haas> Jie: New new protocol id?
[14:40:19] <Jeffrey Haas> Hannes: No.
[14:40:27] <Jeffrey Haas> Hannes: These should be direct.
[14:41:09] <Jeffrey Haas> Stefano: Disagree with Hannes.  Distinction is useful.  BGP-LS is focused on igp.  We're talking all sort of tables.
[14:41:23] <Jeffrey Haas> The distinction is flow of info across protocols.
[14:41:35] Paolo Saviano leaves the room
[14:41:53] <Jeffrey Haas> Presentation: BGP extneions for service-oriented mpls path programming. shunwan zhuang
[14:42:14] <Jeffrey Haas> Slide: Introduction
[14:42:56] <Jeffrey Haas> Slide: Use cases for unicast service mpls path programming
[14:44:38] <Jeffrey Haas> Slide: Architecture of mpls path programming
[14:45:18] <Jeffrey Haas> Slide: BGP Extensions requirements for service-oriented mpls path programming
[14:47:09] <Jeffrey Haas> Slide: Download of MPLS Path
[14:47:23] <Jeffrey Haas> (sorry if he's gone quiet, he's got mic management issues)
[14:48:28] <Jeffrey Haas> (presenter reading slide verbatim)
[14:48:39] <Jeffrey Haas> Slide: Download of MPLS Path (Cont.)
[14:49:20] <Jeffrey Haas> Slide: Download of MPLS Path (Cont.) (2)
[14:50:32] <Jeffrey Haas> Slide: Extended unicast tunnel attribute
[14:50:40] Paolo Saviano joins the room
[14:50:47] dan cohn joins the room
[14:51:05] <Jeffrey Haas> Slide: Extended PMSI
[14:51:08] <Jeffrey Haas> Slide: Next step
[14:51:18] <Jeffrey Haas> Questions:
[14:51:38] <Jeffrey Haas> Saikot: Assume want to put label stack in path attribute instead of nlri
[14:51:42] <Jeffrey Haas> if you do, you destroy packing of nlri
[14:52:14] fenner leaves the room
[14:55:00] rjs leaves the room
[14:55:01] <Jeffrey Haas> Lou berger: what to do if path programming fails?
[14:55:33] rjs joins the room
[14:56:01] mikemlb joins the room
[14:57:34] Jeffrey Haas is distracted and missed the current argument
[14:58:04] Paolo Saviano leaves the room
[14:58:08] <Jeffrey Haas> John: goto spring;
[14:58:23] <Jeffrey Haas> Presentation: draft-litkowski-idr-bgp-timestamp
[14:58:44] <Jeffrey Haas> Stephane Litkowski
[14:58:51] <Jeffrey Haas> Slide: PRoblem Statement
[14:59:34] <Jeffrey Haas> Slide: (2)
[14:59:55] Paolo Saviano joins the room
[15:00:16] <Jeffrey Haas> Slide: Proposal
[15:00:46] <Jeffrey Haas> Slide: Architecture for propagation time measurement
[15:01:40] <Jeffrey Haas> Slide: Proposal
[15:02:15] slm joins the room
[15:02:21] <Jeffrey Haas> Slide: (2)
[15:03:44] <Jeffrey Haas> Slide: (3)
[15:03:58] <Jeffrey Haas> Slide: Inter-as
[15:04:59] dan cohn leaves the room
[15:05:03] <Jeffrey Haas> Slide: Compared to BMP
[15:06:26] <Jeffrey Haas> Slide: Next steps
[15:07:12] <Jeffrey Haas> Questins:
[15:07:20] <Jeffrey Haas> Saikot: Mess up packing.
[15:07:24] <Jeffrey Haas> Stephane: Yes.
[15:07:38] <Jeffrey Haas> Saikot: L2VPN in particular!
[15:08:10] <Jeffrey Haas> If only TS changes, we should not resend.
[15:08:14] <Jeffrey Haas> BMP is really right place to go.
[15:08:39] Simon Romano leaves the room
[15:08:41] <Jeffrey Haas> Stephane: ONly adding TS on a *few* nlri.
[15:09:25] <Jeffrey Haas> John: Point out several options.  Optional transitive - it'll leak far.
[15:10:36] <Jeffrey Haas> Robin: Timestamp is on receipt. more useful on receive *and* send?
[15:12:58] <Jeffrey Haas> Sue: look at benchmarking drafts/ talk to chair
[15:13:17] <Jeffrey Haas> Saikot: Measuring one probe route doesn't give you good system wide observation
[15:13:22] <Jeffrey Haas> Presentation: Flowspec interfaces
[15:13:38] <Jeffrey Haas> Slide: Problem statement
[15:13:46] <Jeffrey Haas> Slide: Use cases
[15:14:44] <Jeffrey Haas> Slide: interface-set extended community
[15:15:32] <Jeffrey Haas> Slide: example
[15:16:09] <Jeffrey Haas> Slide: example (2)
[15:16:42] <Jeffrey Haas> Slide: Security considerations
[15:17:48] <Jeffrey Haas> Slide: Discussion outcomes
[15:20:45] <Jeffrey Haas> Slide: next steps
[15:20:55] <Jeffrey Haas> Questions:
[15:21:28] <Jeffrey Haas> Jon Mitchell: there are already implementations that let you apply specs groups.
[15:21:35] <Jeffrey Haas> Not sure if you can encode ingress/egress.
[15:21:48] <Jeffrey Haas> same point on automatic groupings. ebgp may be "internal".
[15:22:24] Wes George leaves the room
[15:22:34] <Jeffrey Haas> Andy Karch: base rfc should be run on ingress only.
[15:22:41] <Jeffrey Haas> where would you do it on egress?
[15:23:06] <Jeffrey Haas> for labeled traffic, what does this mean for multiple tables (vpn routes?)
[15:23:30] rjs leaves the room
[15:24:02] <Jeffrey Haas> Presentation: AS Migration - wes george
[15:24:12] <Jeffrey Haas> Slide: Changelog.
[15:25:02] <Jeffrey Haas> Slide: Open items for wglc?
[15:28:40] slm leaves the room
[15:29:33] Pratima Kini leaves the room
[15:30:00] mikemlb leaves the room
[15:30:12] Paolo Saviano leaves the room
[15:30:23] Wes George joins the room
[15:30:34] Jeffrey Haas leaves the room
[15:30:39] Robert  Raszuk leaves the room
[15:31:33] Wes George leaves the room
[15:33:00] slideer a leaves the room
[15:33:04] streamroom a leaves the room
[15:46:57] Fix Recorder joins the room
[15:49:15] Fix Recorder leaves the room
[15:54:12] fix rec joins the room
[15:56:00] fix rec leaves the room
[15:58:42] ppp ppp joins the room
[15:59:44] ppp ppp leaves the room
[16:15:26] Jeffrey Haas joins the room
[16:15:29] Jeffrey Haas leaves the room
[16:31:40] dredgie@wtfismyip.com leaves the room
[16:44:14] mikemlb joins the room
[16:45:40] mikemlb leaves the room
[17:05:51] Wes George joins the room
[17:06:03] Wes George leaves the room
[17:21:40] slm joins the room
[17:24:29] rjs joins the room
[18:08:40] slm leaves the room
[18:10:32] rjs leaves the room
[18:17:19] slm joins the room
[18:29:39] rjs joins the room
[18:36:14] rjs leaves the room
[20:31:00] slm leaves the room
[20:57:57] slm joins the room
[22:44:26] slm leaves the room
[23:21:54] slm joins the room
[23:31:02] slm leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!