[RESOLVED] Corporate Bandwidth.

[patweb]

Hello guys,

I asked for some new switches for 'Christmas' but got shot down. I am sure we have bandwidth problems due to improper allocation of server resources (putting too much on one box). I now need to 'prove' my conclusion to suffice management.

I don't have the hardware to shift services around, which may be necessary anyway. 70% of all my hardware is on switches (unfortunately, they are all uplinked via 100MB ports).

I need good evidence that my servers are taxed (other than the fact that my DS server cannot handle authentication requests to Exchange 2000, causing mail client failures, which is the current PAIN).

Thanks for any help!

[BigFred]

Hi Pat

My sympathies are with you. I don't have a direct answer save to look at CPU loading on the machines. You might also look at disk space, I've had a situation where I started to run out due to some very large log files which clogged things up.

Finally if management aren't prepared to give you a budget for your use to purchase items that you with you knowledge and experience deam necessary, then they have to live with it when everything fails.

PS you might notice that I no longer work where I previously did, no budget is part of the reason.

[patweb]

Ouch,

I was able to capture some critical evidence of high network activity using the performance monitor. At those same instances, the processor and memory resources were low. I think I have enough information, now it is about presentation.

I have two obvious courses of action-

#1- To get a single switch with a large backbone.

and

#2- To reapportion the server tasks over multiple servers.

We have one file server that is also the print server, the DNS server, the WINS Server, the DHCP server, and on the side, a Directory server. It was the Dial in server too, until I found that when clients connected to it, it registered it's dialin address in DNS, which caused some huge problems.

[Rapmaster]

If the CPU was being taxed and the memory was low... why are you concluding that the problem is network related...?

[patweb]

Oops, the Performance monitor shows resources in use. I should have said the memory and processor UTILIZATION was low. Bad me.

I found out that I only had one Global Catalog server and so, when the network activity is tacked out on that server, there was no backup plan. The server activity is still high, but now there are other servers fulfilling that function.

That doesn't exactly solve the problem, but it is a good workaround until the budget comes out.