Thanks Loa.

Those are good points. One more reason for carrying IOAM data fields and IOAM-DEX data fields in PSD!

P.S. We had IOAM as well as IOM-DEX data fields carried in Post-Stack Data from the very early days of the MPLS IOAM draft due to such issues.

Thanks,

Rakesh

On Wed, Jun 26, 2024 at 4:56 PM Loa Andersson <loa@pi.nu> wrote:

Rakesh, Tnx for this, I think we are close with what you say here and
what I said in my review of draft-mb-mpls-ioam-dex
[https://datatracker.ietf.org/doc/review-mb-mpls-ioam-dex-06-rtgdir-early-andersson-2024-06-17/].

The -dex drafts would fit much nicer if we carried the AD as PSD.

I also have a comment on your bullet 2 below.

2. IOAM data field format (32-bit) does not fit well into 32-bit ISD
    LSE due to S bit (e.g., in 31-bit Format D), and it can result in
    limitations, kind of defeats the use of standard IOAM formats

    There is an additional problem in Format D also the bit zero is "taken",
    it is set to "1", to avoid confusing the it with a bSPL. You only have
    30 bits available, this does not make it easier.

    And, if you are going to transport a 22-bit sequence number that is by
    definition mutable, and can't be transported in bit 1-19. It has to in
    the 11 bit 10-22 and 24-31. Since it is 22 bits it will take 2 Format D
    LSEs and make it even more different from the nice 32-bit aligned fields
    in the original dex-draft.

    I think this is a pretty good example of "case 2" in my previous
mail, it
    possible to solve as >ISD, but becomes pretty awkward. Points to a PSD
    based solution.

/Loa

Den 2024-06-26 kl. 21:08, skrev Rakesh Gandhi:
> Hi Joel,
>
> Thanks for your review comment.
>
> Based on my understanding, some of the main motivations for adding
> IOAM direct export data fields in PSD instead of ISD would be:
>
> 1. IOAM data fields such as 32-bit Sequence Number or 32-bit
> Timestamp in an ISD LSE can lead to undesired ECMP behavior on
> nodes that use labels for ECMP hashing
> 2. IOAM data field format (32-bit) does not fit well into 32-bit ISD
> LSE due to S bit (e.g., in 31-bit Format D), and it can result in
> limitations, kind of defeats the use of standard IOAM formats
> 3. IOAM direct export supports extensibility to optionally add
> “unlimited” number of data fields. When added as ISD, node RLD
> would make it difficult to process the entire label stack
> 4. IOAM data fields for direct export (e.g., timestamp, sequence
> number, flow identifier, Namespace-ID, IOAM Option-Type, etc.)
> represent metadata (extra baggage) in the received packets
> * NPU simply exports these metadata (extra baggage) and does not
> really process them, there is not much motivation to place
> them in ISD.
>
>
> Based on the discussion, will update draft-gandhi-mpls-mna-ioam-dex-01
> <https://datatracker.ietf.org/doc/html/draft-gandhi-mpls-mna-ioam-dex-01>.
>
>
> Thanks,
> Rakesh
>
>
>
> On Wed, Jun 26, 2024 at 10:06 AM Joel Halpern <jmh@joelhalpern.com> wrote:
>
> Rakesh, I tried to understand the explanation in your draft for
> why the iOAM DEX needs to be post-stack data. I could not parse
> it. Could you write an email that explains why you think it is
> needed?
>
> Thank you,
>
> Joel
>
> On 6/26/2024 9:12 AM, Rakesh Gandhi wrote:
>> Thanks Jags.
>>
>> The PSD solution for MNA defined in draft-jags-mpls-ps-mna-hdr-03
>> is used for IOAM and IOAM DEX in draft-gandhi-mpls-mna-ioam-dex-01:
>>
>> https://datatracker.ietf.org/doc/html/draft-gandhi-mpls-mna-ioam-dex-01
>>
>>
>> IOAM and IOAM DEX use-cases for MNA are defined in
>> draft-ietf-mpls-mna-usecases-10:
>> https://datatracker.ietf.org/doc/html/draft-ietf-mpls-mna-usecases-10
>>
>> The PSD solution is useful for IOAM and IOAM-DEX, I support the
>> request for WG adoption for draft-jags-mpls-ps-mna-hdr-03 (as
>> co-author).
>>
>> P.S. I understand this email was not the WG adoption poll
>> initiated by the chairs.
>>
>> Thanks,
>> Rakesh
>>
>>
>>
>>
>>
>> On Tue, Jun 25, 2024 at 10:52 PM Jaganbabu Rajamanickam
>> (jrajaman) <jrajaman=40cisco.com@dmarc.ietf.org> wrote:
>>
>> Hello Chairs,
>>
>>     We would like to request WG adoption for
>> draft-jags-mpls-ps-mna-hdr.
>>
>>     Updated the draft with the initial review comments and the
>> latest MNA header format.
>>
>>    Welcome your review comments and suggestions.
>>
>> Thanx,
>>
>> Jags
>>
>> _______________________________________________
>> mpls mailing list -- mpls@ietf.org
>> To unsubscribe send an email to mpls-leave@ietf.org
>>
>>
>> _______________________________________________
>> mpls mailing list --mpls@ietf.org
>> To unsubscribe send an email tompls-leave@ietf.org
>
>
> _______________________________________________
> mpls mailing list -- mpls@ietf.org
> To unsubscribe send an email to mpls-leave@ietf.org

--
Loa Andersson
Senior MPLS Expert
Bronze Dragon Consulting
loa@pi.nu
loa.pi.nu.@gmail.com