LSVR IETF 108 WG Notes
======================

Thursday, July 30 2020, 13:00 - 13:50 UTC
Chairs: Victor Kuarsingh, Gunter Van De Velde
Minute Taker: Jacob Uecker
Jabber Logs: https://www.ietf.org/jabber/logs/lsvr/2020-07-30.html

Agenda:
-------
https://www.ietf.org/proceedings/108/agenda/agenda-108-lsvr-03

(Chairs)
 * Status
 * Implementation
 * Updates on Drafts

LSVR WG Status:
---------------
* YANG is MIA
* Next steps = LSVR Re-charter for IETF109

Keyur gives update that LSVR Yang model is being worked. First version should be up and running and submitted to IETF soon.

Presentations:
--------------

* Keyur Applicability Doc Update
 o Version 5 was submitted in March
 o Comments from Boris were incorporated into version 6 (examples and text updates).
 o Version 6 was submitted by Acee

* Keyur BGP SPF Doc Update by Keyur
 o Version 9 was submitted in May
 o Mulitiple implementations since IETF 107
 o Comments have been addressed from routing directorate
 o New contributor
 o Yang model in progress

* Keyur Implementation Draft for BGP-SPF by Keyur
 o Two implemtnations now
   1. ArcOS
   2. FRR Routing Stack
 o Review implementation report (see slides for chart)
 o Alvaro Question: - request to add implementation to draft.
     Response: there’s a separate implementation draft. Add a new section?
     AR: link or pointer would work
     Gunter: added to shepard draft and wiki suggestion
     AR: need as much visibility as possible
     Jacob: reachout out for testing out implementation
     Acee: as far as implementation drafts go, he has some suggestions. would also like to review the results
     Anurag: mentions in the draft about dampining SPF calculations. is that to allow for better scalability?
     Answer (keyur): from an implentation standpoint, wanted to pace it incase there’s a lot of updates.
     Anurag: is it stateful?
     Keyur: Stateful behavior have 3 phases. Different implemtnations can do it different. the 3 default phases in BGP were replaced with SPF comuptation to install router the RIB
     Anurag: is this a local rib or adj out?
     Keyur: this is an implementation specific detail. the draft allows flexibility
     Anurag: is there a recommendation in the draft?
     Keyur: no recommendation in the draft.
     Acee: comment on the backoff. all IGPs have some kind of backoff. in the draft there’s a recommendation.

* Randy Presentation for L3DL:
 o Primary Goal: Discovery of L3 topology and liveness. Not a routing protocol.
   * Jorg Ott did an early review of the draft
   * -06 was reviewed
 o 2 python implementations
   1 is opensource (LSoE)
   1 is in ArcOS (not open source) in GoLang

* Randy Presentation on L3DL Signing
  o Some simple security additions
  o Open PDU has Auth Type and Length with Signature info
  o Key in open PDU is public key
    * 2 General verifications:
      1. TOFU. Trust on the first use (open) and subsequent PDUs use that
      2. PKI Based. Adds a trust anchor. This is not x509
  o question from John Scudder: one thing that he runs into. what’s your manadory to implement algorithm.
  o Randy: currently not one.
  o This draft describes the L3LC ULPC PDU which describes what the upper layer protocol.
  o Descriptions of UPLC information for BGP