Modeling continuous changes of the user's dynamic behavior in the WWW

Abstract

Understanding the characteristics of the users' workload is an important aspect when designing and providing web services. The majority of current workload characterization techniques introduce some limitations when representing the dynamism of the client behavior and the continuous changes in its role. This fact implies that the majority of the existing workload generators model these characteristics in a simple an improperly way. This paper focuses on the dynamism of WWW in general, and the new techniques to characterize the user behavior. Our work is addressed to develop a dynamic workload generator that considers the new behavior of the web clients, and the continuous changes in their role.

Generating Probabilistic and Intensity-Varying Workload for Web-Based Software Systems

“…Peña-Ortiz et al [11] provide an overview of outstanding and historical workload generators including an evaluation in terms of their features and capabilities. We explicitly modeled the workload generator on a conceptual level including the execution semantics and implemented the resulting tool Markov4JMeter as an extension for the popular workload generator JMeter [3].…”

“…Implementing such a tool from scratch would have required us to implement a bunch of low-level functionalities which do already exist in a number of workload generation tools (cf. [11] for an overview of existing tools). Instead, we integrated our approach into the popular open source workload generator Apache JMeter [3], and could thus focus on the implementation of those functionalities specific to our approach.…”

WEB WORKLOAD GENERATORS - A Survey Focusing on User Dynamism Representation

“…In a previous work (Peña-Ortiz et al, 2005) we proposed the GUERNICA approach to model the dynamism of the web workload based on users' behavior models. This paper extends that work in several ways.…”

From Reconfigurable Architectures to Self-Adaptive Autonomic Systems

“…Nowadays information systems can be seen as aggregates of complex architectures, spanning from grids including thousands of geographically distributed systems dependent upon multi-site collaboration [8]- [10], and fast, reliable access to shared resources and data [11], to small and specialized embedded systems [7]. We achieved such a complexity in the infrastructure, no matter what kind of architecture it is based on, because of the increasing demanding for high performance computation, high reliability and for the need in providing always different, more complex, and high quality services to the users [12], [13].…”