Hi,

I am one of the authors of the path-marking draft.

Initially, the draft's primary goal was to define the TLV independently of the 
"markings" within. We then decided to have some markings because without them 
implementations would have a hard time figuring out what to do.

This will increase the scope of that document a bit, if the chairs/AD are in 
sync with that, we can go ahead.

Camilo

On 1/12/25, 11:37, "Ketan Talaulikar" <[email protected] 
<mailto:[email protected]>> wrote:


Hi All,


I have been following the discussion regarding my DISCUSS points on
draft-ietf-grow-bmp-bgp-rib-stats-16.


Regarding Discuss-1 (Route vs. Path Semantics)


I agree with the suggestion to move types 24 and 25 (Primary/Backup route
counts) from this document to draft-ietf-grow-bmp-path-marking-tlv. This
will allow the Working Group to fully deliberate the semantics of "route"
vs. "path," especially concerning Add-Paths and multipath aspects.


Assuming this is the chosen path:


1. Do Types 26, 27, and 28, which also refer to a path, also need to be
moved to the path-marking document for consistency?
2. For clarity across all BMP stats, does the document confirm that "route"
is functionally equivalent to "prefix/destination"? My interpretation of
existing stats suggests this is the case.


Addressing the removal of 24/25 (and potentially 26-28) would resolve my
discuss-1 ballot point.


Regarding Discuss-2 (Statistic Scope and Clarity)


Regarding discuss-2, I agree that establishing general guidance for BMP
evolution is a broader WG issue. Focusing solely on this document, the
following points need clarification:


a) Statistic View Clarity: Do all statistic descriptions accurately
indicate the view (Adj-RIB-In, Adj-RIB-Out, Loc-RIB)? The document is
sometimes ambiguous. Specifically, Types 22, 23, and 26 through 34 do not
explicitly state Adj-RIB-In, and Types 38, 39, and 40 do not explicitly
state Adj-RIB-Out. The descriptions should reflect the scope since the
Section 4 table is not being registered with IANA.


b) Loc-RIB Scoping: Types 26, 27, 28, 31, and 32 are applicable to the
Loc-RIB. Should new, dedicated stat types be defined for the Loc-RIB view
to maintain consistency with previous RFC conventions, or should the
existing type descriptions be updated to capture the Loc-RIB view as well
explicitly?


c) Global vs. Per-AFI/SAFI Types: Are both global and per-AFI/SAFI types
defined for all applicable statistics? Why are some types (e.g., 26, 27,
28) defined as per-AFI/SAFI only, without an accompanying global type?


Addressing these specific points and making necessary updates would resolve
my discuss-2 ballot point, which, in my view, would make the document ready
for publication.


Thanks,
Ketan


On Sun, Nov 30, 2025 at 12:09 AM Srivastava, Mukul <[email protected] 
<mailto:[email protected]>>
wrote:


> >>>> i propose that these two stat types are removed from
> draft-ietf-grow-bmp-bgp-rib-stats mainly for consistency to
> draft-ietf-grow-bmp-path-marking-tlv and to avoid dependencies among the
> two documents
> [MS] I too have suggested that in past. I feel discussions are not
> converging. I am ok to remove this.
>
> Thanks
> Mukul
>
> *From: *Paolo Lucente <[email protected] <mailto:[email protected]>>
> *Date: *Saturday, November 29, 2025 at 12:57 PM
> *To: *Ketan Talaulikar <[email protected] 
> <mailto:[email protected]>>, The IESG <[email protected] 
> <mailto:[email protected]>>
> *Cc: *[email protected] 
> <mailto:[email protected]> <
> [email protected] 
> <mailto:[email protected]>>, [email protected] 
> <mailto:[email protected]> <
> [email protected] <mailto:[email protected]>>, [email protected] 
> <mailto:[email protected]> <[email protected] <mailto:[email protected]>>,
> [email protected] 
> <mailto:[email protected]> <
> [email protected] 
> <mailto:[email protected]>>
> *Subject: *[GROW] Re: Ketan Talaulikar's Discuss on
> draft-ietf-grow-bmp-bgp-rib-stats-16: (with DISCUSS)
>
>
> Hi Ketan, Med, Authors,
>
> Following up on the two open discussion points:
>
> discuss 1) The only two defined stats that touch the concept of
> "primary" and "backup" are types 24 and 25; in
> draft-ietf-grow-bmp-path-marking-tlv path statuses are being defined --
> and there is more to it than just primary and backup. Evolving from my
> previous email, i propose that these two stat types are removed from
> draft-ietf-grow-bmp-bgp-rib-stats mainly for consistency to
> draft-ietf-grow-bmp-path-marking-tlv and to avoid dependencies among the
> two documents; instead we can define stats for all defined path status
> in the path marking document; this, i guess, would also close this
> discussion point;
>
> discuss 2) On the specific guidance point for future documents, please
> see
>
> https://urldefense.com/v3/__https://mailarchive.ietf.org/arch/msg/grow/6pqYmYyy2V7eVuNHkERiLd5qnrM/__;!!NpxR!jgm75lbeYnBHk5v45dt23ZCzxjHufZIxAs58VQNxugAThMQIUaTHIHZkoF6L-N2G-GmojqZZI9wXgLs$
>  
> <https://urldefense.com/v3/__https://mailarchive.ietf.org/arch/msg/grow/6pqYmYyy2V7eVuNHkERiLd5qnrM/__;!!NpxR!jgm75lbeYnBHk5v45dt23ZCzxjHufZIxAs58VQNxugAThMQIUaTHIHZkoF6L-N2G-GmojqZZI9wXgLs$>
>
> . Away from the greasy technical details, in short, the BMPv4 document
> would be a more suitable place than this document where to provide
> guidance and straighten a few aspects out.
>
> Paolo
>
>
> On 25/11/25 21:52, Paolo Lucente wrote:
> >
> > Hi Ketan,
> >
> > On the two discussion points:
> >
> > discuss 1) Complementing answers from Jeff: while it's not the role of
> > this document or draft-ietf-grow-bmp-path-marking-tlv to make any
> > definition (ie. route vs path, primary vs backup etc.), we have two
> > documents that speak about things with a certain degree of affinity:
> > maybe we can avoid both to use similar terminology independently; we
> > could explain the terminology in one document
> > (draft-ietf-grow-bmp-path-marking-tlv would be the place to do that,
> > IMO) and place a reference in the other and let it re-use the
> terminology.
> >
> > The immediate con that comes to mind is that we introduce a dependency
> > among a document already in IESG court over one that has still a bit of
> > mileage to do in the WG (although i think we are almost done with it).
> >
> > A further idea could be to lock the two documents up by adding a "path
> > status" field in relevant stats types defined in
> > draft-ietf-grow-bmp-bgp-rib-stats referencing the path code points
> > defined in draft-ietf-grow-bmp-path-marking-tlv; the main con i see is
> > that - guess we would agree on a static format for stats (see next
> > point) - it would break auto-parsing of stats in a BMP collector.
> >
> > discuss 2) There is a couple of points to unpack:
> >
> > BMP messages include a per-peer header where there are peer flags.
> > Turning and twisting some of these, one can say whether content of the
> > BMP message belongs to Adj-Rib-In pre/post policy, Adj-Rib-Out pre/post
> > policy, Loc-Rib. Of course one can't mix-and-match stats for different
> > vantage points as part of the same Stats message; one Stats message per
> > covered vantage point is needed -- sub-optimal but this is how BMP
> > operates today and, especially for periodic messages, maybe good enough


_______________________________________________
GROW mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to