Re: Separation of protocol and information model

[Bobby Krupczak <rdk@krupczak.org>]

Hi!

>applications know about each device's information model?  Both Andy's and
>Phil's comments seem to indicate an assumption that management applications
>will be intimately familiar with the details of the protocol, but less so
>with the details of a device's information model.  

This assumption runs contrary to what many have observed in the SNMP
management community.

>We assume that a
>management application will also be intimately familiar with the details of
>the information model of each device it manages.  We don't assume that a
>management application will be able to manage a new device by simply feeding
>it an XML schema for the device.  Unless the device supports a standard
>schema, we assume it will take a good bit of coding.

I concur.  Many many applications have find understanding of the
protocol and that *only* provides them with syntactic
interoperability.  As soon as anything interesting is desired
(interesting in the management sense), a deeper understanding of the
semantics of the particular MIB specification is required.

However, keeping the semantics out of the protocol has served the SNMP
community well in that the protocol (despite its warts, and
shortcomings) has allowed a great deal of extensibility in terms of
agents, supported MIBs, etc.  This separation aint perfect, but its a
good start.

As an example, I want to use the netconf protocol and structure for
non-config management (rather than re-inventing the wheel) and would
like it to at least be syntactically interoperable with netconf
management software.  Putting the semantics in the XML schemas allows
us to re-use the protocol and XML framework.  I cant be the only
schmuck who wants to do this.

Bobby

--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>