Study On User Experience Based Multi-QoS Constrained Multi-Process Scheduling Algorithms

With the development and population of information technology, Internet is becoming a new computing platform. The applications based on the Internet are distributed and cooperating processes. And now they face more sorts of clients than before. So the requirements are becoming diversity. The diversity of requirements is not only focus on the fuctional requirements, and also QoS requirements. At the same time the number of Internet users is increasing sharply. This leads to the high concurrency of service requests. Then numerous process instances are in execution. Based on the analysis above, the large scale processes on the Internet platform have two features:the diversity of QoS requirements and the large number of processes.This paper is focus on the diversity of QoS requirements and the large number of processes problems which the large scale processes are faced. The main objective is to satisfy more users'QoS requirements and enhance the user experience. Based on this objective, the user experience oriented multi-QoS constrained multi-process scheduling strategy is proposed. The scheduling strategy can schedule a large number of processes and enhance the user experience. This scheduling strategy includes a multi-QoS constrained multi-process scheduling algorithm and a user experience measure algorithm. The first algorithm addresses the multi-QoS constrained multi-process scheduling problem. This method is mainly to meet as many users'QoS requirements, rather than to improve the executing efficiency of one user. The algorithm decomposes the QoS requirements of the process to the task of the process. And then scheduling the tasks accord to the QoS constrained. The second algorithm introduces the description and measure method of the user experience. The users'feedback is taken into account in the scheduling process of this method to improve the user experience. By this way, the resources can be selected according to the user experience degree (UED). So the users'requirements can be met better and the user experience can be improved. Experiments show that the multi-QoS constrained multi-process scheduling algorithm can satisfy more users'QoS requirements and effectively improve the success rate of scheduling. Furthermore, by using two algorithms introduce in this paper, the users'feedback can be taken into the scheduling process and the user experience can be improved.The algorithms proposed in this paper are the core technology of large scale cooperation applications. They can improve the execution effenciency and the throughput. Moreover, the user experience can be enhanced obviously. On the other hand, the scheduling strategy in this paper is one of the key technologies in the cloud computing platform. So it has a wide prospect of application.