On Tue, Apr 04, 2017 at 10:49:34AM +0000, Bogaert, Bart (Nokia - BE/Antwerp) 
wrote:
> On Tue, Apr 04, 2017 at 10:09:01AM +0000, Bogaert, Bart (Nokia - BE/Antwerp)
> wrote:
> > Hi,
> > 
> > When looking at the system time management part of this model I notice 
> > that for ntp only a config true section is foreseen.  But assume that 
> > we learn at run-time the NTP server (e.g. using DHCP), there is no 
> > config false data part to reflect what has been learned dynamically.  
> > Storing this in the config true part does not seem to be correct to me 
> > since this might get lost because of a copy-config action (in case a 
> > NC client never configured something there).  Are we not missing a
> server-list in the state data?
> > 
> 
> Yes, but the good news is that the revised datastore architecture solves
> this issue without requiring any changes to the model. ;-)
>
> [Bart Bogaert] Ok, but there is no NC server implementing this yet so this
> means that everybody needs to solve it in their way if this is needed now?
> Doesn't this raise an interop issue if different approaches are taken?

There is not standard yet to report an NTP server learned say via
DHCP.  Reporting a dynamically learned NTP server is simply not part
of the system model - so yes you can't expect interoperability for this.

I doubt that the system model will be changed to add a specific state
object for this given the direction we are moving to.

/js

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: +49 421 200 3587         Campus Ring 1 | 28759 Bremen | Germany
Fax:   +49 421 200 3103         <http://www.jacobs-university.de/>

_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod

Reply via email to