complex discovery problem

Raul Miller (rdm@tad.micro.umn.edu)
Fri, 21 Feb 1997 22:15:28 -0600 (CST)

I've got a rather large network to map, involving both internal and
external gateways. I've been frustrated, so far, using tkined and
scotty tools to map this out.

Part of my problem is that the snmp infrastructure has been very
neglected. Another part of my problem is human (person officially in
charge of network is very busy, to busy to coordinate with me, and
maybe doesn't really want to understand what needs to be done here).
Net result: I've got half a dozen snmp communities I'm aware of, and
expect to stumble on more as time progresses. Many values
(e.g. sysContact) are unconfigured.

[Oh, and I need to go out and grab mibs for a bunch of devices, but so
far I've been stalled when I try to get tkined to recognize them.]

(1) tkined doesn't have an abstraction for representing all the
interfaces of a router as one object.

(2) snmp occasionally has problems here, too.

Anyone have a favorite technique for recognizing which n of m
interfaces represent the same underlying machine? [For the case where
the machine is a router (and the routes are fairly stable) I can
perform a match based on the routing table. Of course, this doesn't
work for multi-homed hosts which don't route.]

I've seen some cases where I get different responses, from the same
ip, to something like sysDescr when I probe with different
communities. Ho hum...

Things I'd like to see in tkined.

(1) command line tool (I presume if I said "mib load sun.mib" inside
tkined session, tkined would start properly reporting the results of a
walk through enterprises).

(2) router/multi-homed host abstraction.

(3) consistent support for public mibs (why does mibtree under
tnm/examples have so much better coverage than tkined's mibtree? this
is when examples/mibtree is for a prior version of scotty -- it
doesn't explicitly say scotty2.1.5 so won't work without being
editted) and private mibs (why no cisco under snmp browser when it's
under snmp private?). I dunno -- I see a lot of repeated code doing
basic data movement (e.g. packing and unpacking) -- maybe tkined needs
some better high level abstractions in other areas, too.

Hints? Suggestions? Flames?

Thanks,

-- 
Raul
--
!! This message is brought to you via the `tkined & scotty' mailing list.
!! Please do not reply to this message to unsubscribe. To subscribe or
!! unsubscribe, send a mail message to <tkined-request@ibr.cs.tu-bs.de>.
!! See http://wwwsnmp.cs.utwente.nl/~schoenw/scotty/ for more information.