I was writing on my other Blog the other day about the Apollo DomainOS system, and it got me thinking about thin clients. My experience with thin clients started back in 1995. As a consultant for a local company that had Apollo DomainOS systems and was transitioning to HPUX, I was newly introduced to something called X-Windows.
X-Windows is basically the graphical core of almost every UNIX/Linux/BSD system on the market. It’s the alternative to the Windows API on anything that doesn’t run Windows.
What X-Windows could do at the time that WIndows couldn’t was give a remote user a graphical front-end to a remote server. In other words, run a process on the server and display it on the client, although in a strange twist of wording, the X-Server was actually the process that ran on the “client” machine. This opened up a whole new realm of possibilities. A high-dollar licensed application – and some engineering-type apps for UNIX workstations were very high-dollar – could be run on a server and displayed elsewhere. So users could time-share a single license and save on costs.
The next step, and here is where the thin-client came in, was the invention of the X-terminal. The X-terminal was a pizza-box with a CPU, graphics card, memory, but no disk drive. It had just enough brains to boot up and find a server running XDMCP which would serve up a whole X-windows session from the server to the X-terminal. Any user could log in at any X-terminal – same session, same applications, same desktop, same experience. Still fairly expensive at the time, but cheaper than putting a proprietary UNIX workstation on every desk.
Next came the advent of Windows-based X-windows software. For a pretty reasonable price, you could load software on your Windows PC that would let you use it partially as an X-terminal. So now, with a single computer on your desktop, you could access all the new Windows office apps that everyone wanted so bad, and still launch an X-terminal session to access the UNIX engineering-type apps such as Mentor Graphics and Framemaker.
What really killed the thin-client idea back then, which was basically the end of the 90′s, was the rapid drop in cost and increase in speed of the desktop PC. Up until that time, UNIX workstations far outpaced PC’s for pure power for engineering apps. It wasn’t long before PC’s caught up and it was possible to run the same apps with the same performance directly on the PC.
Tomorrow I’ll hopefully continue this post until we reach today.
