Fast dynamic clustering SOAP messages based compression and aggregation model for enhanced performance of Web services

Abstract

The Simple Object Access Protocol (SOAP) is a basic communication protocol in Web services, which is based on eXtensible Markup Language (XML). SOAP could suffer from high latency and bottlenecks that might occur due to the high network traffic caused by the large number of client requests and the large size of XML Web messages. Previous works have proposed static and dynamic clustering models for SOAP messages to support compression based aggregation tool that could potentially reduce the overall size of SOAP messages in order to reduce the required bandwidth between the clients and their server and increase the performance of Web services. In this paper, dynamic clustering based aggregation model has been implemented based on Term Frequency-Inverse Document Frequency (TF-IDF) and Euclidean Distance methods for estimating the high degree of similarity among SOAP messages and then grouping them into a dynamic number of clusters based on lower distance to support Huffman compression based aggregation tool in combining several compressed XML Web messages in one compact message. Our proposed model has achieved better results especially in medium and large subsets of used dataset in comparison with dynamic fractal clustering and in medium, large and very large subsets with vector space model that used the same dataset. Moreover, the experiment results show a significant improvement in reducing the required processing time for clustering XML Web messages in each group of dataset.