Re: [mpls] A question about the benefit of not using ISD MNA

[Stewart Bryant <[e-mail]>]

Q

On 26 Feb 2024, at 17:39, Greg Mirsky <[e-mail]> wrote:



Hi Xiao Min,

thank you for remaining about the history of this work, but my question was a bit more about the technical merits of two solutions. AFAICS, the MPLS Network Action is the mechanism for enhancing MPLS capabilities (including the signaling of entropy information). I don't see any benefit of having two ways of realizing the same functionality, especially when a better solution is already in a quite advanced state of being finalized. If the WG decides to recognize an earlier experimental work that led to a better technical solution, then it can consider document track other than Standard.

Regards,

I am not sure about MNA, or for that matter this draft when it comes down to simple universal deployment.

The thing that took us down the SFL path so many years ago was that pretty much any router can count the number of times it received a particular label value as part of its existing forwarding functionality.

That does not apply to MNA and I am not it applies to this proposal.

The downside of SFL is that you need to do control plane work to expect the label, but in all PM applications you need to coordinate the endpoints to make the measurement, and that means a control plane type of action.

Stewart

Sent from my iPad

On Mon, Feb 26, 2024 at 12:35 AM <[e-mail]> wrote:

Many thanks Adrian!

Dear Greg,

Please also note that Weiqiang's email [1] reflects the authors' view.

Best Regards,

Xiao Min

[1] https://mailarchive.ietf.org/arch/msg/mpls/S4a5kbmBsDmHEcKhF0lsVO_0kSo/" target="_blank" rel="nofollow">https://mailarchive.ietf.org/arch/msg/mpls/S4a5kbmBsDmHEcKhF0lsVO_0kSo/" target="_blank" rel="nofollow">https://mailarchive.ietf.org/arch/msg/mpls/S4a5kbmBsDmHEcKhF0lsVO_0kSo/

Original

From: AdrianFarrel <[e-mail]>

To: 'Greg Mirsky' <[e-mail]>;'mpls' <[e-mail]>;

Cc: 'MPLS Working Group' <[e-mail]>;[e-mail] <[e-mail]>;

Date: 2024年02月25日 01:22

Subject: RE: [mpls] A question about the benefit of not using ISD MNA

For clarity for those following this thread, it applies to draft-ietf-mpls-inband-pm-encapsulation

 

The chairs discussed this point with the authors. There are several points to consider…

 

1. Per section 8, there are some existing and interoperable implementations.

2. If MNA was available today, it seems that it would be easier to direct the authors to use it.

3. Two of the lead authors are aware of MNA and working with you on a draft to use MNA for exactly this function (draft-cx-mpls-mna-inband-pm).

4. At some time in the future, the MNA approach could possibly supersede the approach in this draft.

 

So usual questions for the work group to think about at or around working group last call would be:

• Are there implementations?

• Is there are body of support?

• Is there anything broken in the current solution?

• Does the current solutions predicate against developing an alternative solution?

• Is this a case where the market could decide?

• Does anyone find the patents disclosed and their license terms such that an alternative solution should be developed? (Possibly worth checking whether that IPR would also apply to draft-cx-mpls-mna-inband-pm.)

 

The authors, of course, will want to respond to your question themselves.

 

Cheers,

Adrian

 

 

From: mpls <[e-mail]> On Behalf Of Greg Mirsky
Sent: 24 February 2024 01:11
To: [e-mail]; mpls <[e-mail]>; MPLS Working Group <[e-mail]>
Subject: [mpls] A question about the benefit of not using ISD MNA

 

Dear Authors,

I've read the draft and I have a question:

What benefits do you see to use the proposed in your draft solution instead of using MPLS Network Action In-Stack Data approach?

I thought about it and, frankly, I don't have an answer. Thank you for your consideration.

 

Regards,

Greg

_______________________________________________
mpls mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/mpls