Re: [mpls] Poll 2: New SPL vs re-purposed ELI for carrying Network Action Indicators

[Robert Raszuk <[e-mail]>]

Hi Greg,

Nope. It can clearly be supported by MPLS networks. But it does not need to
>> be part of the label stack. Subtle but extremely significant difference.
>>
> GIM2>> Can you point out to me the document that proposes encoding the
> IOAM Trace header and Trace Data into the label stack, supporting IOAM in
> MPLS using ISD?
>

No I can not as I did not make such claims.

To clarify when I make observations about label stack size I am
referring to current MNA framework not IOAM itself.


GIM2>> I've read the opinion that Bruno's draft can support up to eight
> MNAs.
>

Well that would clearly be a non-starter if you would assume such encoding
would be part of ELC.

To me the situation is simple and there are two options and indeed would be
great fi Bruno or other co-authors would clarify which one they envision as
applicable (my current reading of that draft is that both - as they leave
the actual entropy+slice_id encoding to the operator as local decision.
They are not mutually exclusive as long as the network element knows how to
apply them).

*Option 1 - *

ELC defines 256 action vectors and their numerical value follows (so here
you have the ability to encode 2^8 + 2^20) action chains. Additional
arguments would sit behind the label stack (in a PSD or an analogy to it).

*Option 2 - *

ELC defines one special action vector which binds a local action chain to
control a plane programmed set of actions.

Both options offer scalable and extendable frameworks. In both there is no
explicit action encoding.

Cheers,
R.