Legacy systems must support key business initiatives
(Originally appeared in Computerworld magazine March 27, 2000)
Legacy systems have survived mergers, acquisitions, divestitures, re-engineering efforts, technical revolutions, industry realignment and Y2k. These systems, some dating back to the 1960s, remain the mainstay of information management capabilities, even as companies focus on e-commerce opportunities.
With the bulk of a company's information knowledge base locked up in these legacy systems, IT must interface, integrate, migrate and/or retire them before they hinder ongoing business strategies.
Some people think "legacy" is synonymous with Cobol, but there are hundreds of legacy languages, many of which (such as assembler and C) are harder to decipher than Cobol. And legacy systems aren't just restricted to the mainframe. They've expanded to include Java, XML, network environments and a host of evolving categories.
The foremost legacy systems challenge is the need to articulate their value and identify what role they'll play in future information initiatives. Legacy systems are easy to ignore, until IT is forced to confront them. Y2k made us pay attention to legacy systems, and now e-commerce and back-end integration requirements are forcing us to pay attention again. Creating an order-processing Web site is a manageable task, but ensuring that orders are posted, inventory is in stock, fulfillment is ensured, distribution is verified and payment is received requires back-end systems integration. Linking e-commerce applications to legacy systems is a challenge facing numerous industries in the business-to-consumer and business-to-business areas. And this has put the legacy systems challenge back on the IT agenda.
Addressing these challenges requires understanding legacy data and system functionality at an enterprise level and down to a granular level, so any project team can interface with, capture and reuse legacy data and business rules when needed. With a common reuse, integration and migration strategy, project teams could quickly distinguish between valuable, redundant, obsolete and irrelevant data and business rules. A framework for meeting these challenges is essential.
As a first step, organizations should create a "systems knowledge base," or map, of an enterprise-computing environment. They could then, for example, have accurate depictions of the systems, data and business rules invoked under various transactions. Analysts would be able to determine the legacy components needed to interface with, be reused under or be replaced by an e-commerce application. The systems knowledge base would reside in a commercial repository or database, be loaded and updated using commercially available analysis tools and accurately depict all production environments.
This knowledge base would include all physical systems components, business data and rule definitions and relations. Analysts could use and update the information as they plan and deploy system upgrades, migrations, integrations and e-commerce projects. While the tools for capturing and consolidating information within this knowledge base could be built using mostly commercial technology, a strategy would have to be developed to deploy this information across projects.
The legacy systems challenge must be tackled at an enterprise level because the installed base of systems and related data is too interdependent to tackle from a one-department perspective. This requires a comprehensive strategy for dealing with legacy systems across business units. Executives must craft a phased transition plan where immediate value can be gleaned from the systems knowledge base, while focusing on long-term goals for these systems.
Long-term goals include, for example, shifting to component-based development paradigms. A central architecture team must drive and monitor progress toward these goals using the systems knowledge base as an enabling tool. Applications management, e-commerce project teams and architecture planning teams should synchronize efforts under this common strategy. This will ensure that legacy systems support - and don't hinder - critical business initiatives.