SIGMETRICS 2001 / Performance 2001

Analyzing Robot Behavior in E-Business Sites

Authors

Virgilio Almeida
Dept. of Computer Science, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil

Daniel Menasce
Dept. of Computer Science, George Mason University

Rudolf Riedi
Dept. of Electrical and Computer Engineering, Rice University Belo Horizonte, Brazil

Rodrigo Fonseca
Dept. of Computer Science, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil

Wagner Meira Jr.
Dept. of Computer Science, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil

Flavia Peligrinelli
Dept. of Computer Science, Universidade Federal de Minas Gerais,

Abstract

Understanding the nature and characteristics of e-business workloads is an essential step to improve quality of service. Using a multi-layer hierarchical workload model, this paper presents a characterization of the workload generated by autonomous agents and robots. The characterization of the robot workload focuses on the statistical properties of the arrival process and on the robot behavior graph model. A multi-scale time analysis is used to understand the impact of robot workloads on the performance of e-business sites. Based on the workload characterization, we develop an analytical model for the interaction between robots and e-business sites. Using the model, we show the resource utilization associated with robots' requests. Multiple time-scale analysis point out that server overload may occur at fine time scales, even though average utilizations at larger time-scales do not show evidences of overloading. These results indicate not only the need for a better understanding of the behavior of robots, but also help to quantify the impact of robots on the quality of service provided by a site.

[Last updated Fri Mar 23 2001]

Web Accessibility