Mainframes and legacy systems have presented IT executives with a dilemma for much of the past decade. In the race to move IT operations to the Web in the 1990s, many companies considered dumping the aging systems altogether. Yet, these "dinosaurs" usually contained mission-critical data the companies wanted to preserve and integrate into newer Web applications, and so they survived until the dot-com implosion.
Since then, many companies haven't had the resources—or often the inclination—to replace mainframes and other older IT assets that hum steadily along and are the incarnation of decades of investment. Those systems still handle high volumes of transactions—and the data they generate—securely and reliably. This leaves IT shops searching for methods to get to the data housed in the systems so that it can be used in Web-based applications—and they'd like to do it without rewriting code. Complicating the picture is the fact that mainframes were designed as discrete single functions, creating problems for companies that need to give Web-based applications access to several different silos on back-end systems.
Now, with growing corporate interest in service-oriented architectures (SOA), companies are looking at using Web services to more easily integrate data from older back-end systems such as IBM mainframes and AS/400s, Unix/OpenVMS systems and Hewlett-Packard e3000s into newer applications.
