RE: [NGO] comments on draft-li-ngo-access-mib-00

[Li Yan <liyan_77 at huawei.com>]

Hi martin,
I consider that the MIB data and the NETCONF data should be separated.
Because MIBs and NETCONF data model are two different data models, they are
independent each other, their content may overlap. Therefore, <mib-get> can
only retrieve the MIB data，while <get> can only retrieve the true NETCONF
data. Otherwise, when a NETCONF client want to retrieve all NETCONF data by
a <get> operation, a <filter> parameter must be used to exclude all MIB
data.
>From the point of view of implementation, the mechanisms of retrieving the
MIB data and the NETCONF data are totally different. So I use operations
instead of the <filter> parameter to determine which mechanism of retrieving
data should be adopted.

RowStatus object should be ignored. The 'create' value of operation
attribute is treated as 'createAndGo', and the 'delete' value of operation
attribute is treated as 'destroy'.

Thanks for your comments.

Yan
> -----Original Message-----
> From: Martin Bjorklund [mailto:mbj at tail-f.com]
> Sent: Tuesday, January 09, 2007 5:54 AM
> To: ngo at ietf.org
> Subject: [NGO] comments on draft-li-ngo-access-mib-00
> 
> Hi,
> 
> I don't understand why a new rpc method <mib-get> is needed.  Why can't
> the normal <get> method be used?
> 
> <mib-set> I could possibly understand, but not the way it's defined.
> Specifically, how is the operation attribute supposed to be used in
> combination with RowStatus objects?
> 
> In general I think that these MIB request can be done using the normal
> netconf operations, as long as the XML mapping is well-defined.
> I.e. given a standard MIB, what will the XML look like (namespace URI,
> element structure, attribute uses, datatype represention).  This part
> isn't covered in the draft (just a reference to smidump).
> 
> 
> /martin
> 



_______________________________________________
NGO mailing list
NGO at ietf.org
https://www1.ietf.org/mailman/listinfo/ngo