Re: Comments on SNMP/SMP Coexistence...

[Keith McCloghrie <[e-mail]>]

> Shouldnt 4.1.2 (1) note that it assumes the SMP Manager is acting in a
> bi-lingual role.
> 
> Wouldnt another alternative be for the proxy AGENT described in 4.1.2 to 
> act as a truly transparent proxy AGENT, returning compliant SMP PDUs?
 
As Marshall says, a truly transparent proxy agent would be problematic.
The beauty of the 4.1.2 procedure is that it produces unambiguous PDUs 
which are SMP-compliant (as far as the SMP manager is concerned), even 
though no SMP-compliant (non-proxy) agent would ever produce them.
This is true since noSuchName, badValue, and readOnly are retained by SMP
as enumerated values of error-status, even though none of the procedures 
in the SMP-protocol docuemnt generate them.  Thus, a SMP-only manager 
must be able to handle them.  

Would it help to have the SMP Protocol Ops document include a mention 
of when noSuchName, badValue, and readOnly might be generated and what 
they mean ?

> Is 4.1.2 describing the AGENT, or the 'as if it were a proxy agent' in 
> the MANAGER (as noted in 4.2 para 2).

As Marshall said, 4.1.2 discusses the behavior of a proxy agent.
In addition, it is, of course, one way to implement the 'as if it
were a proxy agent' advice given in 4.2. para 2.

Keith.