Dynamic Data Mediation in Enterprise Application Integration Scenarios

Introduction

In the mid-1990s, a new term called enterprise application integration (EAI) was established, which introduced several methods and software components for efficiently integrating software in an enterprise. While in the past, the need for integration was driven by lower-level needs to share information among applications, today, business drivers fuel the need for integration (Vollmer & Peyret, 2005) – the need for business to respond to regulatory challenges, improve business process execution, grow employee productivity, and improve customer service is a hallmark driving organizations to integrate applications. In response to these new requirements, the integration market landscape is shifting. Service-Oriented Architectures (SOAs), business process management (BPM), composite applications, and other new application requirements have become the driving force in the market (Rymer et a., 2004).

The currently available enterprise application integration software address integration problems for example in the following ways (Friesen et al., 2007): a) graphically supporting the mapping of systems’ interfaces to each other (e.g. SAP NetWeaver Exchange Infrastructure), b) reducing complexity using intermediate data-exchange languages (e.g. eXtensible Markup Language – XML), or c) reducing the number of connection adapters needed through the introduction of hubs (e.g. Enterprise Service Bus). These efforts entail significant costs and typically due to the “lack of automated support in defining integration, it takes a long time for a human engineer to define semantically correct integration” (Bussler, 2003).

Current industrial EAI trends and technologies, like SOA, Enterprise Bus and Web Services, are up to now quite mature. However, if we try to increase the level of automation in integration scenarios, we confront several problems and challenges (Bouras et al., 2007), such as a) data and message level heterogeneities between interoperating services, b) insufficient search and discovery of published Web Services in a common registry, and c) inadequate Web Process composition with regard to the desired functionality and operational requirements. The problem that still exists, which the traditional, syntactic EAI technologies are weak to solve, refers to the formalization and the documentation of the semantics related to the interfaces and the data structures of the deployed Web Services. This lack of formal semantics of applications and services to be integrated makes it difficult for software engineers and developers to interconnect heterogeneous applications and thus creates obstacles in the automating EAI activities (Haller et al., 2006). There is no doubt that these needs impose the use and interpretation of semantics in EAI and that semantically enriched approaches will hopefully mitigate these problems.

In this book chapter, we present a semantically-enriched approach for dynamic data mediation in Enterprise Application Integration scenarios, based on Ontologies, Semantic Web and Semantic Web Services Technologies. Our approach, which is presented in the next section, focuses on the resolution of message level heterogeneities between collaborative enterprise services exposed from the participating business systems, facilitating automatic, dynamic data mediation during execution time by providing formal transformations of the input and output messages (of the participating Web Services) to a common reference model, i.e. an enterprise data ontology.

In addition, Section 3 provides an overview of the enterprise data ontology that we have developed and utilized as part of a multi- layered and –faceted interoperability ontology, called Enterprise Interoperability Ontology, which provides a shared, common understanding of data, services and processes within enterprise application integration scenarios.