[SNMP4J] diskIOWrites - cached responses

[Eugene R. Snider]

In many cases this ambiguity is intentional. Remember that SNMP was 
originally designed and is still intended to function on dedicated 
special purpose appliances with limited resources which may or may not 
have sophisticated (or even simple) database facilities.

The SNMP architecture presumes that the manager has the horsepower 
necessary to transform the appliance data and that the NOC engineering 
team can figure out how to do that. Consequently the resolution of these 
is handled in the documentation which is interpreted by humans and used 
to implement the manager infrastructure.

In the world in which SNMP arose of course appliances were produced by 
large companies which had documentation teams so this was a very 
reasonable model.

This also meant that the simple thing in SNMP is the agent 
implementation ( to reduce platform cost ) and the only requirement of 
an appliance is basically that it deliver an instance datum  upon request.

Which, as we all know, is harder than it sounds....

Gene

Frank Fock wrote:
>> That's deliberate.
>>
>> Conceptually, the behaviour of snmpwalk is
>>    "get me all the information from the subtree
>>     rooted on the specified OID"
>>
>> If the OID is a leaf instance, then "all the information"
>> is the corresponding value for this instance.
>>   
> That it works like described above was my first idea too, but
> I could not find it documented in the snmpwalk specification
> for the OID parameter.
>
> Frank
>
>
>
>
> _______________________________________________
> SNMP4J mailing list
> SNMP4J at agentpp.org
> http://lists.agentpp.org/mailman/listinfo/snmp4j