Re: [ippm] Notes on the draft-fz-ippm-alt-mark-deployment

Giuseppe Fioccola <giuseppe.fioccola@huawei.com> Thu, 05 October 2023 09:03 UTC

From: Giuseppe Fioccola <giuseppe.fioccola@huawei.com>
To: Greg Mirsky <gregimirsky@gmail.com>, IETF IPPM WG <ippm@ietf.org>, "draft-fz-ippm-alt-mark-deployment@ietf.org" <draft-fz-ippm-alt-mark-deployment@ietf.org>
Thread-Topic: Notes on the draft-fz-ippm-alt-mark-deployment
Thread-Index: AQHZ9qZARfPWUbYjM0eZRHnP7PJ54bA5myqg
Date: Thu, 05 Oct 2023 09:03:34 +0000
Message-ID: <3e1016536e4c4ca5a23942c2174e2e2f@huawei.com>
References: <CA+RyBmW0EF_ZxPjaGT59WLok=ADZ0ZALHxbq1P1ov9=5LGGpPQ@mail.gmail.com>
In-Reply-To: <CA+RyBmW0EF_ZxPjaGT59WLok=ADZ0ZALHxbq1P1ov9=5LGGpPQ@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: multipart/alternative; boundary="_000_3e1016536e4c4ca5a23942c2174e2e2fhuaweicom_"
MIME-Version: 1.0
Archived-At: <https://mailarchive.ietf.org/arch/msg/ippm/HbF0QRvVjuRmskxdS7-hYQcRl5Q>
Subject: Re: [ippm] Notes on the draft-fz-ippm-alt-mark-deployment
Precedence: list

Hi Greg,
Thank you for your thoughtful inputs.
Please see inline [GF].

Regards,

Giuseppe

From: Greg Mirsky <gregimirsky@gmail.com>
Sent: Wednesday, October 4, 2023 11:36 AM
To: IETF IPPM WG <ippm@ietf.org>; draft-fz-ippm-alt-mark-deployment@ietf.org
Subject: Notes on the draft-fz-ippm-alt-mark-deployment

Dear Authors,
thank you for taking up on this work. I believe that it is an important work, essential to the deployment of the Alternate Marking method (AMM), as an example of the on-path telemetry. I've read the draft and found that it is well-written and thoughtfully organized. I have several notes and would greatly appreciate your kind consideration:

  *   It seems like throughout the document, the AMM is considered for only packet loss and packet delay measurements. Although that is what have been demonstrated in early deployments, I think that AMM can be used for more. I imagine that an AMM profile can be defined that includes the collection of additional operational state and telemetry information on, for example, packets that distinguished with a batch of marked packets. That additional information can characterize, for example, queue utilization or trace packet through a transit node. We can add more examples in the future versions of the draft.
[GF]: This is a good point. I agree to mention additional applications such as tracing or queue utilization, which can be enabled by the Alternate-Marking.

  *   AFAICS, there were comments from the SPRING WG questioning the benefit of using SRH for the AMM in an SRv6 case instead of using the solution defined in RFC 9343. Let's see how that discussion unravels.
[GF]: Yes, it needs to be clarified. Indeed, we will update this draft according to the discussion in SPRING WG.

  *   What are your view for the AMM in an SR-MPLS network? Do yo consider that the Synonymous Flow Label (SFL) be also applicable in SR-MPLS or that AMM could use the MPLS Network Action (MNA) approach?
[GF]: I think MNA approach would be an option for SR-MPLS. Maybe, we can add a reference to the relevant document here (i.e. draft-cx-mpls-mna-inband-pm).

  *   Although there are proposal to apply AMM in SFC and Geneve (NVO3), the work might need a new place to proceed.
[GF]: I agree with you. I remember it was discussed to move it forward in IPPM, right?
Again, thank you for taking on this work!

Regards,
Greg

[ippm] Notes on the draft-fz-ippm-alt-mark-deploy… Greg Mirsky
Re: [ippm] Notes on the draft-fz-ippm-alt-mark-de… Giuseppe Fioccola