Re: [mpls] What percentage will carry indicator/ancillary data in the future

Hi Kireeti,
thank you for the heads up on the coming updates to the draft. Comparing
the detailed and brief description of flags it seems to me that
short descriptions of 0b01 should be replaced by slightly reworded 0b11,
and 0b10 and 0b11 ought to be swapped (or the detailed). It might look like:
00 — no PSD
01 — both hbh & e2e SHOULD PSD
10 — both hbh & e2e MUST PSD
11 — only e2e PSD

What do you think?

Regards,
Greg

On Mon, Oct 11, 2021 at 11:22 AM Kireeti Kompella <kireeti.ietf@gmail.com>
wrote:

> Here’s what I’m thinking for the next iteration of the draft.  We’ll burn
> two bits for hbh and e2e.  I propose the following encoding:
> 00 — No PSD (go back to sleep)
> 01 — if you are a transit or end node, you SHOULD look at the PSD
> 10 — if you are a transit or end node, you MUST look at the PSD (use with
> care)
> 11 — if you are a transit node, nothing interesting for you in the PSD; if
> you are an end node, look at the PSD.  (i.e., no hbh, only e2e PSD).
>
> Note that in the naive approach:
> 00 — no PSD
> 01 — only hbh PSD
> 10 — only e2e PSD
> 11 — both hbh & e2e PSD
>
> The action for 01 and 10 is pretty much the same.
>
> Thoughts?
>
> Kireeti.
>
> On Oct 11, 2021, at 10:48, Kireeti Kompella <kireeti.ietf@gmail.com>
> wrote:
>
> Hi Greg,
>
> There were separate indicators for h-by-h and e2e.  The e2e was removed
> following discussion in the DT; it will return in the next revision.
>
> Of course, the PSD carried and the processing needed will be different in
> the two cases as they carry different types of data.  Processing h-by-h
> will also have to deal with traversing a potentially large label stack (and
> occurs on many nodes), whereas processing e2e PSD will deal with a small
> label stack and just one node (the egress).
>
> :k
>
> On Oct 11, 2021, at 09:27, Greg Mirsky <gregimirsky@gmail.com> wrote:
>
> Hi Loa,
> I have a clarification question. As I understand it, FAI is to signal
> ancillary data that is processed in hop-by-hop or end-to-end manner (John
> proposed to have a separate indicator for each use case). The impact on
> transit nodes is, likely, different. Do you think we need to consider these
> cases separately?
>
> Regards,
> Greg
>
> On Sat, Oct 9, 2021 at 10:43 AM Loa Andersson <loa@pi.nu> wrote:
>
>> Design Team,
>>
>> I have one thing that I been thinking about.
>>
>> If we consider future mpls encapsulated packet, what percentage will
>> carry indicators and ancillary data?
>>
>> /Loa
>> --
>>
>> Loa Andersson                        email: loa@pi.nu
>> Senior MPLS Expert                          loa.pi.nu@gmail.com
>> Bronze Dragon Consulting             phone: +46 739 81 21 64
>>
>> _______________________________________________
>> mpls mailing list
>> mpls@ietf.org
>> https://www.ietf.org/mailman/listinfo/mpls
>>
> _______________________________________________
> mpls mailing list
> mpls@ietf.org
> https://www.ietf.org/mailman/listinfo/mpls
>
>
>
>