Talk Of Client/Server's Death Is Premature

"Rumors of my demise are greatly exaggerated!" If client/server technology could speak, these might be the first words it would utter. But if you ask IT professionals whether they use client/server technology, the most likely answer you'll receive is that client/server failed. Of course, this answer stems from the use of traditional two-tier database applications that account for some of the most inflexible systems in the enterprise.

The truth is that any distributed application is designed as either two-tier client/server or peer-to-peer. Both terms define the relationship between two components in a distributed application. In client/server, the client uses the services of a server to perform some unit of work. Peer-to-peer is essentially just a more complex version of client/server in which either component can act as the client or the server at any time. This enables interactive communication between distributed components.

The n-tiered distributed applications that characterize the Internet or "Weblication" industry are merely chains of two-tier client/server applications, with each link providing more modularity. Hence, the lesson learned from the failures of client/server applications is that the client should not be tightly bound to the server end point. That is, the early designs of those applications in which the client application called directly into the service interface got IT departments into trouble.

The problem with this architecture was that it didn't provide the flexibility for rapid changes on the server-a necessity in today's applications. Any changes to the server meant IT managers had to either redeploy the client application or risk application crashes or corrupted data. Also, the tight bindings did not allow the application to scale long term, forcing enterprises to redevelop these applications for large numbers of users. Today, we place application servers between the server and the client, allowing us to change the server architecture as needed for scalability and business processing without affecting the running client applications.

Some may see this column as introducing a trivial point, but I think it is important for people to understand that client/server technology did not fail. In fact, it is alive and well and powering the Web, which in itself is one of the most widely used client/server applications in the world. It is important to recognize the real lesson learned from the failures of the early client/server applications: Foresight in design is important to the future of the company's IT health.