Exploring K-Means Clustering and skyline for Web Service Selection

During the last decade, an exponential growth of web services is observed over the Internet. This offers a big challenge for the web service based systems to make the optimal selection of the desired web service. In this work, we have used a two layer architecture for web service selection, prefiltering followed by selection. The use of K-Means clustering technique for grouping the web services with similar Quality of Service (QoS) under a common umbrella is explored. This act as prefiltering step for candidate web services to filter out unrelated web services. From the set of filtered web services, a non-dominated set of web services is obtained using skyline technique. The first step ensures to include only those web services, which are related based on QoS information. The second step operates on the reduced problem set and identifies the best web service among the group. The real world web service dataset is used to test the approach and an improvement in the web service selection is observed.