A General Cache Partition Model for Multiple QoS Classes: Algorithm and Simulation

Previously, Web cache partition by page priority or popularity was individually studied. In this paper, we propose a caching scheme that integrates both QoS classes, priority and popularity. The algorithm is based on the linear combination of the two criterions. In this case, the weight coefficients for the two classes have significant effects on cache performance. The hit rate is also affected by the number of partitions which depends on the combination of different property levels. Simulation programs are developed to examine the effects of the parameters. Results from the simulation show that the new algorithm overcomes the disadvantages of component conflict of partition by individual QoS class. The algorithm can be generated to the case that there are more than two QoS classes