UniForum '95's track on "Migrating from Mainframes to Open Systems" stems from one major fact of business life: companies today spend hundreds of billions of dollars a year maintaining their mainframe-based or legacy systems, and would like to get out from under that burden. Open systems offers the opportunity, but obstacles and restrictions stand in the way. Will a client/server system really save money, and when? How and when should applications and data be moved off the mainframe, and at what risk? What are the options for re-engineering a current system to reap the real advantages of open systems?
The track's four sessions are designed for the IS manager who is considering or planning a migration to client/server, and needs the information available from both users who have gone through the process and vendors who have tools that may help.
The sessions are:
Way says he sees a need to look at the problem from a perspective of data rather than programming. "There's a new view that we're never going to deal with this legacy system issue if we don't look at data first," he says. "It's the foundation. After you have the data restructured, then you can think about new programming techniques and new ways of manipulating the data you have already accumulated. But if you just come at it from the program perspective, which everybody tends to do, you're lost before you start. That's why there have been no successful large-scale migrations, and that's why mainframe unit sales are still going up."
Yesselman says his session will examine the business and technical drivers toward migration to open systems, and various strategies for making that migration. "Do I cut off my mainframe and start from scratch, do I incorporate the mainframe as part of my distributed client/server system, which applications do I port, which applications do I re-engineer, what are the benefits and costs, and what paths may I take?" In some cases, it may make more sense to leave an application on the mainframe and in others the right solution may be to take it off the mainframe and run it on a UNIX machine. "We want to look at the strategies involved in the migration," Yesselman says. Understanding the costs of client/server computing, including "an awful lot of hidden costs" is vital, he points out, noting, "The cost of open systems and client/server computing are reasonably high and in some cases more than legacy systems."
Matthews says his panel will cover "a blend of what the options are, and some tools and experiences that can help you along the way." The session will include two case studies - from KPMG Peat Marwick and Wells Fargo Bank - as well as one consultant talking about different paths and options, and one vendor talking about tools that might help get ready for such a migration. "The major things are how you are going to take your legacy applications over to UNIX, how you are going to integrate your legacy applications with the client/server environment, and how you take advantage of the client/server tools once you get into that environment," Matthews says.