Tom's Guide | Tom's Hardware | Tom's Games
![]() |
![]() |
![]() |
We're running some DOS-based software from a domain server. Some clients load and run faster than others. Talked to our software vendor and they seem to think it's a hardware issue (too long cabling, bad ports, etc., etc.). My questions are these:
1) What's the longest length to effectively run cat-5 ethernet cable? Some of these cables I believe to be over 100'.
2) How do we work around extended lengths of cable, if that's an issue?
3) It was a 'homemade' network job, and I suspect some of the pinouts may not be right on the cable ends. Is there a way to test cables to make sure they were installed correctly?
4) A couple of years back, the software company had someone onsite and they found cat-5 coiled up under a desk. Am I to understand this would slow line speeds?
5) We've got a 3Com router and a 3Com hub that everything's running through. Might these be bad, and how could we tell?
6) The DSL connection's coming straight into the Win2000 server, which has a firewall, then goes out to the router/hub setup. Is there a better way to do this?
I probably got a few more questions that'll pop up, but this is a start. Thanks for all your help.

1. 328feet
2. not a issue but you would use a hub or switch to extend the distance
3. lots of cable testers out there but they are not cheap. Might be best to download the correct pinout diagram and check both ends of the line. Improper pinout can cause loss of signal. Like if you uncross pairs in the pinout you can degrade performance.
4. no unused cable is not a problem. only if the combination of patch and cable run to switch exceeded 328feet would there be a problem.
5. you would look at the device logs or the functions they are controlling fail
6. Proper configuration would be to have the router be the gateway entry and all pcs would have that entry. So instead of going thru the server they would go directly to the DSL router. But perhaps there is something the server is doing with this link more then just adding overhead and slowing the link down.Now concerning "Some clients load and run faster than others." I would need to see your topology design. In other words if you have a backbone switch that the server is connected to but the workstation that is slow is connected to a 10mb hub and then the server it could be the hub slowing down your network.
How is your topology setup?

I can answer some of them,
1) You'll get different responses, but as a general rule it's best not to go over 100 meters (approx. 330ft.). I had one at 380ft one time without a (noticeable) problem.
2) Put a hub/switch on the line. In other words, instead of a 200ft line run 100ft plug it into a hub then run another 100ft.
3) The cheapest way is to plug a machine in and see if it works. You can spend some money and get a tester if you want. Look at the Fluke brand.
4) I've never heard of that. Not sayiing it isn't true just haven't heard it.
5) 3Com should have some sort of application that would let you 'see' the router. If worse comes to worse you might want to investigate SNMP.
6) I'm sure there is, but that way sounds decent to me.

![]() |
BT Broadband and connecti...
|
IPX trouble
|

This post is quite old and has been locked from receiving new replies. Please create a new posting instead.
| Ads by Google |