Showing papers on "Legacy mode published in 1994"

Legacy systems — the achilles heel of downsizing (panel session)

Abstract: The application systems of most large organizations are deeply mired in the “sins of the past”. Typically, such legacy systems are very large (for example, 107 lines of code), geriatric (more than 10-years old), written in COBOL, and use a legacy database service (for example, IBM's IMS or no database management system at all). These application systems are often mission-critical (that is, essential to the organization's business) and must be operational at all times. They are not only inordinately expensive to maintain, but also inflexible (that is, difficult to adapt to changing business needs), and brittle (easily broken when modified for any purpose). Perhaps worse is the widespread fear that legacy information systems will, one day, break beyond repair. In summary, legacy information systems consume 90% to 95% of all available application systems resources, and most organizations have no idea how to reduce this resource drain. This state of affairs prevents organizations from moving to newer software architectures, including client/server architectures, modern DBMSs, parallel DBMSs, distributed DBMSs, and fourth-generation languages (4GLs). Techniques for modernizing such legacy systems are clearly required. This panel will address migration strategies that deal with this problem. The panelists will comment on possible strategies including “cold turkey” (rewrite legacy system from scratch), and “chicken little” (in place migration of code base). We also expect to address needed tools that would expedite progress in this area. Real-world experiences (“war stories”) will be highlighted.