[tcpm] TSVART review of I2NSF Re: Is ECN a valid TCP header flag?

[tom petch <[e-mail]>]

From: Scharf, Michael <[e-mail]>
Sent: 24 September 2021 11:25
To: tom petch; tcpm
Subject: RE: [tcpm] Is ECN a valid TCP header flag?

Thanks for the heads-up!

As already noted by Tom, there are multiple documents in the I2NSF working group.

As TSV-ART reviewer, I have pushed already back against similar issues in draft-ietf-i2nsf-capability-data-model: https://mailarchive.ietf.org/arch/msg/i2nsf/94nz77vNHDpP8DkbXb2hzuuBjRk/

<tp>
Ah yes I had not seen that review.  I agree with most but not all of the review.  I found that I had to read five I-D, one now an RFC, to see how they hang together.  Thus I see capability as higher level and not needing the detail, whereas nsf-facing and consumer-facing are lower level and need the detail of which aspects of e.g. TCP header need to be specified in the I-D.

The five I-D were reviewed by five different YANG doctors which I find unfortunate; there is so much overlap that I think that you have to read all five.  Changing one may be an improvement but may make it out of line without changing the others (and one is an RFC).  I see 'capability-data-model', now -18,  as the starting point and have worked to bring the others in line with that if only in terminology.

Tom Petch

p.s. my ESP thinks that these e-mail are junk - grrrrr.

It would probably help if more contributors to TCPM have a look at these documents in I2NSF.

Michael


> -----Original Message-----
> From: tcpm <[e-mail]> On Behalf Of t petch
> Sent: Friday, September 24, 2021 12:06 PM
> To: tcpm <[e-mail]>
> Subject: [tcpm] Is ECN a valid TCP header flag?
>
> This question came to me seeing the IESG discussion of an OPSAWG I-D
> which modelled, in part, TCP, in YANG, but did not model it the way TCPM
> does.
>
> There is a lot of this about, as in I2NSF, which has the most detailed
> model I see of TCP although not in a way that is likely to be used by
> those operating TCP.
>
> Thus draft-ietf-i2nsf-nsf-facing-interface-dm provides identities such as
>      identity tcp-flags {
> which is then used as
>
>    base tcp-flags;
>      identity cwr {
>      identity ecn {
>      identity urg {
>      identity ack {
>      identity psh {
>      identity rst {
>      identity syn {
>      identity fin {
> with descriptions and references.  My knowledge of this is limited but I
> suspect that the reference for 'ecn' should be RFC3168 and not RFC793
> but then should it be 'ece' and not 'ecn'?  I suspect that a TCP expert
> might see rather more idiosyncrasies.  Is RFC793 still the right
> reference or should it be 793bis? Both are used.
>
> Details of TCP also appear in
> draft-ietf-i2nsf-capability-data-model
> draft-ietf-i2nsf-consumer-facing-interface-dm
> draft-ietf-i2nsf-nsf-monitoring-data-model
> as do details of IP and a wide range of application and transport
> protocols; I suspect that those with an attention to detail and an
> interest in transport could have a field day here.
>
> Tom Petch
>
>
>
>
>
>
>
>
> _______________________________________________
> tcpm mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/tcpm