We have a client running OI and ARev3 on NT.
A consultant has been bought in to find why the network runs erratically. Out view has been that the network is unstable.
The consultant says that when he runs OI, the Windows System Monitor show 100% utilisation and this will cause the OI and network to be unstable. We feel, from experience, that this doesn't stack with out observations.
Does anyone have a reason and explanation for the 100% utilisation figure. Does it have any relevance? Does it imply a network problem?
We have run reports/processes on this network which run at wildly varying times. For example, a process was run last weekend on 3 identical PCs. They took 15, 45 and 90 minutes.
It doesn't stack up.
I've seen similar utilization figures under NT. OI is a 16bit engine and the 'thunking' eats up a lot of CPU. This would slow down other processes other than OI running on the same machine, including file I/O.
Do we need to do anything to improve the speed or is it not relevant?
Unless you're running OI and ARev on the server or a workstation that has shared resources (folders, printers, etc.), the CPU utilization should have little relevance.
I'd suspect router/switch saturation or a bad NIC/router/switch somewhere. It's also possible people are playing games on the network or making heavy use of the Internet, thus filling up the network bandwidth.
Warren,
Thanks for the information.
There was unlikely to be anyone playing games or making heavy use of the Internet.
Is there anyway of proving the router/switch is the problem?
You'd need a 'packet sniffer' to see what's going on with then network packets or hook into the diagnostic/statistical functions of the switches. I'm not heavy into the technology so I can't get more specific than that - a networking specialist can help you there.