Design and Implementation of a Tool for Measuring the Performance of Complex E-commerce Sites

E-commerce applications are growing at unprecedent rates, resulting in overloaded sites with poor quality of service. The workload intensity of an e-commerce site is not totally predictable given that external events can generate load spikes that exceed by far the average load. Therefore, e-commerce site managers need to be able to understand the performance of the site and be able to tune it to cope with varying traffic patterns. In this paper we present PROFIT, a new tool for profiling the performance of e-commerce sites. PROFIT measures both throughput and response time and breaks down the response time in terms of components (e.g., Web server, application server, and database server) and services (e.g., search, browse, select, add to cart, and pay). To illustrate the use of the tool, the paper shows an analysis of performance and security in e-commerce applications, measuring the impact of the Secure Sockets Layer (SSL) protocol on the request response time.