Performance Evaluation of XML Query Processing in Centralized and Distributed Environment

Abstract

eXtensible Markup Language (XML) has been used to transfer data among a wide variety of systems. The increasing usage of XML data and increase in query workloads makes XML unrealisable for centralized storage. Therefore, a distributed query evaluation strategy is well-suited to access these types of collections without having to ship large volumes of irrelevant data across the network. A centralized planning and distributed execution strategy can be used in processing XML queries in distributed manner. In this paper, a technique that processes query in distributed environment is presented, with a pruning technique in local distributed servers and the central server federates the results sent by the distributed servers. A series of evaluations were that compares the performance of centralized and distributed techniques using same set of queries on two different datasets. The results show that the proposed technique, D-DGReLab+ outperformed other centralized techniques, TwigStack and QTwig.