云计算环境下基于优先级的IO和网络密集型应用调度策略

当前云计算环境中,当大于CPU核数的IO和网络密集型应用并发执行时,传统的资源分配策略没有考虑到应用的特性,导致资源利用率偏低,应用执行效率低下.针对这种现状,本文对IO密集型应用和网络密集型应用进行分析,根据它们可量化的特性,提出并设计了基于优先级的IO和网络密集型应用调度策略.针对可量化的小应用提高优先级,获得更大的CPU时间片,让小应用尽早完成,然后将所有CPU时间片分配给大应用,减少进程之间的切换调度,提高了效率.大量实验表明,该策略可以有效提高应用的执行效率,减少资源的消耗.

Scheduling Strategy of IO and Network Intensive Applications Based on the Priority in Cloud Environment

At present,in cloud environment,when the times of parallel IO and network intensive applications are more than the number of CPU cores,the traditional resource allocation strategy is not taking the characteristics of the applications into account,which leads to the low utilization rate of resources and the low efficiency of applications execution.To solve this situation,the IO intensive applications and network intensive applications were analyzed,and the scheduling strategy for IO and network intensive applications was put forward based on the priority.The priority of small applications was improved to get larger CPU time slices,so that small applications can be completed as soon as possible,and then all the CPU time slices are allocated to large applications,to reduce the process of switching between schedulings so as to improve the efficiency.A large number of experiments show that the strategy can effectively improve the efficiency of the applications execution and reduce the consumption of resources.