Re: [Gen-art] Genart last call review of draft-ietf-tictoc-1588v2-yang-09

[Linda Dunbar <[e-mail]>]

Rodney, 

Thanks for addressing my comments. 

It is not like "right" or "wrong" for using ENUM. Just when IEEE1588 adds a new type, the new data model to include the new type won't be backward compatible to your current one.  It is more like a better choice, not like your current design has any flaws. So, it is not necessary to change now. Maybe to consider for next data model specification. 

It would be great to have additional description in the YANG module especially the "default-ds" & "current-ds", the names are not self-explanatory. 

Thanks, Linda Dunbar
-----Original Message-----
From: Rodney Cummings [mailto:[e-mail]] 
Sent: Wednesday, September 05, 2018 12:26 PM
To: Linda Dunbar <[e-mail]>; [e-mail]
Cc: [e-mail]; [e-mail]; [e-mail]
Subject: RE: [Gen-art] Genart last call review of draft-ietf-tictoc-1588v2-yang-09

Hi Linda,

I'll attempt to address both of your comments, and my co-authors can chime in as needed.

Regarding use of enum...

I agree that identity would be preferable in order to allow augments and other modules to add new values. The reason this module uses enum is to explicitly disallow that sort of addition.
The reason why is that the list of number/name pairs is published in the IEEE 1588 standard, and the IEEE 1588 working group is the only entity that can add new values. The number for each name is used for IEEE 1588 messages (not only YANG management), and that's why the
1588 standard itself needs to enforce its own registration to avoid incompatibilities.

That being said, it is certainly possible that future editions of the IEEE 1588 standard will add new number/name pairs to the list. When that occurs, IEEE 1588 will revise the YANG module to align with the new 1588 edition, and that YANG revision will add new enum values according to the requirements of RFC 7950 section 11, which states:
      An "enumeration" type may have new enums added, provided the old
      enums's values do not change.  Note that inserting a new enum
      before an existing enum or reordering existing enums will result
      in new values for the existing enums, unless they have explicit
      values assigned to them.
Since IEEE 1588 assigns explicit values, and IEEE 1588 cannot change old values (i.e., cannot break existing products), this requirement is straightforward.

Therefore, I recommend retaining the current use of enum.

Regarding default-ds and current-ds...

Those specific terms are used in the IEEE 1588 standard document itself, including use of the abbreviation "ds" for "data set". 1588's data sets are used by the protocol itself (e.g. in messages), and also for management (thus in YANG). The terms date back to the early 1990's, so most 1588 implementers just "know" what these terms mean.

That being said, I agree that additional description in the YANG module would be helpful.

I would say that the original intent for these data sets from a management perspective was:
- default-ds: configuration of local 1588 data for the instance
- current-ds: state of local 1588 data for the instance

Most of the other data sets in 1588 represent information learned remotely (from received messages), configuration/state of a port, or data for optional features.

Why do I say "original intent"? Unfortunately, the IEEE 1588 document did not provide an explicit definition of "default" and "current", and the document still doesn't provide such a definition. As with any standard, when a concept is ambiguous, implementers sometimes add enhancements in unintended ways. In this case, it was possible for a 1588 implementer to add state data to default-ds, or add configuration data to current-ds. For YANG, that sort of addition might be done in a vendor-specific augment in order to represent product features that have existed for years. For example, we could potentially add "config false" to current-ds, but if we do, there might be complaints from folks who ship a product that configures current-ds, with justification based on the ambiguity in the
1588 standard document itself.

That history is the reason why we effectively dodged addition of a better description of default-ds and current-ds.

I'd appreciate your advice on this.

Rodney Cummings

> -----Original Message-----
> From: Linda Dunbar <[e-mail]>
> Sent: Tuesday, September 4, 2018 5:42 PM
> To: Linda Dunbar <[e-mail]>; [e-mail]
> Cc: [e-mail]; [e-mail]; 
> [e-mail]
> Subject: RE: [Gen-art] Genart last call review of draft-ietf-tictoc-
> 1588v2-yang-09
> 
> One more comment with the structure of the YANG Module:
> 
> The data model specified used several "enum" type, making it very 
> difficult to expand in the future.
> 
> For example, "delay-mechanism-enumeration" currently has "e2e", "p2P", 
> and "disabled". If you want to add one more value, the new data model 
> is not backward compatible.
> 
> Should consider using "identity" and use "identityref". When expand in 
> the future, data model is still backward compatible.
> 
> Linda Dunbar
> 
> -----Original Message-----
> From: Gen-art [mailto:[e-mail]] On Behalf Of Linda 
> Dunbar
> Sent: Tuesday, September 04, 2018 5:30 PM
> To: [e-mail]
> Cc: [e-mail]; [e-mail]; 
> [e-mail]
> Subject: [Gen-art] Genart last call review of 
> draft-ietf-tictoc-1588v2-
> yang-09
> 
> Reviewer: Linda Dunbar
> Review result: Almost Ready
> 
> I am the assigned Gen-ART reviewer for this draft. The General Area 
> Review Team (Gen-ART) reviews all IETF documents being processed by 
> the IESG for the IETF Chair.  Please treat these comments just like 
> any other last call comments.
> 
> For more information, please see the FAQ at
> 
> <https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__trac.ietf.org_trac_gen_wiki_GenArtfaq&d=DwIFAg&c=I_0YwoKy7z5LMTVdy
> O6YC 
> iE2uzI1jjZZuIPelcSjixA&r=WA71sf2o7Dw7CbYhFt24DPjt3lJuupswWYdnboKbZ8k&m
> =4uF
> XFzSFUJzVZFHQNy9WTW6pJuumHOJ1hyKNAe-
> hPYM&s=4idNt4twmS2wCSQ2GRIHlrdgUd9wsckQx3u9H85y1XM&e=>.
> 
> Document: draft-ietf-tictoc-1588v2-yang-??
> Reviewer: Linda Dunbar
> Review Date: 2018-09-04
> IETF LC End Date: 2018-09-07
> IESG Telechat date: Not scheduled for a telechat
> 
> Summary:
> This document specify the YANG data model for IEEE1588-2008.
> The document is written very clear. I have some questions, such as 
> What is the relationship between Current-DS and Default-DS?
> It seems to be that the "default-ds" has most of the information for 
> the clock.
> Is Current-ds simply supplement?
> 
> Major issues:
> 
> Minor issues:
> 
> Nits/editorial comments:
> 
> 
> _______________________________________________
> Gen-art mailing list
> [e-mail]
> https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__www.ietf.org_mailman_listinfo_gen-
> 2Dart&d=DwIFAg&c=I_0YwoKy7z5LMTVdyO6YCiE2uzI1jjZZuIPelcSjixA&r=WA71sf2
> o7Dw
> 7CbYhFt24DPjt3lJuupswWYdnboKbZ8k&m=4uFXFzSFUJzVZFHQNy9WTW6pJuumHOJ1hyK
> NAe- hPYM&s=9jRwKb0u3LDMczyKgP8Es7qIHh6Fil57BSmR0ZpgOy4&e=