clarifications of SNMPv2 related RFCs.

Bert Wijnen <wijnen@vnet.ibm.com> Fri, 31 January 1997 22:08 UTC

Received: from cnri by ietf.org id aa04729; 31 Jan 97 17:08 EST
Received: from portal.ex.tis.com by CNRI.Reston.VA.US id aa21797; 31 Jan 97 17:08 EST
Received: (from majordom@localhost) by portal.ex.tis.com (8.8.2/8.8.2) id QAA28943 for snmpv2-outgoing; Fri, 31 Jan 1997 16:53:20 -0500 (EST)
Message-Id: <199701312158.QAA06231@relay.hq.tis.com>
Date: Fri, 31 Jan 1997 22:57:47 -0000
From: Bert Wijnen <wijnen@vnet.ibm.com>
To: snmpv2@tis.com
Subject: clarifications of SNMPv2 related RFCs.
Sender: owner-snmpv2@ex.tis.com
Precedence: bulk

Ref:  Your note of Fri, 31 Jan 1997 00:54:02 -0800 (PST)

Subject: clarifications of SNMPv2 related RFCs.

Keith answer my posting:
>
> > I then thought abot the following strange thing that I do not yet
> > grasp.
> >
> > In RFC1903 I see the TCs for TDomain and TAddress
> >
> > In RFC1906 I see snmpUDPDomain and snmpUDPAddress and others.
> >
> > So if I have a MIB I would like to be able to define 2 columns
> > in a table, one of type TDomain and one of TAddress, like:
> >
> >    myTdomain  OBJECT-TYPE
> >        SYNTAX   TDomain
> >        ..etc..
> >
> >    myTaddress OBJECT-TYPE
> >        SYNTAX   TAddress
> >        ..etc..
> >
> > If an instance of myTdomain takes for instance the value of
> > snmpUDPDomain, then the same instance of myTaddress takes the
> > value of an octet string formatted according to the snmpUDPAddress.
> > Sofar So good. But the question then is.... what is the use of
> > snmpUDPAddress TC?? Should it not be based on TAddress with a
> > refinement? And even then.... how do you handle that programmatic?
>
> SnmpUDPAddress was used in 1906 as a formal way to define the format,
> and as a specific TC for use when an address for a specific domain is
> required.
>
> TAddress is a general TC suitable for use when an address can be for
> one of several domains.  In the event that the domain is snmpUDPDomain,
> then the value of myTaddress is identical to the value it would have
> had if the syntax were SnmpUDPAddress.
>
Well, this is how I had interpreted it. But then when I read the
RFC text again, I wasn't sure that that was the intended use for
TAddress. That is why I had subject: "clarifications...."
I think it would help if RFC1906 added some text to the DESCRIPTION
clause of SnmpUDPaddress and others to refer to TAddress in RFC1903.
Or maybe that should be a REFERENCE.
It is also kind of strange that this is SnmpUDPDomain.... because
it is not specific to SNMP is it. It is in fact probably better named
TAddressUDPDomain. That would have made it clearer.

> I don't see the problem ?
>
Well, from a programmtic point I can see a problem that when you parese
a MIB that has TAddress TC for an object, then it is programatically
inpossible to figure out which displayHint to use from some other TC.
The TDomain will indicate snmpUDPdomain, but nowhere can I
programmatically connect that to the TC SnmpUDPDomain to find a
displayhint, can I?

Bert