Interoperability between enterprise applications requires an understanding of the obstacles to interoperability in order to provide methods for overcoming these obstacles. We address the problem from different angles: first, we investigate the benefit of multi-lingual ontologies to overcome language problems between users of an enterprise application. Second, we propose a method to integrate legacy components into new enterprise applications. Third, we research the impact service-oriented computing to enterprise application integration. Finally, we present our results with agent-oriented platforms for integrating autonomous applications.
• meta modeling: objects are instances of classes, classes are instances of meta classes, meta classes are instances of meta meta classes etc. The abstraction hierarchy is virtually unlimited allowing to model multiple representation schemes into a single uniform framework • advanced query language: the query language is based on logical deduction. Save recursion over negation is supported as well as aggregation functions. Answers can be returned in XML or other text-based formats
Web-services are rapidly becoming the de-facto technology for developing and integrating highly distributed business applications that are capable of supporting cross-organizational business processes. Web-services can be defined as self-describing, interoperable and reusable business components that can be published, dynamically combined and invoked through the Internet, even when they reside behind a company’s firewall.
Web Service Technology is facing a number of daunting obstacles before it can actually deliver its tantalizing promises of massive reuse and seamless cross-enterprise integration. One of the most severe obstacles is the integration of newly developed constellations of web services with existing enterprise legacy systems. These systems are hard to maintain and adapt to new business requirements, but support the value-adding business operations of many brick-and-mortar companies. Existing webservice based solutions for legacy integration are based on the rather unrealistic assumption that legacy systems can be wrapped and reused integrally without any major modification. In addition, they are predominantly technology focused while largely neglecting business process semantics. Lastly, a method for integrating cohesive fragments of legacy system and data with new web-services is currently lacking. At the INFOLAB we have been developing a comprehensive method, named BALES, that encompasses a wide range of techniques and a supporting prototypical toolset for systematically designing and constructing web-service based, cross-organizational business applications with wrapped legacy systems[13]. The method is business-process driven and supports the notion of selective integration of semantically enriched portions of encapsulated legacy application logic and corresponding data.
Service-Oriented Computing (SOC) has been emerging as a new computing paradigm that utilizes services as fundamental elements for developing loosely coupled distributed applications/solutions. The platform neutral nature of SOC paradigm holds the promise for interoperability support based on the way the distributed software applications are developed and deployed in the SOC environment. In order for the realization of the full potential of SOC, many issues need to be addressed, which are best presented and described in the Extended Service Oriented Architecture (ESOA) [5]. The ESOA is a layered architecture that utilizes the basic SOA constructs as its bottom layer, on top of which it layers a composition layer and a service management layer. We have been working on the ESOC architecture that provides separate tiers for composing and coordinating services and for managing services in an open marketplace by employing grid services, and use it to streamline and integrate our research activities. Currently we are mainly focusing on the middle layer of the ESOC in the following two areas: •
Service composition: the big challenge of service composition lies in the fact that the generation of the composition must be dynamic in terms of service selection and composition construction; and the result of composition must be configurable, manageable and reusable to meet the ever changing business requirements. To support dynamicity, configurability, manageability, and reusability of service composition, we have been working on the following projects: o Business rule driven service composition: Service composition needs to be flexible in the way that is defined, scheduled, and constructed. It also needs to adapt to changes so that compositions do not need to be re-generated whenever changes occur. To support various types of service composition, we first need to identify the rules and policies that determine the structure of the composition and prescribe alternative services and provider selection. we then need to separate the rules that govern the composition from the composition specifications. This means developing a rule based mechanism that manages the entire life cycle of service composition. Some initial results can be found in [6, 7]
Model driven approach for service composition: The benefits of adapting system analysis and design methodology in service composition is that we gain much more insights in the process of service composition development so that we can maintain traceability between models and design. In our MDA approach we use UML as the method for designing service compositions. This will enable us to develop technology independent composition definitions, which can subsequently be mapped to specific standard (e.g. BPEL) automatically. Furthermore together with UML we use Object Constraint Language (OCL) to express business rules that can govern and steer the process of service composition. Some results can be found in [8].
Service transaction management: The aim of this project is to develop theory, mechanism, and
infrastructure for supporting collaborative transactions in the web service environment. This will
lead to a comprehensive framework providing: (
Agent-orientation is a new software engineering paradigm. An agent is a software module, just like objects and components are. One major break with the object-oriented tradition is that the modules are considered to be autonomous. This has consequences for the design. The main effort in the design process is how to bring together modules (agents) that were developed independently from each other, typically on the basis of different ontologies, and that will remain independent as well. In other words, the focus is on the design of the coordination [11], rather than on the functionality of the isolated components. The question how this coordination is to be achieved and maintained is precisely the main question in interoperability.
In [10] an agent-oriented design method has been introduced that uses agreement and refinement. The interoperability question cannot be addressed without considering the infrastructure, or “society” in which the agents meet each other. What should be present, minimally, to make interoperability possible in the first place? What coordination mechanisms should be supported? Reaching agreement (against the background of a certain infrastructure) is another necessary capability for software entities to achieve interoperability. This cannot be done without making explicit what is to be agreed upon. In our research, we have explored this problem under the heading of contract-driven coordination [12]. An expressive formal logic language for contracts is described in [14]. The language is based on deontic logic; this is because dealing with autonomous (rather than completely controllable) systems implies that norms, goals and commitments are not always kept, and the interoperating systems should be able to cope with these violations.
[5]: Mike P. Papazoglou and Dimitrios Georgakopoulos (2003): Introduction. In CACM 46 (10): 24-28, Special
Issue: Service-oriented computing. [6] B. Orriens, J. Yang, and M.P. Papazoglou, “A Framework for Business Rule Driven Service Composition”. In
VLDB-TES2003, Sep, Berlin, 2003. [7] Jian Yang and Mike P. Papazoglou, “Service components for managing the life-cycle of service compositions”.
Information Systems 29(