Re: [Detnet] AD Review of draft-ietf-detnet-pof-04

[Roman Danyliw <[e-mail]>]

Feedback from idnits:

  Miscellaneous warnings:
  ----------------------------------------------------------------------------

  == The document doesn't use any RFC 2119 keywords, yet seems to have RFC
     2119 boilerplate text.

> -----Original Message-----
> From: Roman Danyliw
> Sent: Friday, October 27, 2023 12:36 PM
> To: [e-mail]
> Subject: AD Review of draft-ietf-detnet-pof-04
> 
> Hi!
> 
> I performed an AD review on draft-ietf-detnet-pof-04 to help load balance the
> IESG PubReq queue.  Thanks for this document.  I have the following feedback,
> largely to improve clarity.
> 
> Please review and respond to Henning Rogge's review at
> https://datatracker.ietf.org/doc/review-ietf-detnet-pof-04-rtgdir-early-rogge-
> 2023-10-23/ -- thanks Henning!
> In particular, please review his hypothetical case of seq= 10, 5, 7 and 9 per the
> algorithm in Section 4.3.
> 
> ** Section 1.
>    3 errors in-a-row is a usual error threshold and
>    can cause the application to stop (e.g., to go to a fail-safe state)
> 
> What is the basis for assessment that 3 is a common error threshold in
> applications?
> 
> ** Section 4.3. I am unsure of the framing assumed with of Detnet as it applies
> here.  Is there obvious basic assumption around the seq_num coming from a
> layer 2 or 3 protocol (e.g., it has one?  Which one would you choose?)  Please
> add text similar to Section 1 of draft-ietf-detnet-mpls-over-ip-preof-07 without
> the reference to timestamps:
> 
>    In general, the use of PREOF functions require sequencing information
>    to be included in the packets of a DetNet compound flow.  This can be
>    done by adding a sequence number or time stamp as part of DetNet
>    encapsulation.  Sequencing information is typically added once, at or
>    close to the source.
> 
> ** Section 4.3.
>    Note: the difference of sequence number in consecutive packets is
>    bounded due to the history window of the Elimination function before
>    the POF.  Therefore "<=" can be evaluated despite of the circular
>    sequence number space.
> 
> As far as I can tell find, the “history window of the elimination function” is not
> defined.
> 
> ** Section 4.3
>    *  "POFMaxDelay", which cannot be smaller than the delay difference
>       of the paths used by the flow.
> 
> Is “delay difference” than the worst-case value for the end-to-end latency?
> 
> ** Section 4.3.  Editorial.  Per Figure 3, are “fastest” and “slowest” the right
> adjectives to use for a time duration?
> -- “fastest path delay” to “shortest path delay”?
> -- “slowest path delay” to “longest path delay”?
> 
> ** Section 4.5.  Recommend repeating what was already said in Section 4.3,
> “Design of these parameters is out-of-scope in this document.”
> 
> Regards,
> Roman