Network performance implications of multi- dimensional variability in data traffic
Chris Roadknight, I. Howard Marshall · 1999
WWW traffic will dominate network traffic for the foreseeable future. Accurate predictions of network performance can only be achieved if network models reflect WWW traffic statistics. Through analysis of usage logs at a range of caches we confirm that WWW traffic is not a Poisson arrival process, and that it shows significant levels of self-similarity. We show for the first time that the self-similar variability extends to demand for individual pages, and is far more pervasive than previously thought. These measurements are used as the basis for a cache modelling toolkit. Using this software we illustrate the impact of the variability on predictive planning. The model predicts that optimisations based on predictive algorithms (such as least recently used discard) are likely to reduce performance very quickly. This means that far from improving the efficiency of the network, conventional approaches to network planning and engineering will tend to reduce efficiency and increase costs.