Re: I-D ACTION:draft-ietf-tewg-mib-01.txt

[Juergen Schoenwaelder <schoenw@ibr.cs.tu-bs.de>]

On Tue, 16 Oct 2001, Brian Haberman wrote:

>> Kireeti, May I ask as to why you are not using the InetAddress TC
>> that is defined in RFC 2851 and updated in
>> draft-ietf-ops-rfc2851-update-01.txt?  I see no reason in this MIB
>> to create your own Internet Address identifier.

>>>>> Kireeti Kompella writes:

Kireeti> I'm happy to hear that.  Tell me how to have hop addresses
Kireeti> that are one of the following types: IPv4 addresses, IPv6
Kireeti> addresses, AS numbers (2 & 4 octets), and unnumbered
Kireeti> interface indices, and I'll be happy to reuse as much
Kireeti> existing mechanism as I can.

I agree with Brian that you should be using the InetAddress TC
definitions. It is unclear to me who an AS number can identify a "hop
address" - so its kind of hard to tell what the right solution is.
(Perhaps some text should be added to the document which better
explains the tunnel endpoint concept you are using.)

There are some more SMIv2 issues. For example, the teTunnelName should
be not-accessible. It does not make sense to make it
accessible-for-notify so that it can be included in notifications
since the value of the teTunnelName will be automatically part of the
instance identifier of the tePathName parameter. I also suggest to not
use IMPLIED. Also, some objects should probably use Unsigned32 instead
of Integer32 (assuming that e.g. negative bandwidth does not make
sense).

/js