Re: [Netconf] Draft Charter Proposal for NETCONF WG

[Kent Watsen <[e-mail]>]

>> Even updating 7950 I would expect to be resisted given how long it took
>> to appear but for me, it is the only logical first step and not that
>> difficult one, just tedious to proof-read.
>
> I agree.
>

As a NETMOD co-chair, I plan to start a preliminary 7950bis discussion in Chicago.  The scope of such an effort would be the key topic, but hopefully factoring out all the NETCONF and XML specific parts would be near the top of the list.  I know some think that we should not waste resources on housekeeping, but I believe that periodic cleanups are necessary to keep things running smoothly...

Of course, the XML-specific parts factored out of RFC7950 would go into another NETMOD WG draft (mirroring RFC 7951) and the NETMOD WG could ensure to run both the 7950bis and draft-ietf-netmod-yang-xml drafts at the same time (for as clean a cutover as possible).  However, it would be up to the NETCONF WG to pick-up the NETCONF-specific parts factored out of RFC7950, and it would be up to the NETCONF WG to decide to do this work now, or wait to do it in parallel with the NETMOD WG (whenever that might be), or wait until after 7950bis comes out.

While it seems like the NETCONF WG could kick-off this activity now if it wanted to, said activity is nowhere near as important as updating NETCONF WG maintained documents to support the solution defined in the revised-datastores draft (namely RFCs 6241, 7895, 8040).  Without said update, there would be no way for clients to access ephemeral configuration (i2rs requirement) or operational state (open config requirement).  So, this is clearly a higher priority, though it seems that the two could be done at the same time, as is captured by work item #7 in Mehmet's charter proposal.

So, my opinion is that Mehmet's proposal is nearly spot-on with its work item #7.  Yes, it’s a "monster", but it must be done ASAP or else we fail to deliver the datastore solution in a timely manner.  The only thing I hesitate on is that I don't think the WG has enough information to decide if it's better to work on new standalone drafts or -bis of existing drafts (this addresses Andy's point).  To the extent the charter describes changes being made to specific documents, it may be better to remove this part from the charter until more information is known.

PS: the next update to the revised datastores draft will contain an Appendix section for each RFC believed to be impacted, in both the NETMOD and NETCONF WGs, along with a proposal for how that RFC might be changed or updated.  The revised datastore authors plan to request a NETCONF WG presentation timeslot to go over the impact to NETCONF WG drafts, with the intent of asking the WG how to proceed (e.g., standalone NETCONF WG drafts or -bis on existing drafts).  So, we'll be having this discussion then, can the charter wait that long?


Kent