ISBN 9783843923606

978-3-8439-2360-6, Reihe Informatik

Aneta Kabzeva
Multi-Layer Modeling of Service Networks The DAME Approach to Generic Dependency Analysis and Management for Evolution

313 Seiten, Dissertation Technische Universität Kaiserslautern (2015), Softcover, A5

Zusammenfassung / Abstract

With the broad adoption of service-orientation for the realization of distributed systems, especially for business applications, as well as their provisioning and usage over Cloud infrastructures, the organization of today’s IT landscapes is changing. Organizations are focusing on their core competencies, while supporting non-core functionality is consumed on demand as third-party services. The resulting sophisticated relationships between services of different vendors and consumers connect the IT landscape in a complex service network.

Being a complex and dynamic distributed system of service systems, the maintenance and evolution of a service network has called for the definition of new modeling approaches. A formal model can support evolutionary decisions in the network by providing a description of the complex relationships emerging between the independent network participants in the context of applications composed of cross-domain services. However, the variety of stakeholder perspectives on a service network and the different abstraction layers of a service-oriented application make the specification of a generic service network model a challenging task.

This thesis proposes the DAME generic modeling approach for the representation of multi-layer relationships in complex service networks. A set of models that unify different aspects of service network entities and relationships while keeping the possibility for extension and methods that allow the configuration of data collection, analysis, and representation are developed for this purpose. DAME enhances the state of the art in service network modeling by providing a holistic approach including the representation, collection, and analysis of a service network model. At the same time, it considers the freshness of the model and provides dynamic customization for the transformation, evaluation, and visualization of relationship data.