Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Lane Department of Computer Science and Electrical Engineering

Committee Chair

Jim Mooney.


Many systems, including the Web and Software as a Service (SaaS), are best characterized with session-based workloads. Empirical studies have shown that Web session arrivals exhibit long range dependence and that the number of requests in a session is well modeled with skewed or heavy-tailed distributions. However, models that account for session workloads characterized by empirically observed phenomena and studies of their impact on performance and reliability metrics are lacking.;For assessing performance, we use a feedback queue to account for session-based workloads in a physically meaningful way and use simulation to analyze the behavior of the Web system under Long Range Dependent (LRD) session arrival process and skewed distribution for the number of requests in a session. Our results show that the percentage of dropped sessions, mean queue length, mean waiting time, and the useful server utilization are all affected by the LRD session arrivals and the statistics of the number of requests within a session. The impact is higher in the case of more prominent long-range dependence. Interestingly, both the request arrival process and the request departure process are long-range dependent, even in the case when session arrivals are Poisson. This indicates that the LRD at the request level can be a result of the existence of sessions.;For assessing reliability, we propose a framework which integrates (1) the Web workloads defined in term of user sessions, (2) the user navigation patterns through the Web site, and (3) the reliability estimates of the Web requests based on the system architecture; then, we give a detailed reliability model of a Web system based on the proposed framework. We recognize the difficulty of solving the proposed model and use simulation to obtain the results. And last but not least, we use statistical design of experiment to quantify the results and to determine which factors have the highest impact on the system's reliability. Our results show that some two-way and three-way interactions are very important for the session reliability of Web systems.