Characterizing temporal locality and its impact on Web server performance

Abstract The presence of temporal locality in web traces has been long recognized. However, the close proximity of requests for the same le in a trace can be attributed to two orthogonal reasons: long-term popularity and short-term correlation. The former re ects the fact that requests for a popular document appear \frequently " thus they are likely to be \close " in an absolute sense. The latter re ects the fact that requests for a given document might concentrate around particular points in the trace due to a variety of rea-sons, such as deadlines or surges in user interests, hence it focuses on \relative " closeness. We introduce a new mea-sure of temporal locality, the scaled stack distance, which is insensitive to popularity and captures instead the impact of short-term correlation, and use it to parametrize a synthetic trace generator. Then, we validate the appropriateness of this quantity by comparing the le and byte miss ratios cor-responding to either the original or the synthetic traces.