On Wed, Dec 16, 2015 at 6:18 PM, Kent Watsen <[email protected]> wrote:
> > [as a contributor] > > Hi Andy, > > I’m struggling a bit to understand what is motivating you to ask this > question. That is, as a tool vendor, I wouldn’t think that any decision > made here would affect you immediately. My expectations are that any > impact to YANG/NETCONF/RESTCONF would be backwards compatible, such that > implementations would only opt-in when needed - a pay as you grow strategy. > But herein perhaps lies an unstated requirement, that the impact to > YANG/NETCONF/RESTCONF needs to be backwards compatible with respect to > existing deployments. Did we miss it or is it too obvious? > > I agree that supporting these requirements will be unnecessary for many > platforms. After all, we’ve gone decades without needing such visibility, > and that’s not going to change for many platforms for some time, if ever. > > You ask for objective metrics for determining solution applicability. My > thinking is to just let the market decide - is it not good enough? If I > tried to quantify it, I might say that its only useful for networking > devices (as their operational state somehow matters more?) and that it’s > only useful when there is no guarantee that the intended config will become > operational (i.e. applied) in some bounded amount of time. Just throwing > out ideas here, but I like best letting the market decide. > > There seemed to be agreement that most devices will apply intended config within 5 seconds (not counting the 'waiting for line card' corner-case). We usually see exec. times way under 1 sec., but if others are having a problem with delays, I guess they will find this work useful. > Kent > Andy > > > From: netmod <[email protected]> on behalf of Andy Bierman < > [email protected]> > Date: Wednesday, December 16, 2015 at 7:21 PM > To: Thomas Nadeau <[email protected]> > Cc: "[email protected]" <[email protected]>, " > [email protected]" <[email protected]> > Subject: Re: [netmod] NETMOD WG LC: draft-ietf-netmod-opstate-reqs-01 > > Hi, > > I have asked repeatedly for some indication of scope in these requirements. > There is an assumption all possible YANG-based platforms have intended > and applied state that can be different for a long enough interval such > that retrieving > the differences is operationally useful. > > For devices that converge in milli-seconds or even as long as 5 seconds, > I do not see the point of implementing solutions for these requirements. > I would prefer that this draft specify some sort of objective > metric for determining the solution applicability. > > > Andy > > > On Wed, Dec 16, 2015 at 4:10 PM, Nadeau Thomas <[email protected]> > wrote: > >> >> This is a WG Last Call on draft-ietf-netmod-opstate-reqs-01. >> Please post comments on this draft by Wednesday, December 30, 2015 >> at 9AM EST. >> >> Tom/Kent >> >> >> _______________________________________________ >> netmod mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/netmod >> > >
_______________________________________________ netmod mailing list [email protected] https://www.ietf.org/mailman/listinfo/netmod
