Title: Four tier architecture of component selection process using clustering

Authors: Jagdeep Kaur; Pradeep Tomar

Addresses: CSE and Information Technology Department, School of Engineering, ITM University, Gurgaon, Haryana, India ' Department of Computer Science and Engineering, School of Information and Communication Technology, Gautam Budhha University, Greater Noida, Uttar Pradesh, India

Abstract: A component is selected based on the function it provides in component-based software engineering (CBSE). Many techniques have been proposed for component selection under varied situations to select the optimal component from component set of same functionality. This paper proposes architecture for component selection after going through a brief survey of the component selection techniques. Here, architecture is divided into four tiers namely client requirements and system analysis tier, query and decision tier, application logic tier with clustering and component cluster tier. The current selection techniques using clustering suffers from major demerits of specifying the number of clusters beforehand and the selection process depending on subjective judgement of application administrators. The proposed architecture will be better as the need for apriori declaration of clusters will be eliminated and the cluster validation is performed to check the correctness of the clusters. This architecture is validated on a case study of set of sorting and searching components.

Keywords: component-based software development; CBSD; component-based software engineering; CBSE; clustering; XOR; component selection process; four tier architecture; component clusters; client requirements; system analysis; application logic; cluster validation.

DOI: 10.1504/IJSETA.2015.075611

International Journal of Software Engineering, Technology and Applications, 2015 Vol.1 No.2/3/4, pp.155 - 171

Available online: 30 Mar 2016 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article