WALTy: a tool for evaluating Web application performance
Giancarlo Ruffo, Rossano Schifanella, Matteo Sereno, R. Politi · 2004
In this paper we present WALTy (Web Application Loadbased Testing tool), a set of tools that allows the performance analysis of web applications as they are perceived by the end users. The proposed approach is based on a workload characterization created from informations extracted from log files. The workload is generated by using of Customer Behavior Model Graphs (CBMGs) that allow to describe the behaviors of typical users on a web site. 1 A Framework based on Representative Workload Generation Several approaches have been proposed for capacity planning and performance prediction: on one hand, analytical methods or simulation frameworks that works on a model of the system; on the other hand, we have benchmarking or load testing tools that operate on the real system or on a working test plant.WALTy is a powerful load testing tool written in Java that implements a trace-based stressing framework where traces are generated to describe the behavior of each Virtual User (VU) browsing the web site. In general, many stressing clients replicate a (set of) artificial session(s) or navigational patterns where sessions are randomly or manually generated, or extracted from log files. Our tool is implemented to provide every single VU a profile that is represented by means a Customer Behavior Model Graph [2]. This graph based formalism allows to describe the behaviors of typical users on the web application under test. During a test session, a synthetic workload is generated and performance related measures collected. In such scenario, WALTy uses the