NMRG/EMANICS Workshop Position Statement

Name:  Bert Wijnen
email: bwijnen@lucent.com


I have been active in Systems and Network Management for a pretty long 
time now. After having served as the lead-systems-programmer for IBM
mainframe systems (supporting all of IBM Europe, Middle East, Africa)
for a number of years, it seemed that my operational experience and
knowledge of system internals made me a good candidate to look at
the more generic Automated Operations (Dark Room operations) for
IBM EMEA wide internal systems. 

That "Automated Operations for IBM datacenters" work started around 1985
and it got me in touch with IBM T.J.Watson Research to work on the
computing systems to automate a lot of the operational/management
aspects of mainframes and the network that connected them with the
end-users (mostly on dumb terminals at the time). It also got me
exposed and involved in some of the IBM Mainframe NetView products,
which were used for both Network (mainly SNA) and Systems management.

In 1988, IBM got involved in contributing tools/software for managing 
the NSFNET, controlled from the NOC at MERIT in Ann Arbor (USA). IBM
first suggested and tried to manage the NSFNET (IP network) using a
mainframe NetView system with interfaces to (among others) SNMP. I
was instrumental in that piece of work. That exposed me to many of 
the tools/software used in the Unix world. With a team we first 
developed a new object-oriented system (DRAGONS, much like a distributed
JAVA environment) that also used SNMP to speak to the managed devices 
and protocol stacks. 

From that project I got involved in IETF in the Network Management area
and I contributed quite a lot over the years in spec-development of
SNMP and AgentX. At the same time I implemented SNMP (all versions,
both agent and manager side) as a SNMP platform that could be (and has
been) used on basically all systems being shipped by IBM. 

Later (from 1998 onwards) I contributed to the IETF in my role as
co-Area Director for the Operations and Management Area. In 2000
I switched from IBM to Lucent Technologies (unfortunately I did not
write much if any code since then).

In my role of IETF co-Area Director for the Operations and Management 
Area I had to manage the IETF volunteers on deciding between various 
technologies (SNMP vs. COPS, MIBs vs. PIBs, and others). During this time
I got exposed to many different, but also many overlapping and competing
techniques, approaches and viewpoints. The proliferation of Internet
technologies also allowed too many ideas to just ignore others and
so a lot of overlap and competition resulted. Good in the sense that we
learn from competing technologies, Bad in the sense that we also wasted
a lot of resources. And in addition, we put a lot of burden on Operators
who had to understand and use all these technologies. From what I have 
seen over the years, I am now convinced that if we have N technologies
today, that the Operators will have to deal with N+1 technologies every
time that Researchers and Standards Organizations come up with yet 
another technology/protocol/mechanism that will be "the best thing since 
sliced bread and that will obsolete all earlier technologies/protocols".

As such, my main contribution will probably be to take a devils advocate
position and challenge the need for anything new that will be presented.

Bert