Re: [mpls] Notes on draft-jags-mpls-mna-hdr-04

[Greg Mirsky <[e-mail]>]

Hi Tony,
to confirm that your explanation of the scope of the Unrecognized flag
helped me.

Regards,
Greg

On Thu, Jan 5, 2023 at 12:36 PM Tony Li <[e-mail]> wrote:

>
> Hi Greg,
>
>     0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    |                 Label                 | TC  |S|      TTL      |
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    |   Opcode    |        Data             |P|IHS|S|Res|U|O|  NASL |
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    | Opcode = 3  |              NAI              |S|  Data |  NAL  |
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    ~1|            Ancillary   Data               |S|Ancillary Data ~
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>    |1|            Ancillary   Data               |S|Ancillary Data |
>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>
>
>
>    - None, one or more Format C LSEs with Opcode 3 MAY be present in the same NAS.
>    - Each Format C LSE with the Opcode 3 MUST be followed by one or more Format D LSEs carrying associated Ancillary Data.
>
>
>
> That’s correct.
>
>
> If my understaning is correct, I have some questions about the Ancillary Data encoding:
>
>    - Can an AD associated with the particular NAI span more than a single Format D LSE?
>
>
>
> Yes, certainly.  AD length is only bounded by the NAL plus other actions
> within the same LSE.
>
>
>
>    - Can a single Format D LSE carry ADs of more than one NAI?
>
>
>
> Good question.  We have required an action to specify the amount of AD
> that it needs, but we have not specified any kind of granularity. This
> could reasonably be in bits, octets, or LSEs. If it’s in bits or octets, we
> might need to talk about data alignment.  If it’s in LSEs, we will need to
> talk about efficiency.
>
>
> And I still need a help to answer How a node skips an AD associated with the unsupported NAI?
>
>
>
> I think I addressed this in my previous reply.  Please let me know if that
> was unclear.
>
> Regards,
> Tony
>
>