Re: [Idr] IDR WG Charter (Re: Adoption call for draft-dawra-idr-bgpls-srv6-ext [5/2 - 5/16/2018])

[Robert Raszuk <[e-mail]>]

Hello Alvaro,

Thank you for your below note.  Few additional points to think about.

* The comment I made about the BOF was just a retrospect. By today we do
know there is need to transport various opaque to BGP  information around.
Moreover there is clear IETF people who are working on it so also the
aspect of who would continue to actively work seems clear too. In that
light I am really not sure what would be the point of organizing non WG
creation BOF. I think as you suggest routing-wg could be a good venue.

* To your point of  calling it Non-Routing Information Transport (nRIT).
Great abbreviation, but one may correctly say that link state or topology
is routing information. One may even claim that BFD or SR is also routing
information. I think we need to segregate what belongs in BGP or what
belongs in IGP a bit differently. And here I like recent suggestion from
Les - objection to add any information to a routing protocol not directly
relevant to its operation. Quote:

"The restriction was introduced because of general concern that using IGPs
to advertise information not directly relevant to the operation of the IGP
as a routing protocol is sub-optimal and negatively impacts the performance
of the primary IGP functions.

* I think as next steps we have two options (could be proceeded in parallel
... )

- look for IETF community to standardize alternative transports for opaque
information routing protocols (as starter for example we could take Open/R)

- take BGP snapshot of the protocol specifications and define a subset
required to realize distribution of various no directly relevant to IP
reachability information and ask for new port #,  maybe new name to
separate it completely from BGP-4 we know to glue the networks together.

It is up to IESG to decide if we need new WG for the above or any of the
existing WG is so idle that it could take this work on. Maybe from routing
area .. maybe from transport area.

Kind regards,
Robert.


On Wed, Jun 5, 2019 at 2:43 PM Alvaro Retana <[e-mail]> wrote:

> On June 4, 2019 at 9:46:23 AM, Robert Raszuk ([e-mail]) wrote:
>
> [Changed the subject.  Explicitly added rtgwg-chairs/rtg-ads: take a look
> at the last part of this message.]
>
> Robert:
>
> Hi!
>
> As to the 99% of BGP-LS,  that’s a question the chairs and ADs tried to
> ask the WG in 2018.   The WG did not engage on that discussion.
>
> I reread the IDR charter ,,, https://datatracker.ietf.org/wg/idr/about/
>
> I did not find anything there which would justify any of the work related
> to BGP-LS needs in this WG for example - topology or link state information
> transport.
>
> According to the datatracker, the last revision of the Charter was in
> 2010.  It still includes items such as MIBs and four-octet ASNs... I think
> it is time for an update!  This is a conversation that I’ve already had
> with the Chairs and they have the AI to kick off that discussion.  Among
> other things, more clarity is needed on the overall responsibility around
> BGP.
>
> To your point about BGP-LS; I think that a liberal interpretation of the
> Charter and the general understanding that this WG is responsible for BGP
> allows that work.  It would have been ideal if the question had come at the
> time when the work was started….
>
> It would also be ideal if the Charter (all charters) always reflected what
> a WG is working on…but sometimes the process of updating can get in the way
> of getting work done.
>
>
> [Bringing in your other point.]
>
> On June 4, 2019 at 6:30:26 AM, Robert Raszuk ([e-mail]) wrote:
>
> Even charter aside IMO by shifting IDR focus and processing 99% of BGP-LS
> related documents we are doing huge disservice to Internet and routing
> related needs without even bringing the aspect of protocol pollution.
>
> ...
>
> IMO AD should recognize the need for various non routing information
> transport and steer this towards BOF and new WG creation to define new
> protocol for it.. Even if rather from scratch such effort would reuse some
> of the BGP features. And while it was not done up front it is not too late
> now.
>
> I have also had the conversation with the Chairs about the load that
> BGP-LS represents to idr and the general topic of information transport
> (note that BGP is not the only protocol in the Routing Area being used to
> transport stuff…so it is a much wider topic).  The last time was a few
> months ago.  I don’t remember having the discussion on the list.
>
> At that time, the general consensus (between the Chairs and me) was that
> the operation of BGP-LS should be maintained in idr (it is BGP!).  I don’t
> remember any objection about having a discussion about the wider topic.  A
> non-WG forming BOF seems like the right place to do that.
>
> I would be happy to sponsor that discussion at IETF 105.  See more below...
>
>
> I think there are a couple of actions that should come out of this thread:
>
> (1) idr Charter Update.  I’ll rely on the Chairs to move this topic
> forward.  Even though the time is short, it would be ideal to have a
> focused discussion in Montreal.
>
> (2) Non-Routing Information Transport (nRIT).  This is clearly a
> non-idr-specific topic. If you (and/or someone else) wants to propose a BOF
> for Montreal, I will sponsor it — note that the deadline for BOF proposals
> is this Friday [1].  Realizing that a BOF proposal might be too rushed, we
> could also talk about this in rtgwg (their Charter clearly talks about
> being a "venue to discuss, evaluate, support and develop proposals for new
> work in the Routing Area”).  If we want to have a fruitful discussion (and
> not just statements in one direction or another), then I expect to start
> working on this right away.
>
> Thanks!
>
> Alvaro.
>
>
> [1] https://trac.tools.ietf.org/bof/trac/
> _______________________________________________
> Idr mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/idr
>