Benchmark test of clustered transaction processing for electronic commerce

Abstract

The purpose of this paper is to explore clustered transaction processing for electronic commerce (EC), where the cluster is composed of more than one arriving transaction. As transactions in EC are processed in a short time, straightforward application of conventional transaction processing to EC protocols only makes processing efficiency worse because of the overhead of the transaction processing. Motivated by this observation, we introduce the cluster into transaction processing for EC. We investigate a scheduling rule for the cluster in order to execute transactions efficiently. We also introduce the concept of urgent transaction with the highest priority to execute by overriding the urgent transaction in terms of the cluster. Finally, we report a benchmark test of our proposed clustered transaction processing to verify the efficiency of our method for electronic commerce.