Legacy Software Integration in Service-Driven Environments: An Intelligent Agent-Based Framework

Introduction

Enterprise IT systems have typically undergone a long period of evolution since the 1960s. These decade-long projects have eventually produced some of the most complex software systems in existence. For large corporations, mainframe applications programmed in COBOL often form the backbone of the IT structure. The operation and maintenance costs of these applications are usually very high. Worst of all, is the lack of business agility. Introducing new features to meet the constant growth of a corporation takes far too long and costs too much.

Maintaining and upgrading legacy systems is one of the most difficult challenges that many companies face today. They struggle with the problem of modernizing these systems while keeping the day-to-day operations intact. Despite their obsolescence, legacy systems continue to provide a competitive advantage by supporting unique business processes and acting as a repository for invaluable knowledge and historical data.

Pang (2012) presented an approach to improve the business agility of legacy IT systems by modernizing COBOL application development. The special features of this approach included the introduction of a Service-oriented Architecture (SOA) for Web client integration, using a model-driven development approach with code generation following an agile development process. This development approach has been used successfully in many projects since 2004. Despite the success of these projects to produce high quality software, the approach addresses application development in the legacy mainframe platform. Many corporations would like to get away from further development in this platform. They prefer to develop their applications with modern software technology, but leverage existing business processes and data from their legacy systems.

Eventually, the legacy business processes and the database in the legacy systems will be replaced by newly developed software with successive processes and data migration. This development strategy leads to a number of technical challenges. First, we need an enterprise architecture that can incorporate software components running on different platforms. Second, we need a programming model that seamlessly allows the invocation of modules in different programming languages. Third, we must look at the performance issues of running the final applications. Computing resources in the mainframe are expensive. It takes a lot of resources to start a transaction in the mainframe and the number of transactions must be limited in any given application. This chapter presents an architectural framework and implementation methodology for legacy mainframe integration using the Enterprise Service Bus (ESB) and intelligent agents on the modern and legacy environments to coordinate the integration.

This chapter is organized as follows. The background section provides basic concepts and techniques that are prevalent in the industry to handle legacy integration in service-driven environments. The next section details the proposed architecture and implementation methodology that facilitates the integration of service clients with the mainframe legacy systems by detailing the service agents, enterprise service bus, scripting and rule engines, that enable the integration of different systems including Web services, mobile systems, and external systems in the Cloud with the mainframe host systems. A Central Intelligent Agent is described that is used to control the business processes and distribute the data processing to different software components on different platforms. Finally, the chapter presents service development tools in the mainframe and the application development and migration process; and discusses future research directions and conclusions.

Background

Advances in software technologies have shown a lot of promise. Back in the late 1980s, software scientists claimed that Object-oriented (OO) languages were for programming in the large (Wegner, 1989). The OO paradigm, with polymorphism and inheritance, was a solution to resolve and control the complexity of enterprise applications. Corporate executives have been led to believe that if they could port all of their legacy applications to a new platform using programming languages such as Java and Smalltalk, they will have business agility and witness a great reduction in their operating cost. In practice, however, this often turned out to be an illusion.