Folks -


Let's be careful here.

SR-MPLS has been deployed for several years, there are multiple implementations 
which have demonstrated interoperability, and clearly the correct encoding of 
the SID is a key element of that interoperability.



As a co-author, I am happy to listen to relevant feedback, but any textual 
change which has the potential to even suggest that an actual change has been 
made in encoding is clearly undesirable.



John - I note you have already acknowledged any errata (or erratum 😊) would be 
an editorial one - but given the above context and the fact that no one over 
these many years has publicly voiced any concerns argues for caution.

I am sure you have more pressing issues, but as your post has already started 
to cause waves, I would appreciate resolving this sooner rather than later.



Thanx.



   Les





> -----Original Message-----

> From: Acee Lindem <[email protected]>

> Sent: Thursday, December 7, 2023 12:44 PM

> To: John Scudder <[email protected]>

> Cc: Hannes Gredler <[email protected]>;

> [email protected]; Les Ginsberg (ginsberg) <[email protected]>;

> Clarence Filsfils (cfilsfil) <[email protected]>; [email protected];

> [email protected]; DECRAENE Bruno INNOV/NET

> <[email protected]>; [email protected]; Jeff Tantsura

> <[email protected]>; Peter Psenak (ppsenak) <[email protected]>;

> Horneffer, Martin <[email protected]>;

> [email protected]; [email protected]; [email protected];

> [email protected]; [email protected]; lsr <[email protected]>

> Subject: Re: [Lsr] Bug in RFC 8667 definition of SID/Index/Label

>

> Hi John,

>

> > On Dec 7, 2023, at 12:22, John Scudder 
> > <[email protected]<mailto:[email protected]>>

> wrote:

> >

> > Hi Hannes,

> >

> >> On Dec 7, 2023, at 4:38 AM, Hannes Gredler

> <[email protected]<mailto:[email protected]>>
>  wrote:

> >>

> >> We have used similar textblocks for the OSPFv2 and OSPFv3 SR extensions

> and I am not aware

> >> of any questions from implementators around ambiguity.

> >

> > Thanks for the pointer, I’ll take a look at those, too.

> >

> >> IMO there is clear enough language to describe proper encoding of the

> prefix-SID subTLV and

> >> I am not sure why an "erratum" is required.

> >

> > I agree that, after reconsidering the text in light of Les’s reply, it’s 
> > not a

> technical error (or ā€œbugā€ as I put it in the subject line). However, offline

> feedback from a couple of other experienced protocol implementors indicates

> to me that I’m not the only one who finds the presentation of the information

> to be unclear [1] and not as helpful as it could be to someone using the

> document as a reference instead of doing an in-depth read-through.

>

> We’ll probably never BIS these RFCs but I would agree that it would be good

> for one of the RFC authors to provide a clearer definition of the relationship

> between the L flag, V flag, TLV length, and TLV values (label, index, value).

> Since it seems a single flag indicating whether the value is an MPLS label or

> index into an MPLS label range would have sufficed, this description would

> certainly be beneficial to those new to IGP segment routing. Also, it is 
> unclear

> why an index/value ever needs to be 4 octets when the value is bounded by

> the MPLS label space itself.

>

> Thanks,

> Acee

>

>

> >

> > BTW if there’s some nuance to the quotation marks you used around

> ā€œerratumā€ I’m missing it. Errata are a normal part of our process, and erratum

> is just the singular of errata. [2]

> >

> > Thanks,

> >

> > —John

> >

> > [1] This quote doesn’t quite apply, but it’s a humorous way of illustrating

> that information can be provided without being made available as clearly as it

> could be. 
> http://hitchhikerguidetothegalaxy.blogspot.com/2006/04/beware-<http://hitchhikerguidetothegalaxy.blogspot.com/2006/04/beware-of-leopard-douglas-adams-quote.html>

> of-leopard-douglas-adams-quote.html<http://hitchhikerguidetothegalaxy.blogspot.com/2006/04/beware-of-leopard-douglas-adams-quote.html>

> >

> > [2] https://www.rfc-editor.org/errata-definitions/

> > _______________________________________________

> > Lsr mailing list

> > [email protected]<mailto:[email protected]>

> > https://www.ietf.org/mailman/listinfo/lsr


_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr

Reply via email to