Client-server computing is ultimate eventuality.

We now live in a world where people relish the graphical workstation. In fact, the graphical user interface (GUI) is one of the main reasons why computers have become so much fun. We also live in a world where the workstation no longer operates independently. Almost every computer is networked, interacting with other workstations internally. And the concept of networks being on the Internet - where individual users can surf, connect with other users and get mail - is right around the corner.

While networks provide tremendous flexibility and connectivity options, they are interconnected via an ethernet bus which is probably the weakest aspect of the network topology. The bandwidth, or speed at which data can be transported over this medium, is significantly slower than the bandwidth found within the various nodes era network or within the server itself. As more and more networks connect, both inter- and intra-company, it will become increasingly important to find solutions to network traffic problems.

Enter Client-Server Computing

For some time now, the industry's answer has been client-server computing. While that solution has not become a significant reality to date, it will evolve within the coming years. Client-server computing keeps a tremendous amount of traffic off the ethernet bus in a networked environment, and hence allows for all of the benefits of a graphical user interface without compromising performance or integrity.

In the traditional network, a file server generally delivers entire files from a main database to a PC. The application software itself runs on the PC, which allows data to be presented in a graphically pleasing format. However, many property management applications, such as large-scale accounting systems, introduce too much traffic between the server and the workstation for the pure network model to work efficiently.

In a multi-user model, like IBS, the server filters, sorts, processes and updates data within itself, before it delivers data to the workstation. In fact, the workstation often includes either a terminal, or a PC running DOS or windows terminal emulation software. This is ideal for large-scale applications because of the efficient data delivery, however the data may not be presented in a visually appealing manner.

The client-server model is the industry's solution to accommodate both efficiency and aesthetics. Separate software runs at the workstation or "client," while the base software continues to run on the "server" or multi-user system. For example, if a user has a 20,000-tenant database and wants to find all tenants whose rent exceeds $500 and who paid after the 10th of the month, a high-level request can be made from the workstation to the server. The server then finds and processes the data. The server sends only the selected data to the workstation, where it is presented in a graphically pleasing format. This, in the simplest terms, defines the client-server model.

People have once again become too concerned about the platform on which their network runs. The industry's challenge is to re-educate users about the adages of the 1980s: "Select software first and operating platform second." They are still under the misconception that they should choose their hardware first and software second. Whether the server is an Alpha running Windows NT, a Pentium running Novell or a Sun workstation running UNIX, people still believe that the platform they choose is going to change the way the workstation operates. It is simply not true.

The real solution is to let the workstation do what it does best, to provide windows into five or six preferred applications.