Re: [Lime] Progress on LIME Connectionless documents

[Qin Wu <[e-mail]>]

Which one is your proposal,  you just raise the question, we have discussed two possible proposals we authors discussed before,
We compared these proposals and have selected the best proposal.

As I said, if you have better proposal ,please share to the list. Sure, we can discuss in Prague if you have concrete proposal, thanks.

-Qin
发件人: Lime [mailto:[e-mail]] 代表 Greg Mirsky
发送时间: 2017年7月7日 12:02
收件人: Qin Wu
抄送: Benoit Claise (bclaise); Carlos Pignataro (cpignata); [e-mail]
主题: Re: [Lime] Progress on LIME Connectionless documents

Hi Qin,
I've to point that this is not "new" proposal but rather proposal editors of the document, apparently, disregarded.

Regards,
Greg

On Thu, Jul 6, 2017 at 8:58 PM, Qin Wu <[e-mail]<mailto:[e-mail]>> wrote:
I am not criticizing the limitation of two proposals discussed below. My augment is two proposals have been well investigated and compared, the current proposal we took in this draft is good enough based on the below analysis.

If you have better proposal, I am happy to hear. But I am surprised you again have your late comment since the draft has been shipped.

-Qin
发件人: Greg Mirsky [mailto:[e-mail]<mailto:[e-mail]>]
发送时间: 2017年7月6日 23:35
收件人: Qin Wu
抄送: Carlos Pignataro (cpignata); Benoit Claise (bclaise); [e-mail]<mailto:[e-mail]>
主题: Re: [Lime] Progress on LIME Connectionless documents

Hi Qin,
thank you for your consideration of my comment and sharing your arguments. Indeed, I don't think that model that have only Self, Server and Client layers can present some scenarios of multi-layer OAM application in networks. And merely increasing that number to 8 or 16 layers, in my opinion, would not be much better. In YANG terms I view OAM layers, as well as their corresponding network layers, as entries on a list. The entries are indexed and the index provides the key. Entries of the OAM list, obviously, can represent connection-oriented or connectionless OAM.
Perhaps we can have side discussion in Prague. I'll try to prepare mode detailed comments to both documents before then (the last two weeks before the meeting are super-busy as always).

Regards,
Greg

On Mon, Jul 3, 2017 at 8:33 PM, Qin Wu <[e-mail]<mailto:[e-mail]>> wrote:
I think what Greg commented earlier is to make clarity on the definition of server layer, client layer, he didn’t point out the limitation of using level under oam-layer.
Maybe the train has leave the station. ☺

I think the debate is whether we use absolute relationship between tests points at different layer or use relative relationship between test points at different layer.
If we describe the relative relationship between test points using server layer, client layer and the same layer, I think we should make assumption that we know the order of the test points from source to destination, otherwise how to set level if we have more than 3 testpoints at different layer or level. I think index may implicitly indicate the order of these test points.
If we describe the absolute relationship between test points using (-8,8), or (-7,7), we also make assumption for the same protocol layer, we can not breakdown into several sub-layers. It is the downside of using absolute relationship.
Looks like both proposals have limitation, if we really want to tweak proposal, we may consider

a)       we add a protocol layer under oam-layer to stand for absolute stack layer. But technology under test-point-locations may have already indicated this.

b)       We expand the number of levels more than 8, e.g.,let’s say 16

-Qin
发件人: Lime [mailto:[e-mail]<mailto:[e-mail]>] 代表 Carlos Pignataro (cpignata)
发送时间: 2017年7月4日 3:43
收件人: Greg Mirsky
抄送: Benoit Claise (bclaise); [e-mail]<mailto:[e-mail]>
主题: Re: [Lime] Progress on LIME Connectionless documents

Hello, Greg,

The WG held real-time editathon interims in which we went comment by comment, many of those from you: https://datatracker.ietf.org/meeting/past#lime

After that, the editors have been diligently working through all the comments received, including all of yours (from WGLC and from before). There have been several iterations of document revision as all those comments were addressed and edits incorporated. As I had mentioned, addressing a comment does not necessarily imply changing the document — it can mean reply explaining why not.

I think that when you say “I've learned that my comments not taken into consideration” you are mischaracterizing the story. I witnessed the authors considering and discussing your comments, time and again. And when you further add “no one asked my opinion about the changes that went into the current version”, the editors asked the WG as a whole about the changes, we discussed them in person and in interims, and the process announces new revisions on list for interested parties to comment.

That said, and most importantly, if you feel you still have technical comments or unaddressed concerns on the current versions, I’d urge you to share those with clarity on the list. The editors and WG as a whole can review them and reply. There is no need to wait, the sooner things are addressed the better.

The goal of the WG is to produce quality and timely documents.

Thanks,

— Carlos.

On Jun 29, 2017, at 1:28 PM, Greg Mirsky <[e-mail]<mailto:[e-mail]>> wrote:

Hi Carlos,
WGLC from January 2017? Considering that last time I've shared comments with authors at the meeting in Chicago, I'd consider that that WGLC had failed to clear the bar. But if WG chairs and AD believe differently, then the drafts were in the WG LC for five months.

As I've mentioned, I've shared my comments at the meetings, on the list and conference calls. Regrettably, I've learned that my comments not taken into consideration and too many misconceptions stayed in the documents from one version to another after I've pointed them out. I consider it professional courtesy to reach directly to reviewers to confirm that proposed changes address their concerns. Sadly, no one asked my opinion about the changes that went into the current version.

As I was not given time to review the latest version I took rather quick look at the connectionless OAM document and found the following:

   The level in 'oam-layers' indicate whether related OAM test point is

   The level in oam-layers indicate whether related oam test point is in

   client layer(lower layer described in section 3.3<https://tools.ietf.org/html/draft-ietf-lime-yang-connectionless-oam-07#section-3.3>), server layer

   (upper layer described in section 3.3<https://tools.ietf.org/html/draft-ietf-lime-yang-connectionless-oam-07#section-3.3>) or the same layer as the

   current test point under Test point Locations.

I'm having hard time parsing this text and hope you can explain as you seem to be fine with it. And when I've followed on the reference to section 3.3

I've found that what I've commented on, suggested to change and authors verbally agreed to do still in place - {-1, 0, 1} model to indicate server-client layers.

But, as I've pointed many times, relationships between Test Points on different layers may be more complex and may skip some layers (Ethernet OAM with its 8 layers

in MEG may be one example). For the Test Points on the given OAM layer, notifications may be coming from different OAM layers down below. Current model,

as I read the text, does not allow such scenario.



But since I'm the only one who is, and frankly, has been commenting on these drafts throughout the time of the WG, I'll prepare thorough comments for the IETF LC.



I had privilege working and observing work of all authors and learned from them a lot, got to respect their professionalism. Cannot understand what happened this time.



Regards,

Greg

On Wed, Jun 28, 2017 at 3:02 PM, Carlos Pignataro (cpignata) <[e-mail]<mailto:[e-mail]>> wrote:
Greg,

I am astonished at your statement.

WGLCs:
https://mailarchive.ietf.org/arch/msg/lime/hcOi85Eu3dquUJZwWvx3BW_ZGbE
https://mailarchive.ietf.org/arch/msg/lime/4fRb9w-S9Z01x3CCRty7x8uhdmw

Your comments to the WGLCs:
https://mailarchive.ietf.org/arch/msg/lime/Bx6A0eqPRT4qXU8e_caXm4EElQA
https://mailarchive.ietf.org/arch/msg/lime/8_O41P0U4yy2zNfeX4jwxSZk6oU

After that the authors worked with you to incorporate your comments. Subsequent revisions prompted more comments from you, all of which were appropriately disposed of:
https://mailarchive.ietf.org/arch/msg/lime/x_ljuI0fvvNn2TSwj1BqXO6vgVo
https://mailarchive.ietf.org/arch/msg/lime/nQGL4cgWSaCWpNXiLnvgt_OqbSk

After that, the chairs gave you additional time and provided specific review to subsequent comments:
https://mailarchive.ietf.org/arch/msg/lime/oY44xRj9wm6aiqALdyRLuQ3j1BQ

A disposition to a comment means that they could have resulted in a change, or not.

IPR call coincinging with Adoption call:
https://mailarchive.ietf.org/arch/msg/lime/1ojMMWgjloC3FsE3Hp9E2JN2u00

I hope that clarifies?

Thanks,

— Carlos.

On Jun 28, 2017, at 4:47 PM, Greg Mirsky <[e-mail]<mailto:[e-mail]>> wrote:

Hi Carlos and Ron,
thank you for the update on these drafts. I haven't seen IPR poll, WGLC on any of these. Have I missed them? I do have comments I'd like to share. Or should I wait for the IETF LC?

Regards, Greg

On Jun 28, 2017 1:27 PM, "Carlos Pignataro (cpignata)" <[e-mail]<mailto:[e-mail]>> wrote:
Dear WG,

We will be progressing the LIME Connectionless documents, submitting them to our AD.

Please see the respective write-ups at:
https://datatracker.ietf.org/doc/draft-ietf-lime-yang-connectionless-oam/shepherdwriteup/
https://datatracker.ietf.org/doc/draft-ietf-lime-yang-connectionless-oam-methods/shepherdwriteup/

Thanks,

Carlos & Ron.

_______________________________________________
Lime mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/lime



—
Carlos Pignataro, [e-mail]<mailto:[e-mail]>

“Sometimes I use big words that I do not fully understand, to make myself sound more photosynthesis."