On 08/28/2015 03:34 PM, Ladislav Lhotka wrote:
Hi Robert,
On 28 Aug 2015, at 04:26, Robert Varga <n...@hq.sk> wrote:
On 07/21/2015 09:44 AM, Juergen Schoenwaelder wrote:
Lada, you can't simply 'mount' a data model into a different place.
Think about paths in must or when expressions, or think about paths
contraints in leafrefs etc
In ODL we already have a language extension which does this. Semantically it
embeds the conceptual 'root' node into container/list item. That item becomes
the logical root against which all expressions in the embedded models are
evaluated against, e.g. it is its own little world, no constraints coming in or
out of it. While incoming references could be done (by just crossing the
embedding item), we have not found a use case for it yet.
Yes, I think this could work. Two questions:
1. Is it possible to graft the same module multiple times in different places,
perhaps in different revisions and with different features?
2. Can you identify a *set* of modules that are chrooted this way? For example,
can you say you want to have a common root for ietf-interfaces, ietf-ip and
ietf-system, rather than putting each module in its separate little world?
I would say the answer to both questions should be yes, otherwise it
would not work well in a heterogeneous environment and/or presence of
inter-module constraints.
From implementation perspective, we do not declare these in the model
itself -- it acts only as a marker for the client to understand that it
talks to a 'sub-device' and should look for a chrooted RFC6020
/netconf-state subtree to see what is going on in a particular chroot
instance.
Bye,
Robert
_______________________________________________
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod
