Dynamic Hilbert clustering based on convex set for web services aggregation

Abstract

In recent years, web services run by big corporations and different application-specific data centers have all been embraced by several companies worldwide. Web services provide several benefits when compared to other communication technologies. However, it still suffers from congestion and bottlenecks as well as a significant delay due to the tremendous load caused by a large number of web service requests from end users. Clustering and then aggregating similar web services as one compressed message can potentially achieve network traffic reduction. This paper proposes a dynamic Hilbert clustering as a new model for clustering web services based on convex set similarity. Mathematically, the suggested models compute the degree of similarity between simple object access protocol (SOAP) messages and then cluster them into groups with high similarity. Next, each cluster is aggregated as a compact message that is finally encoded by fixed-length or Huffman. The experiment results have shown the suggested model performs better than the conventional clustering techniques in terms of compression ratio. The suggested model has produced the best results, reaching up to 15 with fixed-length and up to 20 with Huffman