一种面向云计算的分态式自适应负载均衡策略

在云计算环境中,大规模并行任务的运行容易造成某些节点负载过重,进而导致整个云计算平台负载不均衡和效率低下。针对此问题,提出了一种面向云计算的分态式自适应负载均衡策略。该策略根据节点的负载度判断节点负载的状态,当节点处于轻度过载或重度过载时,自发地执行过载避免或快速均衡的方法。该策略通过动态调整节点的效益度,使轻度过载的节点能够尽量避免重度过载,重度过载的节点能够快速恢复到正常负载。实验结果表明,分态式自适应负载均衡策略能有效实现云计算系统中的负载均衡,提供高效的性能。

A Sub-state Adaptive Load Balancing Strategy for Cloud Computing

In cloud computing systems,large parallel tasks may lead to overloaded for some nodes and cause the load unbalanced and inefficient for the system.A sub-state adaptive load balancing strategy is proposed for cloud computing system.The strategy judges the state of load by load degree.When a node is in the mild or severe overloading status,it spontaneously runs overload avoiding or quickly balancing method.The strategy adjusts the effectiveness degree of the node dynamically,so that mild overloading nodes can avoid severe overloading and severe overloading nodes can return to normal load quickly.The experiment results show that the sub-state adaptive load balancing strategy ensures the load balancing and efficientive performance in cloud computing systems.