Re: [Netconf] question about "event drafts"

"Eric Voit (evoit)" <evoit@cisco.com> wrote:
> Hi Martin,
> 
> There has been subteam working the four subscription drafts.  With our
> weekly meetings, there have been good enhancements in since Berlin.
> And you are correct in your note at the bottom hinting that some
> individual drafts were released in the middle of shuffling some things
> around.
> 
> I believe we have cleaned up the overlaps with the four WG drafts:
>   draft-ietf-netconf-rfc5277bis-00
>   draft-ietf-netconf-netconf-event-notifications-00
>   draft-ietf-netconf-restconf-notif-00
>   draft-ietf-netconf-yang-push-03
> 
> It is possible that a few overlaps might still exist in yang-push-03,
> but these will be gone in with v04 in the coming week or so.

Ok!  I will review this again once there is a set of documents that
are consistent with each other.

One issue that maybe will be clarified is why we need a new operation
"establish-subscription" instead of extending "create-subscription"
with additional parameters?

/martin

> 
> Other thoughts inline...
> 
> -----Original Message-----
> From Martin Bjorklund, August 31, 2016 11:40 AM
> 
> Hi,
> 
> I just watched the recorded NETCONF session from IETF 96, and read the
> minutes.  In the minutes I read that four new drafts related to event
> notifications are supposed to be adopted, after validation on the ML.
> I haven't seen any email about this adoption, and I don't understand
> which four drafts you referred to.  I can guess that three of them
> are:
> 
>   draft-gonzalez-netconf-event-notifications-00
>   draft-gonzalez-netconf-5277bis-02
>   draft-voit-netconf-restconf-notif-00
> 
> I don't think that these documents are ready for adoption.  It is not
> clear to me how they are supposed to work.  First of all, we have the
> 5277bis document, which apparently doesn't obsolete 5277.  However,
> the event-notifications draft says that it obsoletes 5277.  Also,
> there is still quite some overlap between the documents.
> 
> For example, 5277bis says:
> 
>    This document defines mechanisms that provide an asynchronous message
>    notification delivery service for the NETCONF protocol .  This is an
>    optional capability built on top of the base NETCONF definition.
> 
> And draft-gonzalez-netconf-event-notifications says:
> 
>    This document defines the support of [event-notifications] by the
>    Network Configuration protocol (NETCONF).
> 
> So it seems both drafts define how notifications are sent over
> NETCONF.
> 
> <Eric> Hopefully the new versions are eliminating these replications.
> 
> Further, the intent of draft-voit-netconf-restconf-notif-00 is not
> clear.  RESTCONF already supports notifications, and this new draft
> has:
> 
>   3.1.1.  Dynamic YANG Subscription over RESTCONF
> 
>      Dynamic Subscriptions are configured and manage Subscriptions via
>      signaling.  This signaling is transported over [restconf].  Once
>      established, streaming Event Notifications are then delivered via
>      Restconf SSE.
> 
> I don't understand what this means.
> 
> <Eric> Restconf spec has an issue where it cannot handle multiplexed
> subscriptions.  (i.e., it effectively has the same limitation as
> 5277).  So we need a new mechanism.
> 
> New draft significantly enhances the text above.  For a summary of
> what is coming, check out last week's meeting minutes at:
> https://github.com/netconf-wg/yang-push/wiki/Minutes-2016-09-14 
> Or wait until we post next week.  The text is all written, it just
> needs to be reviewed.  Unicast me if you want a preview.
> 
> 
> I think that this confusion is basically just the result of some
> reshuffling of contents, and that's fine.  But I think that the intent
> of each individual draft must be more clear, and that the text in the
> drafts actually reflects this.
> 
> <Eric>  Agree.  Hopefully we are progressing here.
> 
> Eric
> 
> 
> /martin
> 
> _______________________________________________
> Netconf mailing list
> Netconf@ietf.org
> https://www.ietf.org/mailman/listinfo/netconf
> 