Dynamic Requests Scheduling Model in Multi-core Web Server

A popular web site is expected to deal with a large number of requests simultaneously in the reasonable mean response time. The performance of the web sites mainly relies on hardware performance and the processing strategy of requests. In order to improve the hardware performance, more and more web servers are adopting multi-core CPUs. Moreover, the scheduling algorithm of requests on basis of first-come-first-served (FCFS) is still utilized. Although FCFS is a reasonable and fair strategy for requests sequence, it doesn´t take the distribution of the dynamic requests service time and the characteristic of multi-core CPUs into account. In this paper, in order to solve the above problems, a new dynamic requests scheduling model is presented. The new scheduling model, according to the distribution of the dynamic request service time, schedule the dynamic request based on weight-fair-queueing (WFQ), and exploit the performance of multi-core CPUs by means of hard affinity in O/S. Simulation Experiments has been done to evaluate the new scheduling approach, and the results show that the new scheduling model can reduce the mean response time efficiently.