Server Initiating Schedule Algorithm to Enhance the User's Satisfaction Rate
MA Jun-cai · Journal of Chinese Computer Systems · 2013
In a cluster environment,in order to achieve multiple servers work together to make full use of computing resources,it needs appropriate load balancing strategy.Commonly used server scheduling policy focused on the servers to achieve hardw are load balancing,w hile ignoring the quality of service.The proposed algorithm defines customer satisfaction rate(the ratio of the task completion time w hich the user w ith the actual completion time of the task) to evaluate the quality of service.The algorithm in this paper integrates four factors including user w eight,the average computing time of the type w hich the task belong to in history,the task completion time w hich the user expect and the time w hich the task added to compute the task w eights,and then the servers take the initiative to dispatch according to the level of task w eights.The results show that the algorithm can improve the customer satisfaction rate,and make full use of computing resources.