Legacy Application -- Dinosaur?
In some cases, legacy apps are built on what we called a burning platform. Although sometimes there are packages you can buy to replace them, the problem might be "access" rather than a business or technology mis-match. Most importantly, we need to keep in mind that not all legacy problems are the same. There is a dark-side of the replacement of legacy applications: some of the biggest failures in business IT is in the large scale package replacement. So, Re-Write the application from scratch can sometimes be risky. In a lot of cases, SOA (Service Oriented Architecture) may come in handy, particularly in a mainframe world for many people who think that mordernizations are hopeless and the fact that all applications are not created equal.
Strategy: Non-Invasiveness
Process includes:
- Application Portfolio management
- Presentation Integration
- Programmatic Integration -- Microsoft .com, Java Classes (or some other programmatically consumable way)
Organizations' Top Concerns:
-
TCO (Total Cost of Ownership)
-
Skills (ability to continue to find Mainframe-Style Programmers)
-
Agility (respond fast enough)
Advantages of SOA
SOA doesn't require everyone to subscribe to the same technology religion. It can let users consume those in any technology environment instead of building the application specificly in Linux, Unix, Windows or Java.
How do I determine which application make good target for modernization? Four different categories:
A. Infrastructure
B. Utility Applications
C. Enhancement Applications
D. Frontier Applications
Summary
Reduce cost, reduce risk and reduce cost for maintenance are 3 of the most important benefits to migrate legacy applications. SOA can allow you to deliver business solutions a whole lot faster than otherwise, if you are going to wait for another complete package solution or complete re-write in order to do that.
(References to Garterner Research Notes)