摘要
由于传统的混合服务器差异节点负载均衡调度方法存在调度效果差、负载均衡度低、执行作业耗时长的问题,提出一种云平台下混合服务器差异节点负载均衡调度方法。依据提取的混合服务器差异节点负载目标特征,采用自然数编码方式,设置染色体的路经以及长度个数,对差异节点编号,将所有节点都作为一个目标特征内的染色体,以此检测云平台下混合服务器的差异节点负载情况;利用基于线性的时间序列方式,构建负载模型,根据BP神经网络信号的前向传播和误差反向传播对负载计算,通过负载均衡调度策略实现差异节点负载均衡调度。经过实验证明:所提方法的负载均衡调度效果较好,负载均衡度较高,缩短了系统执行作业所用时间。
Due to poor scheduling effect,low load balancing degree and long execution time in traditional methods,this article presented a load-balance scheduling method of differential nodes in mixed server on cloud platform.According to the load target characteristics of the extracted differential nodes in mixed server,the natural number coding method was used to set the path,length and number of chromosomes,and thus to number the differential nodes.Moreover,all nodes were taken as the chromosomes in one target feature to detect the load of differential nodes in the mixed server on cloud platform.The linear time series method was used to construct the load model.Based on the forward propagation and error back propagation of BP neural network signal,the load was calculated.Finally,the load-balance scheduling strategy was adopted to achieve the load-balance scheduling of different nodes.Simulation results show that the proposed method has good load-balance scheduling effect and high load balancing degree.Meanwhile,this method shortens the execution time of system.
作者
张晨
孙鹏
ZHANG Chen;SUN Peng(Xi'an Kedagaoxin University,Xi'an Shanxi 710109,China)
出处
《计算机仿真》
北大核心
2021年第1期455-459,共5页
Computer Simulation
基金
基于云服务模式的高校综合管理平台的实现研究(18JK1047)。
关键词
云平台
混合服务器
差异节点
负载均衡调度
Cloud platform
Mixed server
Differential node
Load-balance scheduling
