Integration of Reliability Prediction into Service Composition Frameworks

Service composition frameworks hide the complexity of the underlying technologies and enable their users to easily compose new services from existing ones. To anticipate the reliability of composed services before their actual operations, it is possible for service composers to use existing model-based software reliability prediction techniques. However, such techniques either neglect or have only basic integration support into service composition frameworks. This limits their applicability because (1) neglecting integration support leads to high effort and specialist knowledge required to build models necessary for reliability predictions and (2) basic integration support likely reduces the flexibility and reusability when integrating into different service composition frameworks. In this paper, we present RMPI-SerComp to offer enhanced integration support into service composition frameworks for RMPI, a recent model-based software reliability prediction technique. Our approach enables service composition frameworks to achieve a systematic consideration of service reliability during composing, without the need for any specialist knowledge of the underlying prediction method. Via a case study, we validate the applicability of our approach by utilizing RMPI-SerComp to integrate the reliability prediction capability into a service composition framework. Report Date: September 14, 2015 Written Language: English Any Other Identifying Information of this Report: Distribution Statement: First Issue: 3 copies Supplementary Notes: Active Knowledge Engineering Laboratory The University of Aizu Aizu-Wakamatsu Fukushima 965-8580 Japan Integration of Reliability Prediction into Service Composition Frameworks Thanh-Trung Pham and Alexander P. Vazhenin School of Computer Science and Engineering The University of Aizu (UoA) Aizu-Wakamatsu, Fukushima, Japan Email: {pttrung,vazhenin}@u-aizu.ac.jp Abstract—Service composition frameworks hide the complexity of the underlying technologies and enable their users to easily compose new services from existing ones. To anticipate the reliability of composed services before their actual operations, it is possible for service composers to use existing model-based software reliability prediction techniques. However, such techniques either neglect or have only basic integration support into service composition frameworks. This limits their applicability because (1) neglecting integration support leads to high effort and specialist knowledge required to build models necessary for reliability predictions and (2) basic integration support likely reduces the flexibility and reusability when integrating into different service composition frameworks. In this paper, we present RMPI-SerComp to offer enhanced integration support into service composition frameworks for RMPI, a recent model-based software reliability prediction technique. Our approach enables service composition frameworks to achieve a systematic consideration of service reliability during composing, without the need for any specialist knowledge of the underlying prediction method. Via a case study, we validate the applicability of our approach by utilizing RMPI-SerComp to integrate the reliability prediction capability into a service composition framework.