Re: [Int-dir] Intdir telechat review of draft-ietf-detnet-pof-07

[Balázs Varga A <[e-mail]>]

Hi Pascal,

Many thanks for the review and the comments.

Reaction/clarifications per topic:
- mission completed: Thanks. :--)))
- implementation: An RFC leaves some freedom for implementers; furthermore, deployments may vary so much. 
- aggregation: Agree, aggregation is an important capability of DetNet. POF works on aggregate flows as well.
- packet sequencing, native IPv6: [IEEE8021CB] is referred from the draft. IEEE 802.1CB provides a detailed specification 
of the sequencing function. Same functionality was defined by DetNet and it can be used in IP/MPLS networks as well. The 
only difference is which header field carries the sequence number. As native IPv6 encapsulation of sequence numbers is 
pending, it is better not to refer to it from this document. (There is no RFC that could be referred to.)

Thanks & Cheers
Bala'zs

-----Original Message-----
From: Pascal Thubert via Datatracker <[e-mail]> 
Sent: Wednesday, December 20, 2023 10:39 AM
To: [e-mail]
Cc: [e-mail]; [e-mail]; [e-mail]
Subject: Intdir telechat review of draft-ietf-detnet-pof-07

Reviewer: Pascal Thubert
Review result: Ready

I am an assigned INT directorate reviewer for draft-ietf-detnet-pof-06. These comments were written primarily for the benefit of the Internet Area Directors.
Document editors and shepherd(s) should treat these comments just like they would treat comments from any other IETF contributors and resolve them along with any other Last Call comments that have been received. For more details on the INT Directorate, see https://datatracker.ietf.org/group/intdir/about/
<https://datatracker.ietf.org/group/intdir/about/>.

Based on my review, if I was on the IESG I would ballot this document as NO OBJECTION.

General comments:
-----------------
I found the document to be very clear and readable. The goal appears to be limited to an algorithm and if that's so, mission accomplished.

I regret, though, that the real implementation and deployment issues are only lightly discussed, e.g., positioning of the function (only once on the path?
close to exit?), values of delays, or recommendations to operate the feature in a silicon high speed hardware with many ports on multiple line cards.

Clearly, the POF function benefits from aggregation (less flows to look at) but will reorder the aggregate even if the packets are from separate flows.

The function makes an assumption on a packet sequencing. A reference should be provided on how this is done today. There's pending work on tagging IPv6 packets natively (as options) but that work is not progressed yet; worth mentioning?