Hi,

OIDs identifying scalar instances should always have a .0 at the end,
however OIDs identifying columnar instances will have the value of the
index (which identifies the row in the table) appended.
For columnar instances, the value can be a single subidentifier (if
e.g. there is a single index of INTEGER based type) or a sequence
of subidentifiers (if the index is e.g. a variable-length string or
if there are several indexes).

In your case it all depends on what enterprises.388.2.1.8.1.1 actually
is - and you will have to read the MIB to figure this out.

Hope this helps,

-- daniel
http://blogs.sun.com/jmxetc

Joel B Hansell wrote:
> Hi again,
>
> The way I read RFC 1157, instances of a MIB object should alway be
> identified by appending .0 to the object's OID. (RFC1157 section
> 3.2.6.3)
>
> I'm having issues interpreting this statement. Does it apply to all
> Snmp variable bindings sent on the wire, in all request types? For
> example, when I send a trap with some variable bindings in it, should
> the OID of every binding always end with .0?
>
> Does this only apply to Snmp v1 or also to v2c and v3?
>
> The reason I ask is that I've looked at some capture files of SNMP
> traffic from commercially available network elements, and they do not
> always follow this rule. I've seen, for example, v2c notifications
> where the first two bindings are sysUpTime.0 and snmpTrapOID.0, but
> the following 8 binds had OIDs of "enterprises.388.2.1.8.1.1" or
> similar. Am I interpreting the RFC wrong, or is this a case of a
> sloppy SNMP implementation?
>