Re: [Pce] Scoping Items from draft-koldychev-pce-operational

As promised, I’m commenting into this thread as well. Picking Dhruv’s email from the thread because it best captures my feelings on the work.

As I noted in the review I just posted, there seem to be a few (small but important) clarifications and changes to the previous specs that need to be presented concisely and clearly. And there is quite a lot of text that is either discussion, or just observations from experience of what people built and tried to make work.

I would suggest that a first step might be to rearrange this document to make clear to the reader which parts are “updates” of existing RFCs (and group them together), and which parts are commentary.

As to Tom’s point: yes, we can never be sure that more issues won’t be found (although the number of interoperating implementations is growing). Correlation is often achieved successfully through the “updates” metadata, but when the number of changes gets large, we simply “obsolete” with a cleaned-up version. I guess there is a middle option where the “updating” RFC is “obsoleted” to reflect a larger collection of “updates”.

To Siva’s comment that this document “sheds light on the inter-op issues encountered at various multivendor public/private inter-op events”, I think that can be a valuable informational piece of work – it might make an interop report from the events, or it might result in an Informational RFC, or even a section on the PCE WG/community wiki. But formal updates to existing RFCs don’t, IMHO, fall into the category Siva describes. Which leads me to Andrew’s comment that the document “tweaks” the standards and that “the proposed updates to the PCEP standard are not "major" updates” and “The line between updating the standard vs clarifying the standard can be blurry in some cases.” If you are making a change to the procedures that an implementation follows, you are changing the specification and this must not be blurry – a clear and crisp statement is needed.

Hint: What is the cost of two documents compared to the price of wrangling one document that does several things?

Ciao,

Adrian

From: Pce <pce-bounces@ietf.org> On Behalf Of Dhruv Dhody
Sent: 10 November 2022 22:53
To: julien.meuric@orange.com
Cc: pce@ietf.org
Subject: Re: [Pce] Scoping Items from draft-koldychev-pce-operational

Hi, 

It is likely I might be rough on this, but wanted to put my thoughts on the list as well (I said as much in the last IETF meeting). 

My preference (as a WG participant) is for two documents -

(1) A very small standards track I-D that updates RFC 8231 with clear old/new text on what is being updated

(2) A larger informational I-D that matches the name "operational clarification" on how things works with figures and explanations  

For (1) see RFC 8786 as reference! For (2) see RFC 6007 as a clarification document for SVEC. 

IMHO this separation and clear focused I-D serve the WG well :)

We can discuss this during the WG session tomorrow! I have added it to the WG chairs slide as a discussion point! 

Thanks! 

Dhruv 

On Thu, Sep 29, 2022 at 9:37 AM <julien.meuric@orange.com <mailto:julien.meuric@orange.com> > wrote:

Dear PCE WG,

Let's follow up on the discussion started during IETF 114 about 
draft-koldychev-pce-operational [1]. The I-D currently tackles different 
issues about PCEP, some of them being informational, some other updating 
existing PCEP specifications. Among the options we discussed to proceed 
with this work, 2 remain:
1. Keep a single draft, but clearly separate the two types of content;
2. Break it up into 2 drafts.

We'd like to hear the WG's opinion whether you prefer:
a- a single standard track I-D, with both content types sharing fate 
until publication?
b- a clarification I-D on informational track + an I-D updating PCEP on 
standard track (possibly progressing at different paces)?

Please share your feedback using the PCE mailing list.

Thanks,

Dhruv & Julien

[1] https://datatracker.ietf.org/doc/draft-koldychev-pce-operational/

_______________________________________________
Pce mailing list
Pce@ietf.org <mailto:Pce@ietf.org> 
https://www.ietf.org/mailman/listinfo/pce