Hi Carlos,
Thanks. I think it’s those categories where in-band and out-of-band language
is used very loosely when it needn’t be used.
The proposed use of hybrid and in-packet isn’t great. I think ideally you’d
encourage use of descriptors that explicitly spell out what is meant with
respect to the path followed, treatment at a device, encapsulation, header
insertion, etc.
I think your (very good) goal is to remove/reduce poor / ambiguous use of the
in-band/out-of-band language, and that approach should help achieve that.
Tim
On 05/06/2025, 00:16, "Carlos Pignataro"<[email protected]> wrote:
Tim,
Thank you again for taking the time to re-review this document. I really
appreciate both your time and the thoughtful details in your feedback.
We (Tal, Adrian, and I) will be replying in detail — in the meantime, thanks
for this assertion:
I like that the emphasis is now on saying “don’t just use in-band and
out-of-band” as that language is usually loose and can be ambiguous.
As well as for the specifics on the remaining concerns you have.
Best,
Carlos.
On Jun 4, 2025, at 12:14 PM, Tim Chown<[email protected]> wrote:
Hi Carlos.
Apologies for the delay, too many life distractions.
The -06 draft is now better than the -04 I reviewed, in that it is shorter, and
less complex. But I still have a few concerns.
I like that the emphasis is now on saying “don’t just use in-band and
out-of-band” as that language is usually loose and can be ambiguous. I fully
support that aspect of this draft, and it is welcome. Sections 1 and the first
part of 2 are fine.
But then in 2.1 you introduce several terms in a bit of a haphazard way. I
think the document is a little Frankenstein in nature after being reworked a
few times. Not your fault, but I think it needs some fresh structure on top of
the ideas you have.
The differentiation for pathing, for packet treatment, and for active vs
passive is good, but you also muddle in hybrid and in-packet with active and
passive, and the difference between hybrid and in-packet is a bit ambiguous and
unclear. Hybrid sounds to me like hybrid active+passive, if I had to guess.
My suggestion would be to make section 2.1 a new section of its own, 3, and
within that have numbered subsections with headings something like:
3.1 OAM – general approach
3.2 OAM – traffic modification in flight
3.3 OAM – traffic path followed
3.4 OAM – traffic forwarding treatment
And then you can re-use much of your text by talking clearly in there about
OAM – general approach
Describe the use of dedicated (active) probe packets, OAM via pure traffic
observation (passive), and OAM by modifying the data stream.
Maybe ‘hybrid’ is then where active probe packets are modified in transit?
OAM – traffic modification in flight
Describe ways in which traffic can be modified, e.g.:
· Encapsulation/tunnelling
· Header-insertion (IPv6, in particular)
· Rewriting header fields
Some of these will affect packet size, some won’t. And may have
MTU/fragmentation impact, though that’s not in the scope of this draft.
Note here that this might be modification of active probes or of the data
stream traffic.
The modification may be localised to a domain, such that modifications are
undone or changed in some way on egress.
OAM – traffic path followed
Active probes may follow different paths to the data stream traffic
OAM – traffic forwarding treatment
Within a specific forwarding device, active probes may get different treatment,
e.g., QoS, to the data stream traffic.
Maybe define the term ‘data stream’ as meaning the non-OAM data passing through
the network which is the subject of OAM.
Section 2.2 then becomes section 4, and the subsequent sections are renumbered
accordingly.
A small nit in Section 2:
“that OAM can be” -> “that OAM traffic can be”
Oh, and if you want to ack me for comments, feel free.
Best wishes,
Tim
On 30/05/2025, 15:21, "Carlos Pignataro"<[email protected]> wrote:
Thank you, Tim!
We look forward to your re-review.
Best,
Carlos.
On May 29, 2025, at 10:59 AM, Tim Chown<[email protected]> wrote:
Hi Tal,
I’ll look shortly.
Tim
On 28/05/2025, 07:08, "Tal Mizrahi"<[email protected]> wrote:
Hi Tim,
A kind reminder - please let us know if the current version addresses
your comments.
Thanks,
Tal.
On Thu, May 15, 2025 at 7:09 PM Tal Mizrahi<[email protected]> wrote:
Hi Tim,
Thanks again for your review.
We have revised the draft, and we believe the current version
addresses the main comments you raised.
Link to the current draft:
https://datatracker.ietf.org/doc/draft-ietf-opsawg-oam-characterization
Diff compared to version 04 (which you previously reviewed):
https://author-tools.ietf.org/iddiff?url1=draft-ietf-opsawg-oam-characterization-04&url2=draft-ietf-opsawg-oam-characterization-06&difftype=--html
Please let us know if the current version has addressed your comments
