No abstract

A fundamental problem in the area of service engineering is the so-called cross-cutting nature of services, i.e., that service behavior results from a collaboration of partial component behaviors. We present an approach for model-based service engineering, in which system component models are derived automatically from collaboration models. These are specifications of sub-services incorporating both the local behavior of the components and the necessary inter-component communication.The collaborations are expressed in a compact and self-contained way by UML collaborations and activities. The UML activities can express service compositions precisely, so that components may be derived automatically by means of a model transformation. In this paper, we focus on the important issue of how to coordinate and compose collaborations that are executed with several sessions at the same time. We introduce an extension to activities for session selection. Moreover, we explain how this composition is mapped onto the components and how it can be translated into executable code.

This work focuses on techno-business analysis of service platforms and service portfolios. Service platform (SP) hosts services and enabling service functionality. Service providers deliver two main products: end-user services to their customers and enablers to other business actors. 3rd party service providers combine enablers with their own functionality, wrap them in end-user services and deliver them to their customers.SPs are very complex systems technically as well as in user and business aspects. Ignoring or oversimplifying any aspect can decrease the realism and value of techno-business analyses. On the other side, including all the aspects would make the model too complex to be practically useful. We present in this paper a scenario-driven approach for technobusiness modeling and analysis, which is sufficiently simple to be practical and complete enough to give realistic results. It is based on a generic service platform model (GSPM) (represented by ontology, structural and mathematical models) combined with scenario-based modeling of service portfolios and model-based mapping and projection techniques. It enables techno-business analysis at an early stage of service development to serve as a foundation for investment decisions.This analytical framework has been used in a series of practical cases. One of them, provision of mobile service bundles, is presented in this work. We discuss our experience and suggest future improvements in the proposed approach to model driven techno-business analysis.

No abstract

Current trends in distributed computing and e-business processing suggest that many applications are evolving towards Service Oriented Computing (SOC) with technologies such as Web services. Services are autonomous platform-independent computational elements, and we observe an increasing need for core SOC technologies for dynamic discovery, selection, and composition of services. However, such technologies are often based on syntactic descriptions of the services and of their interfaces, which are insufficient to ensure that desired liveness properties are satisfied. In this paper, we propose an approach for the description, discovery, and selection of services based on role modeling and goal expressions that enables the definition of semantic interfaces and the evaluation of liveness properties. The same mechanisms also enable component reuse. We discuss how UML 2.0 can support the modeling of both the services and the desired properties. The approach is illustrated with telephony services.