Moving Legacy Systems Forward

By Joe Gentry

(Back to article)

Large organizations now routinely ask their CIOs to deliver real-time business data to executives, partners, and customers — while still keeping costs in line for shareholders.


Add to that the ongoing pressure to adopt newer technologies designed for the Internet, and it’s no wonder that so many CIOs are asking if they need to rip-and-replace their legacy systems; if it's time to start over with a clean slate?


It's a good question, and one that companies around the globe are pondering. An estimated that 70% of the world’s data still resides on the mainframe. But for a host of technical and business reasons, the clean slate, "rip-and-replace" approach is rarely the right answer.


Preserve and Extend


In fact, preserve-and-extend is the faster, more cost-effective, less risky approach. Why? Because no CIO can simply walk away from the systems that drive the business.


In Internet age, the term "legacy" has taken on a negative connotation. But the truth for an enormous number of companies is legacy applications, such as purchasing, manufacturing, financial and payroll systems, form the very backbone of the business.


These systems house the data and business processes that differentiate a company from its competitors. And they represent years of accrued intellectual property.


Ripping them out and replacing them with newer systems doesn't make sense.


That said, the limitations cannot be ignored. Legacy applications are often disconnected from the enterprise, making key business data difficult to access. They house silos of data that are difficult to integrate with other silos, and they are sometimes difficult to support.


A strategy to preserve and extend legacy applications enable you to capitalize on your enterprise system’s longstanding strengths — reliability, security, and performance — while also addressing its limitations.


From Simple to Sophisticated


There are four ways to do that, ranging from quick-and-simple to sophisticated-and-flexible.


First, you can maintain the system as is — "if it ain't broke, don't fix it." If executives, partners, and customers aren't complaining, you may be able to tolerate what you have.


A second option is to make minor enhancements, such as adding a couple of fields to the database or creating a couple of new reports, to comply with executive requests. But you are essentially preserving what you have.


A third approach is to begin extending and modernizing the legacy application. You might replace the "green screen" terminal with a new, graphical user interface, or a Web browser.


The fourth (and best) option — integrating the legacy system with 25 or 30, or even 100 other applications running on the mainframe or on smaller servers — is the most sophisticated approach and provides the highest degree of flexibility.


It is essentially a decision to move toward a service-oriented architecture (SOA), by extending legacy (and other) applications as Web services. Adopting SOA will have profound implications not only on the ease with which your IT organization can create new and modify existing applications, but also on how your company does business with partners and customers well into the future.


By providing a library of services that can be called upon to deliver features and perform tasks, SOA enables companies to roll out products faster and adapt applications to meet changing customer demands.


Because it is based on open standards and is widely supported across all vendor environments, as well as across your company's own infrastructure, SOA will not lock you into an architecture that could prove difficult to support in the future.


The Power to Adapt


Understanding where in the application lifecycle management process your legacy system stands is the key to deciding which approach to take.


You want to focus on preserving and extending those applications that are most valuable to your business. To do that, you have to determine what impact the application's lack of availability would have on day-to-day operations. 


You must also consider its overall quality and complexity, your team's knowledge of the actual code, and how much time and money you are willing to invest.


Not all legacy applications are worth extending. If, for example, your company recently acquired another firm and is faced with the challenge of combining two custom-written HR applications into one, a cost-benefit analysis might suggest that you are better off going with a packaged HR application designed to run on smaller servers.


The key to success is understanding that different legacy applications will require different strategies. Twenty-five years ago, a company typically had only one automated system, whereas today it is not unusual to have as many as 100.


With multiple legacy applications, you will likely need multiple approaches to modernizing and integrating. Some applications may simply require a SQL interface. Others may use XML and Web services as the foundation to capture data. There are also cases where simply accessing information from a 3270 screen may be the appropriate solution.


What you need is not a "one-size-fits-all" strategy, but the power to adapt to the environment you are in today — as well as to future, unforeseen changes in the business and technology landscape.


Whether you want to start small, or take a leap toward SOA, a strategy to preserve and extend will enable you to move all of your company's legacy applications forward.


Joe Gentry is vice president of Enterprise Transactions Systems at Software AG. From 1996 to 2001, he held senior management positions in product management at Merant and Intersolv and subsequently became vice president of SpaceWorks. In 2001, he returned to Software AG and now is responsible for market and product strategy for the Enterprise Transaction Systems business line.