An iteration‐based interactive analysis method to design dynamic service‐oriented systems. (19th September 2017)

• Record Type: Journal Article
• Authors: Xie, Wuping; Xue, Jinyun; Jiang, Dongming; Song, Lan
• Other Names: Mora M. guestEditor.; O'Connor R. V. guestEditor.; Tsui F. guestEditor.; Marx Gómez J. guestEditor.
• Is Part Of: Software, practice & experience. Volume 48:Number 2(2018)
• Page Start: 300
• Page End: 326
• Publication Date: 2017-09-19
• Subjects: service computing -- service‐oriented design -- dynamic system -- system analysis -- design evaluation; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.2543
• Languages: English
• ISSNs: 0038-0644

Abstract

Summary: Service‐oriented paradigm presents numerous new software development patterns and idioms. Software systems are implemented by composing existing third‐party services deployed in the open environment, which is significantly different from traditional software development methodologies in which systems are built through developing modules after system design in a closed environment. Therefore, it is urgent to raise a new design method to adapt to this new circumstance. We concentrate on reusing as many deployed services as possible then introduce a new life cycle model named Taiji model to illustrate this development process. The iteration‐based interactive analysis method following the model is proposed to design service‐oriented systems based on the view of extracting non‐creative activities from a creative activity through defining new notations or applying new rules. The method includes the interactive analysis process that analyzes requirements with deployed services in a local point of view and the iterative analysis process that redesigns system with new knowledge in a global perspective. Meanwhile, the reusable service threshold value is defined to build the uncertain candidate service set (UCSS) of each module in analysis process. The reliability and flexibility of systems can be improved through the quantitative static structure analysis on the basis of the UCSS of systems. Meanwhile, a practical dynamic service binding method that selects services according to actual states of invoking them is presented on the basis of the UCSS containing them. Finally, we also give a case study to illustrate the feasibility of this method. Copyright © 2017 John Wiley & Sons, Ltd.