Hi Martin,

If we use separate YLs for shared-schema, then during YL update there may be a 
small time gap, where-in all the YLs may not have generated the same 
content-id, violating one of the constraints for the shared-schema mount-point. 

On a separate note, a mount-point is mentioned as shared-schema, so that the 
Client can leverage this information to store the schema only once and use it 
for parsing the data for all mount-point instances. Is there any other reason ?

With Regards,
Rohit

-----Original Message-----
From: Martin Bjorklund [mailto:[email protected]] 
Sent: 14 January 2019 18:36
To: Rohit R Ranade <[email protected]>
Cc: [email protected]; [email protected]
Subject: Re: Schema Mount Yang Library Update

Hi,

I think the only reasonable answer is that this behavior must not be dependent 
on your implementation strategy so the answer must be the same if you choose to 
use a shared YL or separate YLs.  Hence, if a mount point's YL changes, the 
notif is sent from that instance.


/martin



Rohit R Ranade <[email protected]> wrote:
> Hi Authors,
> 
> Any suggestions regarding the question in the below mail ?
> 
> With Regards,
> Rohit
> 
> From: Rohit R Ranade
> Sent: 28 December 2018 09:37
> To: [email protected]
> Subject: Schema Mount Yang Library Update
> 
> Hi All,
> 
> For the shared-schema type, the draft mentions "all instances of the 
> same mount point MUST have the same YANG library content identifier".
> 
> I think to achieve above condition, most vendors will plan to have 
> only one YANG library instance for that mount-point.
> 
> If use multiple instances for Yang library, it is possible that the 
> algorithm may generate a new content identifier for same data as per 
> below statement in Yang library 1.1 draft:
> 
> "There is no requirement that the same information always results in 
> the same "content-id" value."
> 
> 
> If use single instance of Yang library, when a YANG library update 
> happens, for which mount-point instance should a YANG library update 
> notification be sent ?
> What is the guideline for the implementers of this draft regarding 
> this point?
> 
> With Regards,
> Rohit

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

Reply via email to