Modeling web requests: A multifractal approach

World Wide Web (WWW) caching is used to improve network latency and bandwidth usage by storing previously requested files in a cache. Ideally, the cache replacement policy should account for the intrinsic characteristics of WWW traffic, which include temporal locality, spatial locality, and popularity. In this paper, we accurately capture these three characteristics in a stochastic model, which can be used to generate synthetic WWW traces and assess WWW cache designs. To capture temporal and spatial localities, we use a modified version of Riedi et al.'s multifractal model, where we reduce the complexity of the original model from script O sign(N) to script O sign(1); N being the length of the synthetic trace. Our model has the attractiveness of being parsimonious (characterized by few parameters) and that it avoids the need to apply a transformation to a self-similar model (as often done in previously proposed models), thus retaining the temporal locality of the fitted traffic. Furthermore, because of the scale-dependent nature of multifractal processes, the proposed model is more flexible than monofractal (self-similar) models in describing irregularities in the traffic. Trace-driven simulations are used to demonstrate the goodness of the proposed model in terms of generating representative WWW streams and approximating the cache performance of actual WWW traffic.