One aspect of the proliferation of client hardware for Windows terminal services (WTS) worries me. (Arguably, if that’s what is worrying me, then I need to get out more—but that’s beside the point.) I’m afraid that this proliferation might lead us back onto the upgrade path.

One of the key advantages to using WTS lies in its ability to extend the useful life of legacy hardware. Even if that 486 in the corner can’t run Windows 2000 (Win2K), it can run Windows 95, and thus support RDP to connect to Windows NT Server 4.0, Terminal Server Edition (TSE) or Win2K Server Terminal Services sessions. Add MetaFrame to the mix, and the 486 can run an OS such as DOS and yet still run high-performance Win32 applications on a terminal server with MetaFrame.

WTS frees Windows-based terminals (WBTs) from having to do all that much. WBTs are even simpler than that 486, containing only a CPU, RAM, video card, and network connectivity, as well as an OS to manage whichever display protocol links server and client. The latest WBTs also include COM and LPT ports, and perhaps a Universal Serial Bus (USB) port or two to support client-side peripheral devices. Still, though, the demands on the client are minimal. In a WTS environment, the bottleneck is either in the network or on the terminal server, not on the client.

Why, then, are we seeing a race to supply the fastest WBT to the marketplace? While I was walking the floor at iForum, I stopped at the booth of one WBT manufacturer (who shall remain nameless). Because it’s getting hard to tell some of these devices apart, I asked (as I’d asked everyone else) what distinguished his company’s terminals from the other WBTs out there. He told me with great pride that his company’s WBT had the fastest processor on the market.

"Why is that important?" I asked. He didn’t know.

I talked with the product marketing manager of another WBT manufacturer last week and mentioned this point to him. "It’s true," he admitted. "We tell you that we’ve got the best terminal on the market, the one you need, and then a year later we come out with another one and tell you that this is the terminal you need. Sometimes, there really is a reason to upgrade. There’ll be new management features, or something. But sometimes, it’s only the ‘need for speed’ factor."

Understand—I’m not saying that there’s anything wrong with having a fast processor in a WBT. If you’re buying a new box, then sure, get the fastest one out there with the features that you anticipate needing. What concerns me is that consumers will overemphasize the importance of WBT speed and fall right back into the upgrade cycle, always trying to have the latest and greatest hardware on their desks. This is a mistake, isn’t it? Which would you rather do: upgrade every 6 months to the next latest thing in WBT technology, or have a "set it and forget it" box on every client’s desk that you can forget about for the next 20 years? Get the devices you need, but stay out of the PC hardware rat race with your terminals.