面向SLA的云计算负载均衡策略

Load Balancing Strategy of Cloud Computing Oriented to SLA

云计算负载均衡是保障SLA协议的关键问题之一;针对云计算负载均衡问题,提出一种面向SLA的负载均衡策略;该策略引入人工神经网络思想,建立负载均衡模型,采用单层感知器算法(SLPA)将虚拟机负载状态进行分类,然后利用结合了动态加权轮询算法的BP神经网络算法(BPNNA-DWRRA)有针对性地对虚拟机负载权重进行预测更新,最后将任务调度到最小权重所对应的可行虚拟机上;应用CloudSim进行仿真实验,结果表明了该策略的可行性,同时,相比加权最小链接算法和粒子群算法,该策略的平均响应时间分别节省了43.6%和22.5%,SLA违反率分别降低了20.7%和14.4%;因此,所提策略在响应用户任务时,请求响应时间短,SLA违反率低,保障了SLA。

Load balancing of the cloud computing is one of the key issues of guaranteeing SLA agreement.Load Balancing for cloud computing,a load balancing strategy oriented SLA is put forward.Introduced artificial neural network thinking,this strategy establishes load balancing model.VMs load status is classified by single-layer Perceptron algorithm （SLPA）,then BP neural network algorithm （BPNNA）combined with dynamic weighted round-robin algorithm （DWRRA） targetedly predicts and updates VMs lo（a）d weight,finally,the task is dispatched to the feasible virtual machine according to the minimum weight.The CloudSim simulation experiment was performed,and the results show that the proposed strategy is feasible.Meanwhile,compared with the Weighted Least-Connection algorithm and particle swarm optimization algorithm,the strategy of this paper respectively reduce average response time by 43.6％ and 22.5％,and reduce SLA violation rates by 20.7％ and 14.4％.Therefore,the proposed strategy in response to user tasks,make response time short,SLA violation rate low,and SLA is guaranteed.