Interview With Liz Barnett, Giga Information Group

by David Baum

Giga Information Group is riding on the crest of a tidal wave of change. Its research focus is on helping companies integrate their businesses with the Internet. Its analysts use the information gained to offer sober management advice to corporate IT decision-makers. Elizabeth Barnett finds herself in the thick of these issues in her focus on application development and IT management topics, including component-based development, legacy renewal, process management, quality, testing, and reuse. Information Builders Magazine recently spoke with Ms. Barnett to obtain her insight into an increasingly relevant topic: renewing, enhancing and extending legacy technologies.

IB Magazine – What are the primary motivators for today's legacy renewal efforts?

Barnett – One motivator is the tremendous backlog of demand that has built up for new applications for new functionality to be added to existing applications. A lot of companies have spent millions of dollars fixing their Year 2000 problems. Because of this, they haven't had the time or the resources to dedicate to new development. It makes a lot of sense to leverage existing assets to meet these new demands.

Another motivator is the insatiable demand for the Web and all things Internet-related. For most companies, there is no time to lose: release cycles are now six to nine months as opposed to eighteen to twenty-four months to deliver new applications. From a competitive viewpoint, there is an urgency to get application assets online.

IB Magazine – Legacy renewal used to simply mean "screen scraping." What are some of today's popular legacy renewal strategies?

Barnett – For starters, there is still the low-end, non-invasive scenario of screen scraping/giving a face-lift to an old application, such as a new Web front end, for example. You take an application built using old technology and add on some new functionality, but you don't have to open up the old code. In many cases, there is an interface redesign as part of this process. You can think of this as "extending the reach" of legacy applications.

IB Magazine – When is an outright migration justified, in which an application or database is moved to a new platform?

Barnett – Several years ago, we commonly saw companies migrating applications to a new platform or to a new distributed architecture. Typically, this was a means to reduce costs and to potentially get off the mainframe. Today, we're seeing this less and less, but there is still a need for migration in certain cases.

For example, if you have an old application that managed VSAM data, and you want to put the data into a relational database, that would be considered a special kind of migration project. Or if you adopted SAP enterprise resource planning software on a UNIX or NT platform, you might want to co-locate your subsidiary applications on the same platform as your new transactional systems.

IB Magazine – What role does middleware technology play in these legacy renewal efforts?

Barnett – Middleware fits into what I call middle-tier integration where you open up the existing code and start to reuse parts of the application in new ways. Often, these new components are deployed on a middle-tier platform that can act as an integrator. You can think of this approach as "extending the use" of legacy applications rather than "extending the reach." The correct approach is to work from the top, down. You want to determine your strategic architecture going forward, then decide how to map your new applications to your old legacy assets. Of course, you need to be realistic and have a good understanding of your legacy code so you know what's feasible. Legacy renewal should not be a solution looking for a problem.

IB Magazine – Most legacy applications have come under intense scrutiny during Y2K analysis and remediation efforts. How can companies leverage their Y2K work after the turn of the century?

Barnett – Companies have invested immense resources in bringing their legacy code into compliance. They have also invested heavily in the staff members who are responsible for Y2K efforts. The question is how to reuse the technology and experience gained during the remediation process. It's not necessarily recouping the investment, but leveraging dollars that have already been spent.

The good news is, what we've learned about our legacy applications often goes beyond just scanning code to find dates. I sometimes call it "application mining" the process of establishing a repository of knowledge about the code base. Through Y2K efforts, we become better informed about our legacy assets. Some of the systems that had the curtains drawn over them for decades are now well understood. That foundation of understanding can be parlayed into a strategy for maximizing the use of those assets going forward. Companies should make sure that all Y2K inventories and databases are available as resources to new application development teams.

IB Magazine – When it comes to the user interface – to optimal ways of presenting information – is there anything unique that developers of e-business systems need to be aware of?

Barnett – With traditional client/ server, there was a lot of contextual knowledge built into the client side of the application. Web architectures imply the reverse. You give a user access to an application, yet you are managing everything on the application server. There are different design principles that come into play. For example, there is a lot of intelligence and context management that can be built into a client/server application that is much harder to build into a Web-based application. If you are migrating a fat client application onto a thin client, or replacing it with a browser, you can't necessarily pull all that old functionality across. If anything, some of the older legacy architectures will be a better fit with Internet applications than their client/server successors.

IB Magazine – Within the domain of legacy renewal, what are the key capabilities that customers should look for from vendors and consultants?

Barnett – You need a well-rounded organization to do the job well. There are companies – consulting firms and tool vendors – that understand the legacy environment, but they often lack solid Web development and infrastructure skills. Conversely, there are Internet development companies that build fabulous applications and Web sites, but they can't spell COBOL. It is very rare to find a single company that can bring together a team with all the requisite skills to build these applications.

My advice to customers is to scrutinize the skills of the team from both ends of this spectrum. Make sure that, collectively, they have the breadth that you really need. Remember, the long-term goal isn't just to build Internet applications, but to integrate those applications with the rest of the information enterprise. Do not underestimate the complexity that this integration may pose.

IB Magazine – How do companies acquire the skills to undertake legacy renewal projects?

Barnett – Given the costs associated with recruiting new staff, there is a very real incentive to use existing staff to handle legacy renewal projects. Existing employees really understand the business. They own the business processes, and they have the subject matter expertise, so it makes sense to reinvest in these people rather than replace them. For these reasons, I think we're going to see more and more energy placed into retooling the workforce. Of course, for any new technology, businesses will initially rely heavily on outside sources, such as vendors and consultants. But there should always be a focus on skills transfer and mentoring, rather than staff supplementation. The long-term goal is to retain existing staff, particularly those with experience managing and extending corporate information and applications.