Requirements Metadata Driven Dynamic Configuration of Business Processes

Requirements Metadata Driven Dynamic Configuration of Business Processes

Marcus Spies (Ludwig-Maximilians-University of Munich, Germany) and Azzelarabe Taleb-Bendiab (Liverpool John Moores University, UK)
DOI: 10.4018/978-1-60960-485-1.ch009
OnDemand PDF Download:
No Current Special Offers


Services computing is developing towards complex, large-scale infrastructures referred to as the Internet of things and an Internet of services. Cloud computing supports hundreds of thousands of users and applications/services and ensures that these are available on demand, in a reliable and secure way. These infrastructures present many challenges to the design and implementation of systems to manage, protect and tune them. This chapter shows that in order to support the needed interoperability in these infrastructures, such systems must support adaptation based on service semantics and be capable of runtime autonomic behaviour. For engineering such systems, there is a strong need to build on and extend standard software architectures, in particular the service oriented architectures (SOA) and model driven service engineering approaches. Given these requirements, this chapter introduces a model driven approach to a runtime adaptable requirements driven services management system, and presents a theoretical framework building on intention models that are transformed in several steps to dynamic business processes complying with the Web services paradigm and the business process execution language for Web services. These services can be reconfigured at runtime by autonomic behaviours of the management system in line with the intention model. Within the Neptune framework introduced here, the needed modelling languages, model transformations and a runtime environment implementing such an approach are provided. The chapter closes with a case study based on the well known PetShop enterprise application blueprint. The study demonstrates in practice the autonomic adaptation to changes in the services requirements observed at runtime, and concludes with a section highlighting future applications and extensions of the approach.
Chapter Preview


Collaborative business processes orchestrating many partner services are key to contemporary globally synchronized processes, be it in supply chain management, customer relationship management or product lifecycle management (Schroth, 2008). In the technical infrastructure supporting these processes, today’s Service Oriented Architecture (SOA) environments, culminating in an Internet of Services (Ruggaber, 2007), the level at which orchestration is defined is largely dictated by the Web Services Description Language WSDL (Chinnici, Moreau, Ryman and Weerawarana, 2007). WSDL was developed as a web service specific interface definition language borrowing key concepts and structures from CORBA IDL (Common Object Request Broker Architecture – Interface Definition Language). As a consequence, in currently emerging SOA environments, service interaction is defined on the level of data types described in XML schema and interfaces implementing or requesting operations involving these data types. Building on WSDL, services are put into play in real business processes using the Web Services Business Process Execution Language (WS-BPEL) (The Oasis Committee, 2007).

It is clear that interoperability across business domain partitions and real world physical boundaries is a major concern to be addressed. While the present level of service description technologies has enabled business interoperability across a wide range of application, web and data server platforms, orchestrating web services (e.g. crossing the boundaries between Java EE and .NET based services), the level of interaction description, as enabled by WSDL and WS-BPEL has serious shortcomings once dynamic choreographies of services need to be implemented, involving multi-party contracts and service interaction: WS-BPEL and WSDL approaches currently only deal with execution orchestration and syntactic interoperability whereas service interaction across business and organizational boundaries requires dynamic changes to a service choreography and semantic interoperability (Pokraev, Quartel, Steen and Reichert, 2006). Often, in a dynamic service composition, only the generic type of a service is known in advance (like a parcel delivery, or an accounting service, etc). In addition, only generic types of operations (like monitoring a data stream) are known beforehand. This poses a problem for the currently conceived Web Service standards, which only address syntactic interoperability: Ensuring that the data contained in message exchanges is in a common format. It is additionally suggested here that the information exchanged, between a sender and receiver, must have the same meaning for both parties (semantic interoperability) to enable interoperability. Furthermore the sender and receiver must also share an expectation regarding the predicted effect of the exchanged message. This gives rise to a number of requirements that need to be satisfied to enable dynamic E-Business interoperability. Firstly there ought to be an easy translation between system concepts, mapping properties and classes across domains. Secondly, for each interaction, there should exist a defined result, which is predicted (requested) by a contributing system. Finally, for every set of interactions, there must be a predictable result for some system and each set of interactions forms a causal history for the result. Thus, for enabling e-business interoperability on a large scale building on service oriented architectures, runtime dynamism is required, providing service translation and contextual awareness going beyond simple syntactic interoperability. This implies that WSDL and WS-BPEL scripts need to be at least partly composed or recomposed at runtime (Vambenepe, Thompson, Talwar, Rafaeli, Murray, Milojicic, Iyer, Farkas and Arlitt, 2005; Graml, Bracht and Spies, 2008).

Complete Chapter List

Search this Book: