[SunRescue] LAN Commmunication (was Sparc 10 cover plates)

Bjrn Ramqvist brt at osk.sema.se
Thu Dec 23 02:25:37 CST 1999

Greg A. Woods wrote:
> BTW, just exactly how many of you folks who are running big networks
> regularly get out the calculators (or spreadsheets! ;-) and work out
> exactly when and where you have to add in new switches and routers?  :-)

Without shouting out too much about how we work (which I can't tell), we
don't "watch" their network too closely.
I once, just when I started here, allocated an Alpha-machine just to run
a webserver along with MRTG collecting stats from all the FDDI-bridges
running here. Every FDDI-bridge (or -switch) have their own SNMP-able
module which made things alot easier. (Anyone seen Digitals DEChubs?)
So, for almost a year I've been collecting traffic data along _every_
switch or hub available every 5 minutes to satisfy my own mind.
Eventually we will implement RMON functions instead of SNMP, but that's
only a question of time and priority.

Let's say it works like this; if noone complains and "thinks it's
wrong", nothing is wrong either. If we get the slightest idea that
something may be wrong somewhere, we start looking at the traffic stats
and/or hook up the good old sniffer-machine. Works great.
We are however investigating in network monitoring, but not too much has
been done because of lack of time.

Every link between all the crossconnection rooms are fibers. everywhere.
Between every connection room, ew run FDDI, to bridge over to 10Mbit
among all the users. Every "loaded" server is running on FDDI. This was
the situation 1 year ago.
Right now we've upgraded the most demanding users (engineering and such)
to 100Mbit switched networks with Gigabit backbone down to the server
room, which there spreads out to 100Mbit again. In the long term we're
going to investigate in Gigabit for the servers and 100Mbit for all
The most recent addition was some Catalyst 3524XL's.

	/Regards, Bjorn

More information about the rescue mailing list