Authors: M. Buvana; M. Suganthi; K. Muthumayil
Addresses: Department of Computer Science and Engineering, PSNACET, Dindigul, Tamilnadu, India ' Department of Electronics and Communication Engineering, TCE, Madurai, Tamilnadu, India ' Department of Information Technology, PSNACET, Dindigul, Tamilnadu, India
Abstract: With the increase in the number of available services, the use of the service has been increased rapidly and finding the exact services also the challenging task in large-scale mobile ad hoc network. We propose a method named as SD-LDA (linear discriminant analysis-based service discovery) algorithm to find the web services. Our method exploits the clustering structure, where the cluster head node (CHN) is selected to form a distributed directory. This CHN handles the discovery agent in which LDA is used for service discovery by reducing the dimension of service data and ignores the inequality of local data points of the similar class using matrix representation and calculating the overall score of QoS for each service, then ranks the services based on their overall QoS and selects services with highest QoS values. Finally, the performance of our algorithm is evaluated by experiments. The results show that the clustering method has low communication overhead and success of service discovery is increased. In addition, we show that less energy consumption during service discovery, minimise the discovery overhead, and rank identification.
Keywords: service discovery; web services; linear discriminant analysis; LDA; mobile ad hoc networks; quality of service; QoS; MANETs; clustering; energy consumption.
International Journal of Internet Protocol Technology, 2016 Vol.9 No.2/3, pp.90 - 99
Received: 08 Jul 2015
Accepted: 26 Mar 2016
Published online: 29 Sep 2016 *