Optimization of the Dispatch Probability of Service Requests for Two Servers in Parallel Connection

摘要

This paper describes the procedure used to create a queueing model for servers which are currently in parallel connection with different dispatch probabilities. The system response time of parallel Web servers is improved by calculating dispatch probabilities. The system response time is calculated first. The equation for the lowest system response time is obtained by adjusting the equation to meet the dispatch probability requirements. Then the system response time of the servers is measured. Last, the errors in the system response time of the servers are compared by simulating, modeling and measurement.