CACHE_(RP): A Novel Dynamic Cache Size Tuning Model Working with Relative Object Popularity for Fast Web Information Retrieval

摘要

The novel CACHE_(RP) model for dynamic cache size tuning leverages the relative object popularity as the sole parameter. It is capable of maintaining the given hit ratio on the fly by deriving the popularity ratio from the currently collected statistics. Being statistical the accuracy of the CACHE_(RP) operation should be independent of the changes in the Internet traffic pattern, which can switch suddenly. By adaptively maintaining the given hit ratio through cache size auto-tuning in a dynamic manner the model effectively reduces the end-to-end information retrieval roundtrip time (RTT).