A priority-based polling scheduling algorithm in web cluster servers

A solution is urgently expected to meet the higher-quality web service requirements of various applications on the limited bandwidth of internet. This paper presents a new scheme of load balance of web cluster servers, making possible the service of the two-class priority station polling system under the mixed policy of exhaustive and gated services, optimizing the services of the system in time of load change input via adjustment of the times of the access gated services, and strengthening the flexibility and fairness of multimedia transmissions in web cluster system. The theoretical model of this scheme is established with Markov chain and probability generating function. Mathematical analysis is made on the mean queue length and the mean inquiry cyclic time of the common queue and the key station. It turns out that the findings from theoretical analysis correspond well with those from simulated experiments.