Starting in one minute. Last chance to check your mic.
", "time": "2024-09-23T14:01:32Z"}, {"author": "John Scudder", "text": "Since tooltips aren't displaying properly on my client :-/ would one of you be so kind as to remind me what button I'll need to click to control the slides when my turn comes up? I think it's the bottom center one, to the left of the join queue hand, but I'm not sure. TIA.
", "time": "2024-09-23T14:12:04Z"}, {"author": "Luuk Hendriks", "text": "That one says 'Ask slides' indeed
", "time": "2024-09-23T14:12:42Z"}, {"author": "Randy Bush", "text": "my faulty memory is you get new control panel
", "time": "2024-09-23T14:12:45Z"}, {"author": "Natrajan Venkataraman", "text": "@john, thats right ... Bottom middle
", "time": "2024-09-23T14:13:46Z"}, {"author": "John Scudder", "text": "TY!
", "time": "2024-09-23T14:13:56Z"}, {"author": "Natrajan Venkataraman", "text": "@John ... my bad ... its the 4th button from the left
", "time": "2024-09-23T14:15:20Z"}, {"author": "Randy Bush", "text": "to request: the little SIM card to the left of the raised hand.
", "time": "2024-09-23T14:15:47Z"}, {"author": "John Scudder", "text": "Yep, got it. Thanks.
", "time": "2024-09-23T14:16:25Z"}, {"author": "Ketan Talaulikar", "text": "@Kaliraj, it is a bandwidth derived metric the IGP draft has the details on the options available.
", "time": "2024-09-23T14:23:27Z"}, {"author": "Ketan Talaulikar", "text": "", "time": "2024-09-23T14:24:07Z"}, {"author": "Ketan Talaulikar", "text": "Not well and trying to conserve my voice. Apologies. Perhaps @Shraddha can clarify as well.
", "time": "2024-09-23T14:25:15Z"}, {"author": "Shraddha Hegde", "text": "@kaliraj, In IGPs bandwidth metric is derived from link bandwidth. Its not directly the link bandwidth most of the times. Its reference bandwdith/ actual link bandwidth.
", "time": "2024-09-23T14:25:18Z"}, {"author": "Kaliraj Vairavakkalai", "text": "ok, I thought igp-bandwidth can possibly be conveyed in link-bw as-well. and from receiving bgp speakers perspective, it looks the same to me. may be i will connect with shraddha offline to understand more.
", "time": "2024-09-23T14:29:59Z"}, {"author": "Jeffrey Haas", "text": "Once we start talking about metrics in an igp sense, or \"distance to nexthop\" in a bgp sense, we're talking about path to the preferred prefix.
", "time": "2024-09-23T14:32:59Z"}, {"author": "Jeffrey Haas", "text": "Note for John, link local peering need not always carry a link local nexthop.
", "time": "2024-09-23T14:35:06Z"}, {"author": "Ketan Talaulikar", "text": "@Kaliraj, the AMetric conveys the \"distance to nexthop\" and seems to me more semantically correct than deriving from an ExtComm.
", "time": "2024-09-23T14:35:11Z"}, {"author": "Srihari Sangli", "text": "2
", "time": "2024-09-23T14:38:23Z"}, {"author": "Srihari Sangli", "text": "@Ketan, the extcomm carries the bandwidth information and I see the two are 2 different items.
", "time": "2024-09-23T14:38:54Z"}, {"author": "Ketan Talaulikar", "text": "@Srihari, I agree.
", "time": "2024-09-23T14:39:26Z"}, {"author": "Kaliraj Vairavakkalai", "text": "I can connect with you guys offline to understand.
", "time": "2024-09-23T14:40:06Z"}, {"author": "Srihari Sangli", "text": "sure
", "time": "2024-09-23T14:40:54Z"}, {"author": "Natrajan Venkataraman", "text": "@Kaliraj, I also feel that its just propagating the same thing in two different ways ... Bandwidth is a real accumulated value and it cannot change much ... we can discuss internally with Srihari/Shraddha
", "time": "2024-09-23T14:41:24Z"}, {"author": "Kaliraj Vairavakkalai", "text": "ack
", "time": "2024-09-23T14:41:44Z"}, {"author": "Susan Hares", "text": "NHC - has had a WG LC. It will need 2 implementations before a 2nd WG LC. (shepherd).
", "time": "2024-09-23T14:42:37Z"}, {"author": "Jeffrey Haas", "text": "Thanks, Sue. I had meant in response to the addition to the link local issue.
", "time": "2024-09-23T14:43:07Z"}, {"author": "Jeffrey Haas", "text": "It's been stable beyond that.
", "time": "2024-09-23T14:43:19Z"}, {"author": "John Scudder", "text": "Ack. My main point was just that v16 will introduce a normative change to handle the link-local corner case. I would anticipate the WG wants to review that, whether we call it a \"last call\" or not.
", "time": "2024-09-23T14:43:53Z"}, {"author": "Jeffrey Haas", "text": "Tony P would be proud of slide 8.
", "time": "2024-09-23T14:47:59Z"}, {"author": "John Scudder", "text": "^
", "time": "2024-09-23T14:48:07Z"}, {"author": "Jeffrey Haas", "text": "Discussion point MNH slide 9, bgp-car optional-nlri tlvs
", "time": "2024-09-23T14:50:47Z"}, {"author": "Jeffrey Haas", "text": "Discussion point MNH slide 10, comparability to fsv2 error handling situations.
", "time": "2024-09-23T14:51:18Z"}, {"author": "Jeffrey Haas", "text": "Discussion point MNH slide 11, generic metric continuity
", "time": "2024-09-23T14:53:36Z"}, {"author": "Srihari Sangli", "text": "How does one know that intermediate router has honored the C bit and has done cumulation
", "time": "2024-09-23T14:55:51Z"}, {"author": "Nat Kao", "text": "I think MNH also signals the ordering of the NHs.
", "time": "2024-09-23T15:12:47Z"}, {"author": "Robert Raszuk", "text": "How about tunneling encap attribute ... astill very muxch related to fwd-ing
", "time": "2024-09-23T15:14:05Z"}, {"author": "Ketan Talaulikar", "text": "@Robert, this is like TEAv2?
", "time": "2024-09-23T15:14:30Z"}, {"author": "Kaliraj Vairavakkalai", "text": "TEA can possibly be a MNH Constrain called 'tunnel-type'.
", "time": "2024-09-23T15:15:28Z"}, {"author": "Robert Raszuk", "text": "My worry is that MNH may very soon if not day one become an elephant attribute
", "time": "2024-09-23T15:16:40Z"}, {"author": "Jeffrey Haas", "text": "Part of the very legitimate debate here is whether we make one larger complicated attribute, or many loosely coupled ones where the rules are scattered around.
", "time": "2024-09-23T15:18:10Z"}, {"author": "Mankamana Mishra", "text": "will loosely coupled one not give more control over it ?
", "time": "2024-09-23T15:19:01Z"}, {"author": "Robert Raszuk", "text": "Today we rather have the latter case
", "time": "2024-09-23T15:19:07Z"}, {"author": "Jeffrey Haas", "text": "TEA, IMNSHO, has become a very messy \"elephant attribute\" where the contents might not be even in internal harmony.
", "time": "2024-09-23T15:19:20Z"}, {"author": "Robert Raszuk", "text": "More flexibility more difficult control ...
", "time": "2024-09-23T15:19:36Z"}, {"author": "Jeffrey Haas", "text": "And more inconsistency.
", "time": "2024-09-23T15:19:46Z"}, {"author": "Jeffrey Haas", "text": "Route selection consistency has been hard to discuss even with current proposals.
", "time": "2024-09-23T15:20:04Z"}, {"author": "Robert Raszuk", "text": "more rules to get all collisions resolved
", "time": "2024-09-23T15:20:04Z"}, {"author": "Jeffrey Haas", "text": "^
", "time": "2024-09-23T15:20:11Z"}, {"author": "Srihari Sangli", "text": "even if we agree with MNH and get it deployed, the network will be dealing with both types of information for a long time. that's a bigger worry IMO
", "time": "2024-09-23T15:21:13Z"}, {"author": "Robert Raszuk", "text": "But putting all in the same bus means more risk too ... unless we do it well - for example with MNH capabilites :)
", "time": "2024-09-23T15:21:14Z"}, {"author": "Jeffrey Haas", "text": "Much of the discussion overlaps scoping (a broader discussion we've been having) and incremental deployment for new features.
", "time": "2024-09-23T15:21:56Z"}, {"author": "Jie Dong", "text": "Did the author say the information in MNH attribute is not used for control plane path selection?
", "time": "2024-09-23T15:22:21Z"}, {"author": "Jeffrey Haas", "text": "Discussion point float32 LBW, YANG
", "time": "2024-09-23T15:24:59Z"}, {"author": "Kaliraj Vairavakkalai", "text": "@Jie, No, I said the things carried in MNH are Forwarding-attribute related to the EP, most of which may get used to install forwarding to the FIB.
", "time": "2024-09-23T15:25:49Z"}, {"author": "Kaliraj Vairavakkalai", "text": "Info that are pure control plane best-path selection related, that dont need to be carried inside MNH.
", "time": "2024-09-23T15:26:20Z"}, {"author": "Jeffrey Haas", "text": "So, Kaliraj, \"generic metric\" would never be encoded in MNH? Why not?
", "time": "2024-09-23T15:27:30Z"}, {"author": "Shraddha Hegde", "text": "@Kaliraj AIGP and AMetric used for best path selection
", "time": "2024-09-23T15:27:46Z"}, {"author": "Kaliraj Vairavakkalai", "text": "ofcourse, the Forwarding-attribute related to the EP that are carried in the MNH MAY influence path-selection.
\nMy point was, what is carried in the MNH is a forwarding attribute.
", "time": "2024-09-23T15:29:08Z"}, {"author": "Kaliraj Vairavakkalai", "text": "Local-pref is not a forwarding-attribute, so it doesnt come inside MNH. For e.g.
", "time": "2024-09-23T15:29:32Z"}, {"author": "Jeffrey Haas", "text": "The raised point is that if we're putting all nexthop information in one place, we need to reconcile components that are route selection impactful vs. the ones that only impact forwarding behavior.
", "time": "2024-09-23T15:30:16Z"}, {"author": "Srihari Sangli", "text": "LP does change the which nexthop traffic to be sent to. So is AIGP.
", "time": "2024-09-23T15:30:20Z"}, {"author": "Jeffrey Haas", "text": "My suggestion is hold some of these points for last part of interim for voice debate.
", "time": "2024-09-23T15:30:47Z"}, {"author": "Kaliraj Vairavakkalai", "text": "Yes, but LP is not a forwarding-property. So it is not in MNH
", "time": "2024-09-23T15:30:50Z"}, {"author": "Jie Dong", "text": "Thanks Kaliraj, just want to better understand the scope of MNH, and it would be helpful to explain the concept and scope of \"forwarding attribute\" in the draft
", "time": "2024-09-23T15:31:00Z"}, {"author": "Kaliraj Vairavakkalai", "text": "sure, thanks Jie. will describe and excemplify \"Forwarding property\"
", "time": "2024-09-23T15:31:35Z"}]