Home // International Journal On Advances in Software, volume 4, numbers 1 and 2, 2011 // View article

Metrics for Evaluating Service Designs Based on SoaML

Authors:
Michael Gebhart
Sebastian Abeck

Keywords: service design; soaml; evaluation; metric; quality attribute; quality indicator

Abstract:
In the context of service-oriented architectures, quality attributes, such as loose coupling and autonomy, have been identified that services should fulfill. In order to influence services with regard to these quality attributes, an evaluation is necessary at an early development stage, i.e. during design time. Existing work mostly focuses on a textual description of desired quality attributes, formalizes metrics that require more information than available during design time, or bases on a theoretical model that hampers the practical applicability. In this article, quality indicators for a unique categorization, loose coupling, discoverability and autonomy are identified. For each quality indicator formalized metrics are provided which enable their measurement and application on service candidates and service designs based on the Service oriented architecture Modeling Language as standardized language for modeling service-oriented architectures. To illustrate the metrics and to verify their validity, service candidates and service designs of a campus guide system as developed at the Karlsruhe Institute of Technology are evaluated.

Pages: 61 to 75

Copyright: Copyright (c) to authors, 2011. Used with permission.

Publication date: September 15, 2011

Published in: journal

ISSN: 1942-2628