Time-domain analysis of Web cache filter effects

Abstract

This paper uses trace-driven simulation to study the traffic arrival process for Web workloads in a simple Web proxy caching hierarchy. Both empirical and synthetic Web proxy workloads are used in the study. The simulation results show that a Web cache reduces both the peak and the mean request arrival rate for Web traffic workloads, while the variance-to-mean ratio of the filtered traffic typically increases, depending on the input arrival process and the configuration of the cache. If the input traffic is self-similar, then the filtered request traffic remains self-similar, with the same Hurst parameter, though with reduced mean. Finally, we find that a Gamma distribution provides a flexible and robust means of modeling aggregate workloads in hierarchical Web caching architectures, for a broad range of workload characteristics and Web proxy cache sizes. To demonstrate the generality and effectiveness of the modeling approach, we present a detailed example of filter effects and traffic superposition in a two-level Web caching hierarchy with heterogenous input workloads. The Gamma modeling results match well with the results from trace-driven simulations.

Workload characterization in Web caching hierarchies

“…First, we generate a set of synthetic Web proxy workloads to use in our study. We validate these workloads to ensure that they have the intended workload characteristics, and are similar to empirical Web proxy workloads used in our earlier studies [5,9,10,18,23]. Second, we conduct a set of trace-driven simulation experiments, using an applicationlevel Web proxy caching simulator and the generated workloads.…”

“…Web workload characterization studies have focussed on Web client [7], Web server [3], and Web proxy workload characteristics [2,5,18]. Common workload characteristics observed include a high degree of one-time referencing, a Zipf-like document popularity distribution, heavy-tailed file and transfer size distributions, and a temporal locality property in the document referencing behaviour [3,9,18].…”

“…Few papers explicitly address the structural changes in the request arrival process in multi-level Web caching hierarchies. Our own previous work [5] studied time-domain cache filter effects using an empirical Web proxy workload, but only for a single-level Web proxy cache. This current paper extends our work to caching hierarchies, and generalizes our results to a broader set of (synthetically-generated) Web workload characteristics.…”

“…We refer to the changes in the document popularity profile as frequency-domain cache filter effects, and the changes in the request arrival process as time-domain cache filter effects. The frequencydomain effect has been fairly well-studied in the literature [10,12,14,23], while the time-domain effect has received little attention [5]. It is the latter (time-domain) effect that is the focus of this paper.…”

“…The precise manifestation of the cache filter effect is highly dependent upon the architecture of the Web proxy caching system, the cache size used, and the cache replacement policy [5,10,23]. Since the goal in Web proxy caching is to improve overall "system-level" performance, it is important to design an appropriate caching system architecture.…”