Re: [netmod] Adoption Poll: draft-rtgyangdt-netmod-module-tags-02

[Robert Wilton <[e-mail]>]

Hi,

I conditionally support WG adoption of this draft, with the condition 
being that I think that there are technical areas of this draft that 
should be changed (I've mentioned these previously).

My interpretation of the approach currently documented in the draft 
assumes that the tags are treated as a special type of configuration 
that don't quite adhere to normal configuration rules.  In particular, 
the approach assumes that the server will pre-populate the running 
configuration with the server default tags.  A client can then fetch and 
modify these and push them back to the server.  However, I think that 
this approach will both bloat the configuration and violates the 
principal that the contents of the configuration should be owned by the 
client, not the server.  And probably raise various other corner case 
conditions as well (e.g. what do you do if the server software is 
upgraded and the tags change? how do you merge the current client and 
new server defaults together?)

Instead, I think that the tags should make use of the NMDA architecture 
and be treated entirely as regular configuration:
- <operational> would always report the current set of tags in effect 
(system default values, overridden by any tags add/removed/modified via 
configuration).
- <running> would only contain any configured changes to the tags 
(additional tags could be added, or existing tags could be marked as 
being changed or deleted).
- No reset IPC is required, since it is just regular configuration.

I think that the draft should also support NMDA by augmenting YANG 
library bis.

Thanks,
Rob


On 06/02/2018 23:47, joel jaeggli wrote:
>
> Hi,
>
> This is the start of a *two* week poll on making 
> draft-rtgyangdt-netmod-module-tags-02 a working group document.
>
> https://tools.ietf.org/html/draft-rtgyangdt-netmod-module-tags-02
>
> This document was most recently discussed at IETF 100.
>
> Please send email to the list indicating "yes/support" or "no/do not 
> support".  If indicating no, please state your reservations with the 
> document.  If yes, please also feel free to provide comments you'd 
> like to see addressed once the document is a WG document.
>
> This poll ends on February 8.
>
> Thank you!
>
> Joel Jaeggli and IETF NETMOD Co-Chairs
>
>
>
> _______________________________________________
> netmod mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netmod