The Revenues Driven Resource Allocation Algorithm for Cluster-based Web Server*

In order to optimize the revenues attained in the cluster based Web server by dynamically allocating the pool´s size of each service class under QoS constraints, the R-PSA (revenue driven pools´ sizes allocation) scheme is provided. In this scheme, the pool´s sizes allocation of different service classes is linked with the revenues in cluster based Web server under QoS requirement constraints, and the revenues target function is formulized according to queue theory, and the problem of maximizing revenues of system is solved by dynamic programming algorithm combining heuristic schemes. The simulated experimentations show the efficiency of R-PSA under different workload