Authors: Mostefai Abdelkader; Mimoun Malki; Sidi Mohamed Benslimane
Addresses: Department of Computer Science, University of Dr Tahir Moulay, Saida, Algeria ' Department of Computer Science, University of Djilali Liabes, Sidi belabess, Algeria ' Department of Computer Science, University of Djilali Liabes, Sidi belabess, Algeria
Abstract: The identification of candidate web service in legacy software is a crucial task during the process of reengineering legacy software towards Service Oriented Architecture (i.e. SOA). In this paper, we propose an original approach that formulates the problem of locating services as search problem and we justify the adoption of functional cohesion measures and GA heuristic (i.e. Genetic Algorithms) to find a good enough solution, a set of modules (i.e. procedures or functions) contributing in the computation of searched service, given that deriving the optimal solution is shown to be NP. Two experiments are conducted on a two medium-sized legacy software and results show that it achieves better precision and recall than features location approaches based on information retrieval tools when used to locate services in legacy software.
Keywords: legacy software; web services; genetic algorithms; cohesion measure; service oriented architecture; SOA; heuristics; service location.
International Journal of Computer Applications in Technology, 2013 Vol.47 No.2/3, pp.152 - 161
Published online: 05 Jun 2013 *Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article