Re: [sfc] The SFC WG has placeddraft-wang-sfc-multi-layer-oaminstate"Candidate for WG Adoption"

[<[e-mail]>]

Hi Carlos,






I didn't quote any of your words, but if I misunderstand you in any way, I apologize to you.



I'm not wordsmith, especially in English, so if you want me to understand you exactly, pls express your points explicitly and clearly.


Please see more of my comments inline ...







原始邮件



发件人：CarlosPignataro(cpignata) <[e-mail]>
收件人：肖敏10093570;
抄送人：[e-mail] <[e-mail]>;[e-mail] <[e-mail]>;[e-mail] <[e-mail]>;
日 期 ：2018年10月27日 21:37
主 题 ：Re: [sfc]  The SFC WG has placeddraft-wang-sfc-multi-layer-oaminstate"Candidate for WG Adoption"



Xiao,
 Thank you for your quick response!

 Please find some follow-ups, using the same numbering scheme:

 1. That’s a strawman. I do not believe anyone said anything about “golden rules” — if I missed the reference please quote it, otherwise let’s stay factual. 

 However, the framework is a comprehensive document and choosing to simply ignore it is just buying future problems for the WG. That document has a Layering Models, OAM Components, OAM functions, and a Gap Analysis in sections 2-5 respectively. For example,  it would be baseline-useful to at least show which area of OAM within the framework your document is trying to solve for. Earlier versions of the multi-layer work mentioned the framework. 
[XM]   I think the alignment between the framework draft and this draft is necessary, and I encourage the co-authors of draft-wang-sfc-multi-layer-oam to think about your proposal carefully, then accept it or provide their own proposal.




2. You wrote:


To the question whether we need a new protocol for SFC active OAM, my answer is yes,




I understand your preference is to invent a new protocol because that’s what your document does. However, more importantly, you have not explained *why*. Let me ask again. What is the technical rationale for this major design choice? It could very well  be that it is needed, I’m not saying it is absolutely not. It’s hard to tell without a technical explanation. But you seem to be starting with a *solution* and reverse-engineering requirements. On the framework, Section 5 has a gap analysis. Section 6.4 has  OAM toolset applicability. 
[XM]   Actually I'm NOT co-author of this draft, so I don't understand why you say "your document". As has been indicated by the chairs, "This document has been presented at several SFC working group sessions", so I doubt that they have not explained "why".




3. Again, please do not mis quote me. I did NOT say this draft should mention IOAM. That’s covered in the framework. I was asking a technical question: you said it is complementary and I am asking if you could please expand on that and explain. Here’s  my question:
[XM]   IOAM can be considered a hybrid OAM, and this draft proposes an active OAM solution. In my opinion, both of them are necessary for SFC.




Lastly you wrote:


I'd like to see more coorperation, even if compitition always exists.




In my experience, innovation is a collaborative effort. Cooperation and collaboration are always positive and useful and one of the reasons for having these standardization venues. We should leverage this cooperation as much as we can. I agree and good  point. 




Trying to understand the context of the comment: By looking at the mailing list discussion history, and at the author, contributor, and acknowledgement lists, I assume you are referring to draft-wang-sfc-multi-layer-oam  needing more cooperation. I’d recommend looking at IOAM and the framework doc. for examples. 

[XM]   I meant the coorperation between the sets of OAM drafts, as well as the coorperation between the wg participants, such as you and Greg.


 Cooperation and collaboration reminded me of this: One important request: can this document please add an “Implementation Status” section?

[XM]   It's up to the authors of this draft. To my understanding, "implementation status" section is not a must, I didn't read such a section in many published RFCs.


 And one final comment and question:

[XM]   I hope the authors would like to explain them to you.


 It appears as if this document is basically two documents. 

 First, revisions 00-06, it was a model with text like this:
https://tools.ietf.org/html/draft-wang-sfc-multi-layer-oam-02#section-6

 6.  Gap analysis     This document tries to complement the SFC OAM framework and all the    SFC OAM functions are the same with the SFC OAM framework. 

 Then, in rev 07-onwards, it seems to have morphed into a set of Requirements trying to justify a solution:
https://tools.ietf.org/rfcdiff?url2=draft-wang-sfc-multi-layer-oam-07.txt

 And: https://tools.ietf.org/rfcdiff?url2=draft-wang-sfc-multi-layer-oam-12.txt&url1=draft-wang-sfc-multi-layer-oam-06.txt

 This is noticeable even in the discrepancy between the filename and the old vs. new title. Where’s multilayer?

 Is this an OK characterization?




Thanks again!

 
Thumb typed by Carlos Pignataro. Excuze typofraphicak errows





Best Regards,

Xiao Min

On Oct 26, 2018, at 23:39, "[e-mail]" <[e-mail]> wrote:
 
 



Hi Carlos,
 



 


Actually I appreciate the discussion initiated by you, even after the deadline, some of my points are as follow:


1. To the adoption process, I have a different view. I don't think an adopted framework document means something like golden rule, as the chair Joel has indicated in his response "If there are mismatches between this and the adopted working group document, it will be up to the working group to resolve them."


2. To the question whether we need a new protocol for SFC active OAM, my answer is yes, as to whether the new protocol models after MPLS LSP Ping or not, I personally don't care about it.


3. I agree to your point that this draft should mention IOAM. I'd like to see more coorperation, even if compitition always exists.



 


Best Regards,


Xiao Min



 


 










What is complementary about it (other than the same complementarity between IOAM and any other OAM solution)?