Tips from the Windows Pros: Grassroots Networking

Despite their becoming so inexpensive and simple to install, networks are extremely complex systems "under the hood." It's hard enough to solve the problems that creep up from time to time in an existing, functional LAN, but new LANs are worse because everything is untested, and a little problem in any one part can mess up the whole thing. Where do you start looking for the problem?

The answer is an exercise in delayed gratification! It's exciting to see all the new equipment, parts, and cables all over the place, but as much as I'd like to hook it all up and see what happens when I turn on the switch for the first time, I've found that it's best to start small.

Whenever I build a new network, I put two computers side-by-side on one desk. They can be two regular computers for a peer-to-peer network, or a Windows 200x Server and a regular workstation for a Server-based network. I wire them together in the simplest possible way, usually with two short patch cables and a hub.

This technique gives me the smallest possible, least complex system to start with. It's much easier to solve a networking problem when you can see both computers' screens at the same time.

When I have these two computers completely configured and tested, I start adding components one at a time: a network printer, an Internet connection, a tape backup system, an uninterruptible power supply, and so on.

When something goes wrong during this technique, I know it must have something to do with the last component I added, and I'm not searching for a needle in a haystack.

Finally, when I have all the parts working, I take the two computers to their final locations and see whether they still work with the real-world wiring. Then I add workstations to the network one at a time. Attaching them this way is not as much fun as assembling the whole thing at once, but I've found that staying up all night diagnosing problems on a new network is even less fun.