Re: [Lsr] Last Call: <draft-ietf-lsr-isis-srv6-extensions-14.txt> (IS-IS Extension to Support Segment Routing over IPv6 Dataplane) to Proposed Standard

["Les Ginsberg (ginsberg)" <[e-mail]>]

Alvaro (and everyone) - 

I think we can do better than this.

Prefix-attributes sub-TLV is necessary when a prefix is leaked between levels - and more specifically when leaked upwards in the hierarchy.
(We have the "D" bit in the TLV itself when leaked downwards.)

While I would prefer that we simplify things and simply require the sub-TLV all the time, I think we can be more generous and still be functional.

1)Prefix-attributes SHOULD be included in Locator TLV
2)Prefix-attributes MUST be included when TLV is leaked upwards in the hierarchy
3)Prefix-attributes sub-TLV MUST be included when the advertisement is "redistributed" from another protocol

Note that because the sub-TLV is not mandatory, if #2 and #3 are NOT followed, receivers will be unable to determine the correct source of the advertisement and may do the "wrong thing". And the receivers will be unable to detect the violation.

Finally, RFC 7794 was published over 5 years ago.
Vendors make their own choices as to what protocol extensions they choose to support. But given the usefulness of the information in prefix-attributes sub-TLV I would encourage implementations which do not yet support the sub-TLV to add it.

   Les


> -----Original Message-----
> From: Alvaro Retana <[e-mail]>
> Sent: Wednesday, May 12, 2021 7:17 AM
> To: Gengxuesong (Geng Xuesong) <[e-mail]>; Peter
> Psenak (ppsenak) <[e-mail]>; [e-mail]
> Cc: Van De Velde, Gunter (Nokia - BE/Antwerp)
> <[e-mail]>; draft-ietf-lsr-isis-srv6-
> [e-mail]; Les Ginsberg (ginsberg) <[e-mail]>;
> Shraddha Hegde <shraddha=[e-mail]>; [e-mail];
> [e-mail]
> Subject: Re: [Lsr] Last Call: <draft-ietf-lsr-isis-srv6-extensions-14.txt> (IS-IS
> Extension to Support Segment Routing over IPv6 Dataplane) to Proposed
> Standard
> 
> Peter:
> 
> 
> Hi!
> 
> As Xuesong suggested earlier, could you/we live with “SHOULD send”?
> The mitigating circumstance (recommend vs require) is precisely the
> lack of support.  I think your original reply to Gunter about how it
> could be hard to mandate the Flags TLV at this point is spot on.
> 
> Thanks!
> 
> Alvaro.
> 
> 
> 
> On May 12, 2021 at 4:49:58 AM, Peter Psenak wrote:
> 
> > as I said, if we want to mandate the presence of the Prefix Attribute
> > sub-TLV, we MUST ignore Locators without it. If we don't, then the MUST
> > on the originator does not mean anything.