From the Editor - May 1995
David M. Kalman

Learning from the Internet

Everyone's talking about the Internet -- ad nauseam. What began as a quiet meeting place for researchers, scientists, and academics now resembles a sprawling suburban shopping mall, complete with tawdry storefronts, greasy loiterers, and poorly trained rent-a-cops. Whether the Internet can bear the onslaught of users, or crash in the world's greatest technological train wreck has now become a popular computer industry parlor game (see John C. Dvorak's article, "Info Overload at Your Fingertips," PC Magazine, March 28, 1995). But regardless of the outcome -- whether the Internet survives the abuse of digital voyeurs and/or the intrusion of government regulation -- it's the Internet architecture that I find so intriguing. This architecture embodies several key ideas that portend the future of business computing on networks. These ideas include:

1. Location independence -- From the user's perspective, the lines between remote and local are blurred. Assuming that you have a fast enough network, applications running on a server could be made to appear indistinguishable from those running locally.

2. Pervasiveness -- The network is everywhere. Users can plug into the wall, or connect via wireless communications.

3. Standards-based -- Information is stored and presented in a way that is compatible across a huge user population. Also, applications and tools become more interoperable as the network itself becomes the platform.

4. Redundancy/resiliency -- If a network node goes down, others can continue to operate. Fault tolerance in business computing is provided by connecting redundant, mirrored servers in various locations.

5. Scalability -- To support the addition of new applications and new information sources, you can add more low-cost servers without disrupting service.

These ideas embodied by the Internet offer interesting possibilities for the architecture of business systems. For example, I can envision an Internet-like network, which is provided and managed by a public utility. Businesses could subscribe to the network, much like they buy telephone services today. There might be multiple RJ11 jacks in your office, one being the public network connection.

Applications for this network could run on cheap, commodity servers that reside anywhere. Applications could span businesses to provide business-to-business integration. Applications could also connect vendors and consumers by putting the vendor's applications on the consumers' desktop computers and set-top terminals. Personal applications could continue to run locally, but they could also run exclusively from the network on a subscription basis. If remote application performance appears indistinguishable from local performance, why not take advantage of the improved manageability, availability, and scalability of the network?

The Internet itself may not serve as the public network of the future; however, the computing model that it represents has dramatic implications. In addition to providing a superior environment for deploying new kinds of applications, a pervasive, resilient, standards-based public network would lower computing infrastructure costs enormously. Companies could dispense with costly network development and support functions, and concentrate on building applications that serve their businesses. Today every corporation must design and wire its own network and build an internal utility company to support it. Meanwhile, the application backlog continues to grow.

Your letter must include your full name, address, daytime telephone number, and e-mail address. We ask for this information so we can verify a letter's authorship. Unless you state otherwise in your message, your Letter-to-the-Editor becomes the property of DBMS, and you grant permission for us to use it on our Letters Page along with your name, city, state, and e-mail address.