Shaping Up The Internet
As the Web grows, information management becomes important–and objects are the key
By John R. Patrick
Issue column appeared: Sept. 16, 1996
In the past two years thousands of companies have embraced Internet technologies to create marketing or customer-service sites on the World Wide Web, or to transact business and reach new customers. Thousands more have begun implementing TCP/IP-based intranets to share information and foster collaboration among their employees. But as these applications of Internet technologies mushroom, so does the challenge of managing all this information.
One solution lies in Web Object Management, a new method of creating, managing, and hosting content on the exploding number of sites.
In 1994, IBM began studying how to manage the growth in network traffic expected when the Web became a true mass medium. IBM scientists from Great Britain, Germany, and the United States began creating a new way to structure, serve, and manage content on the Web.
A key insight–realized early on–was that as Web sites grew, creating and managing tens of thousands of individual files would become unworkable. In a typical Web site, a company might have thousands of documents with the company’s name or logo, all individual flat files tied together in a sort of endlessly branched tree. In this architecture, when prices or logos change, it’s necessary to manually go back and update every individual Web page affected.
IBM researchers realized that a better approach would be to construct pages from reusable parts called objects. Thus, a company logo, headers, charts, tables, etc., can be constructed from common parts. When one object is updated, every document that uses that object reflects the change. Web-site management suddenly becomes much simpler.
While Web Object Management is the core component of this new architectural approach to site development, additional technologies were needed to enable the concept to perform on a global basis. IBM refined the core server itself to what became known as WOMplex, a combination of systems, networks, and software for developing, deploying, and managing Net applications on a global scale.
Essential to WOMplex is Virtual IP Addressing. URLs, the network addresses for Web pages or servers, traditionally have referred to a physical location on the network. The WOMplex creates a URL that is an alias for a cluster of locations, each identified by the same URL to the end user. This “virtualizing” of the server enables load to be distributed simultaneously over many machines.
The WOMplex represents an approach to site design and management that will scale to the needs of any customer. We believe it provides the prototype for tomorrow’s Web.
John R. Patrick is VP of Internet technology at IBM.
