Performance and Availability Analysis of an E-Commerce Site

To encourage customers to prefer e-commerce services over their on-site counterparts, it is necessary that e-commerce services be offered with superior performance and availability. A systematic methodology to enable a quantitative analysis of the performance and availability attributes of an e-commerce site can assist in identifying bottlenecks and suggesting strategies for improvement. The development of such a methodology is the subject of this paper. The analysis methodology derives expressions for performance and availability metrics of a site which are important from a customer´s perspective, namely, the session length and session availability. The methodology is hierarchical, in the first step expressions are derived for each user group based on the navigational pattern of the group. The navigational pattern of a user group is represented by a customer behavior model graph (CBMG), which is mapped to a discrete time Markov chain (DTMC) for analysis. In the second step, expressions for the session length and availability of the site are derived using the distribution of the customer groups and the expressions for each customer group. We illustrate the potential of the methodology with a case study