Re: [netmod] Fwd: New Version Notification for draft-lengyel-netmod-yang-instance-data-01.txt

[Balazs Lengyel <[e-mail]>]

On 6/13/2018 5:26 PM, Juergen Schoenwaelder wrote:

On Wed, Jun 13, 2018 at 03:40:31PM +0100, Robert Wilton wrote:

5) I'm wondering whether there needs to be some sort of identifier about
what type data is held.  E.g. does it represent data that can be consumed as
part of one of the configuration datastores, or does it represent the
equivalent of operational state, or is it data for an RPC, etc.

Yes. I think I requested this before. There should be a way to indicate
which datastore instance-data belongs to.

BALAZS:
I am not against this proposal, I just don't see the use case for it.
At this point I only see 2 use cases:

1. Documentation of config=false data. For this the data store s always operational.
   
2. Initial configuration of config=true data. For this the datastore would be one of candidate or running. It is often not possible to know which.  Anyway all we want to sy, that this is data that can be(should be)  loaded into the running directly or via candidate.
   

I do not see the use case for defining config=true operational-state data?
Who would use it?
In which case would you handle data differently depending on the the datastore for config=true data?
Maybe if there will be dynamic datastores it will be more meaningful.  Could you please describe the use case you are thinking about!

I also do not understand why we have instance-data-set if the set is
limited to exactly one instance-data element. Perhaps room to find a
better name for the outer container or the anyxml (perhaps change
instance-data-set to instance-data and instance-data to simply data).
Or alternatively allow multiple instance-data portions but then meta
data would need to be associated with instance-data and not
instance-data-set, so we would actually need another layer.

BALAZS: The anydata statement could be called data instead of instance-data. I used instance-data because its more specific, and I like that, but I can change it.
I like to keep the file and the data-set separate. In my mind these are two different concepts just as a YANG module and a YANG file are separate. I use the term set because internally it can contain a multiple related pieces of data.

Also, given that this is supposed to be serialized in XML (or JSON I
assume), would it make sense to include an XML preamble in the
examples, i.e.,

BALAZS: OK, XML preamble will be added.

<?xml version="1.0" encoding="UTF-8"?>
<instance-data xmlns=...>
  <name>...</name>
  <datastore>ds:operational</datastore>
  <data>
    <yang-library xmlns="urn:ietf:params:xml:ns:yang:ietf-yang-library">
    </yang-library>
  </data>
</instance-data>

/js

-- 
Balazs Lengyel                       Ericsson Hungary Ltd.
Senior Specialist
Mobile: [phone]              email: [e-mail]