An analytic model of Web servers in distributed computing environments

Web technologies are currently being employed to provide end-user interfaces in diverse computing environments. The core element of these Web solutions is a Web server based on the HyperText Transfer Protocol (HTTP) running over TCP/IP. Web servers are required to perform millions of transaction requests per day at an acceptable Quality of Service (QoS) level in terms of client response time and server throughput. Consequently, a thorough understanding of the performance capabilities and limitations of Web servers is critical. In many applications, the Web server performs significant dynamic server-side processing in distributed computing environments. In these applications, a Web server retrieves a file, parses the file for scripting language content, interprets the scripting statements, and executes embedded code, possibly requiring a TCP connection to a remote application for data processing and transfer. Web server performance in a distributed environment is a complex interplay between a variety of factors. Although testing is essential for assessing performance, there are inherent limitations to a testing approach for validating Web server performance. Consequently, modeling is critical to further understand the performance capabilities and limitations of Web servers that participate in distributed computing. In this paper, we illustrate a model-based approach to Web server performance evaluation, and present an analytic queueing model of Web servers in distributed environments. Performance predictions from the analytic model match well with the performance observed from simulation. The model forms an excellent basis for a decision support tool to allow system architects to predict the behavior of new systems prior to deployment, or existing systems under new workload scenarios.