Re: IETF 55 EOS WG Minutes and Presentation

>>>>> Sharon Chisholm writes:

Sharon> The Oops ID has some interesting features, but how much these
Sharon> would get used in practice really depends on whether or not
Sharon> managers perceive it helps them with 3). If they are not
Sharon> convinced, then it's all for not.

For the average polling engine, I guess new PDUs are rather useless.
If this is what we are worry about, then SNMP is done and we should
all let it rest in peace.

If you write code that does some more intelligent operations (e.g.
checking the consistency of the IP/VLAN configuration on a larger
number of interconnected switches with many ports per switch, then the
efficiency of what we have today really forces you to fetch config
files or to script the CLI.

Sure, making the protocol more efficient by putting some more
intelligence into the agent does not change existing applications nor
does it solve the problem that it is hard to find out what to do with
the MIB data. (The later problem can probably be solved by giving
students boxes to play with plus some support and you will be surprise
how creative some of these young people are. ;-)

/js

-- 
Juergen Schoenwaelder    <http://www.informatik.uni-osnabrueck.de/schoenw/>