Dynamic Fractal Clustering Technique for SOAP Web Messages

The Simple Object Access Protocol (SOAP) is an XML based protocol that is widely used over the Internet as it supports interoperability by establishing access among Web servers and clients from the same or different platforms. However, SOAP Web services suffer the bottlenecks and congestions as a result of Web messages being bigger than the real payload in addition to the potentially increasing demand of the requested Web services. Aggregation of SOAP messages is an effective solution that has been developed to significantly reduce network traffic by aggregating SOAP messages at the server side and then multicast them to the Web clients. The major problem of the aggregation techniques is that they require efficient similarity criteria that can compute the similarity of SOAP messages as group-wise and not just pair-wise. In this paper, a new unsupervised auto class Fractal clustering technique is proposed for clustering SOAP messages into a dynamic number of clusters according to their Fractal similarities. The experimental results showed that the proposed Fractal clustering technique can improve the performance of Web services significantly better than other clustering standards such as the K-means and PCA combined with K-means by enabling the aggregation model to aggregate the most similar messages in one group resulting in better messages size reduction. Furthermore, the proposed Fractal clustering technique potentially reduces the required processing time in comparison with other standards.