A radical change in the relationship between technology and business has taken place in the past two years — a change that promises to multiply the competitiveness and performance of companies that embrace it. Integration is the essence of this transformation: the insistence on integration of IT capabilities with business objectives.
Until quite recently, most advances in business technology were based on IT developments that arose outside the context of business needs and were then shoehorned into a business application. The impulse to develop an Internet presence for its own sake, without business purpose or goals, is among the most common and obvious examples.
Challenging technology’s mover-and-driver role, business is coming to view technological advances, instead, in pragmatic terms: How can the technology serve the business? Business is seeing technology as a way to improve performance; marketing, sales, engineering, and manufacturing departments are chasing their own technology initiatives to serve their own needs — and running away from IT and its sometimes solipsistic and counterproductive initiatives.
This sea change in the relationship between business and technology was inevitable. The needs of business are moving faster than IT departments can respond — particularly for organizations with legacy systems, where maintenance of these systems is estimated to consume 40 to 80 percent of IT expenditures, leaving scant resources for innovation and anticipation of emerging enterprise needs.
As they struggle to keep pace with the requirements of the business, IT departments find that their legacy systems are continuously losing relevance. Needs change constantly and the systems do not adapt; the gap grows wider every day. The needs of business are outstripping the ability of IT departments to serve them — and as a result, the requirements of competition often outpace the ability of firms to respond.
Brittle to Agile
For instance, in the case of legacy midrange iSeries systems, these consist of large, monolithic applications that are complex, expensive to maintain, slow to adapt to the needs of the business, written in legacy languages (such as RPG), and brittle. Their notorious brittleness results from the lack of full integration in these systems. New and old applications are merely connected but unintegrated. Development efforts (within the company or among trading partners) often lack full integration. What business requires and what these systems can deliver remain largely unintegrated. Hence most IT systems that are built around legacy iSeries machines and applications are prone to breakage at connecting points, are expensive to maintain, and are inadequate to the needs of the business.
Nevertheless, legacy systems, legacy applications, and legacy development tools continue to dominate business — as many as 90 percent of businesses depend on them to perform foundational back office operations. These companies find themselves in a compromised competitive position. In the iSeries example, the goal of getting a legacy iSeries system to a contemporary architecture is not just for the sake of having SOA or Web services because the technology is capable of it, but for the sake of meeting business needs. The goal is to “componentize” key business processes to manage change better and more quickly at lower cost. The business benefit is enhanced competitiveness and performance — supported by improved agility on the business’s part to understand, react to, and address customer and supplier needs, changes in the market, and other forces impacting the business — supported, in turn, by the agility of the IT structure. An agile IT structure must support an agile business.
The Enterprise Model
To create these service-oriented architectures, these agile environments, the business needs an enterprise model to frame its development. Currently, business is being offered two distinct alternatives for the iSeries: the Microsoft .NET and the IBM WebSphere models. The .NET model is language independent but hardware and platform dependent; WebSphere is highly language dependent but largely hardware and OS independent.
As many failures as successes accompany implementing either model, but industries are coming to terms with the expectation that it is more expensive either to succeed or to fail with a WebSphere SOA model. It takes longer to implement and requires a more complex path than the .NET model provides. IBM’s strategy for legacy continuance is to apply cosmetics to the monolithic applications which remain in the legacy system. Following IBM’s “roadmap,” to completely evolve, the organization must eventually transform the architecture by rewriting everything in Java. The .NET model, in contrast, supports any MSIL-compliant language, but this will also require substantial new programming for a legacy RPG system — although MSIL-compliant languages exist that will interoperate with legacy applications.
Regardless of the enterprise model chosen, completely rewriting legacy applications may be unrealistic. Legacy information systems often comprise several million lines of code; the enormity of the cost and time required may be prohibitive. Moreover, rewriting legacy applications entails an enormous loss of value because it means discarding along with the legacy applications all of the assets they embody. These assets include the “tribal knowledge” built into them over the years — which can never be adequately replaced — as well as users’ and programmers’ familiarity with the existing applications. The assets that are lost also include the business processes themselves, which typically must be reengineered in a massive effort when legacy applications are discarded and replaced. Perhaps most costly, replacing legacy applications means replacing programming staff as well — and losing all of the experience and expertise these people have accumulated over their years with the company.
To repeat: The purpose of implementing a new architecture is to facilitate integration between business and IT initiatives. An IT initiative that squanders the value of a tremendous business asset such as a legacy system and the people associated with it is not likely to be serving the business’ real interests. What makes better sense is to leverage the value inherent in the legacy system by integrating it into a contemporary architecture, one that has capabilities to facilitate new and future business initiatives.
The sensible way to pursue the benefits of contemporary architectures is through a step function — a path that does not force the company to undertake the entire evolution in a single leap. The challenge is to extend the legacy system into the new architecture at a pace that suits the business’s appetite for speed, risk, and so forth — creating a viable continuum for legacy applications that allows them to participate in new business initiatives. These initiatives are often enabled through Web services, which can change the IT infrastructure swiftly to keep pace with needs. But the real key to maintaining the viability of legacy applications is the development and support of composite applications, which both use and integrate elements of legacy applications. Composite applications combine elements from a legacy system with new ones to achieve integration of the business’s legacy assets with new investments — as well as the potential for integration of the business with its entire supply chain. In short, the solution is extension and integration of the legacy system into new architectures.
Legacy Extension and Integration
Legacy extension and integration means enabling newly emerging business needs by unlocking the value of legacy assets — focuses on extracting and leveraging the value of the legacy portfolio. It is not about criticizing or eliminating existing architecture or systems — except to the end of IT and business agility.
It is critical that businesses pursue the continuum of legacy extension and integration. Fast-changing needs require businesses to create dynamically adaptable processes and functions, which must be able to support business models that did not exist when legacy applications were developed. The benefits of legacy extension are obvious: to integrate key applications by means of composite applications, as well as through technology initiatives such as handheld devices and e-mail; to improve communications between customers, employees, and trading partners with Web services; to build functionality to reflect new business models, and so forth.
Legacy evolution attains these benefits while retaining the baseline benefits of legacy applications. It also allows the business to integrate and leverage development team resources, by, for example, putting VB/C# programmers and RPG programmers on the same team or by distilling development efforts to a common platform.
The goal of legacy extension is to unlock, preserve, extend, and leverage the value that persists in legacy assets: to take what is best from it, combine that with newer technologies, and create evolved systems capable of responding in real time to changes in the business environment. Extension and integration fit into a philosophy of legacy evolution that recognizes in IT there is no such thing as an end state — that business needs and IT tools and capabilities will continue to change. It also recognizes that enormous value persists in legacy systems and this value can be extended into an evolved system.
Legacy extension aims to unlock, preserve, extend, and leverage the entirety of the value invested in the “four pillars” of the legacy portfolio: the people, applications, data, and business processes. The ultimate goal is legacy portfolio transformation: the transformation of all legacy assets into an agile evolved system.
The Four Pillars
As you contemplate the evolution of your legacy systems, consider how your enterprise model will leverage or impact each of the four pillars of your legacy portfolio. Your business applications reflect your business. They are often complex and comprehensive — reflecting their maturity — as well as monolithic and unwieldy, if they are legacy applications. Your conundrum is how to continue to use what you must, to evolve what is necessary to support the business, and to replace what simply does not reflect business needs.
Your business processes should be continuously evolving and should facilitate your company’s agility. They must contribute to superior business performance. Many of today’s business processes did not exist when legacy applications were implemented — you will rely on IT to enable existing and new processes.
Your business data does not belong to a system; they are a business resource and asset. Integrity and security are critical — a good reason to consider persisting the iSeries as a best-of-breed database engine and application server. At the same time, data access and performance must support the reality of heterogeneous environments, and iSeries data must integrate with the rest of the business.
Finally, your people are a crucial and often overlooked pillar of your business. The people in your IT department often possess more knowledge of your business process than any other group in the company. Their “tribal knowledge” has been acquired and developed over time. It is critical that you extend the contributions and knowledge of these human assets we know as legacy application programmers, and that you integrate development efforts by integrating your development teams.
Your aim should be an evolutionary path that results in modern, best-fit, composite, and SOA-ready applications; an agile process that accommodates the pace of business; data that are accessible on multiple platforms; and the extension of your people’s skills and leveraging of their experience.
For companies with legacy iSeries systems, the integration of IT initiatives with business needs implies protection and leveraging of the value inherent in the legacy system through extension and evolution. Rather than abandoning platforms and applications, organizations will integrate them into new architectures — by using tools such as Web services and composite applications — according to the business’s requirements and risk orientation. As a result of integration at every possible level — from in-house development efforts to supply chain initiatives, to the integration of IT and business initiatives –these companies will enhance their performance and competitiveness.
Anne Ferguson co-founded ASNA (Amalgated Software of North America) where she has served as president since 1985. She spearheaded the company’s transformation from a midrange systems utilities provider to one of the world’s fastest growing Windows, Web and .NET business solution companies.