| We will include more MIBs with future versions of the SNMP AM. However,
there will not be a concept of "one stop shopping" from DEC for MIBS.
MIBS are just too dynamic, new mibs are made available every week,
and withdrawn as fast as they are made available. Agents and mibs
do not always match. ETC. ETC.
DEC will distribute those MIBs that we have a relationship with the
vendor that allows us to pre-test a MIB against a specific version
of an agent. Those MIBs that pass this test are candidates for
inclusion on our kit. If customers need support for other mibs they
can get them from the vendors and install them. They are responsible
for the testing, etc.
At Interop we had a dictionary with over 36 MIB extensions that
we tested against. Believe me, you would not want all of them.
There were a lot of problems in that most were prototype mibs.
You will hear a lot about MIB conformance to standards over the
next few years. The industry strongly needs an "Underwriters
Labratory" seal of conformance for MIBs if SNMP is to realize
all the hype around its potential. Various agents need to
implement the protocol correctly and all of them need to follow
the MIB standard before we can start writing generic TCP oriented
FMs.
wally
|
| We should also be suggesting to customers that they request a
floppy with the latest concise mib definition be shipped with all
SNMP based products they order.
With the current multi-protocol router vendors there is a tendency
to add new features to their products, enhance the SNMP agent and
then forget to update their ASN.1 definition. If customers only
accept products which have current MIB definitons supplied with them
then there will be less pressure on management station suppliers to
provide the MIB definitions.
As the product cycle for many of the smaller vendors in now
measured in weeks, and our DECmcc product cycle is about a year, I
don't see how DEC can ever provide current MIB defns with our SSB
released kits.
There is a pressing need for a clearing house for MIBs, and DEC need
to have our compiler used in the verification process. The MOSY
compiler currently used by the Internet community allows many common
errors to pass thru without detection. Now who is going to fund
this effort????
- Mike
- Mike
|
| This problem is an opportunity for a group with an desire to make
money. NOBODY is going to fund this - but if a group invests in this,
there is money to be made. My assumption is based on an informal poll
that says customers will pay $1000 for a MIB update service. One
person would be required, full time, to collect MIBS, test them, call
up the vendors to get fixes, and distribute the 'update service' on
a CD-ROM everytime it goes out. An encryption utility to prevent
'stealing the service' might also be appropriate. There are well over
100 customers of the TCP/IP-AM today - there are over 100 customers of
MSU. If you can sell the service to 50% of today's customers base, you
have funded one person for about six months (BUP pricing), and the
customer base is growing.
bill
|