大数据平台负载均衡策略优化设计研究

Research on optimization design of load balancing strategy for big data platform

大数据平台的负载均衡调度策略存在调度效率低、资源利用率偏小以及优质虚拟资源频繁调度等问题,提出将改进混合蛙跳算法应用于大数据平台云工作流负载均衡调度之中,研究用时间贪心的初始化方法替代传统随机方法优化初始种群质量。此外,以动态调度为视阈,提出基于负载感知的均衡调度优化方法。研究结果表明,在搜索效率、负载均衡度以及工作流完成时间上改进混合蛙跳算法均优于传统静态算法。

The load balancing scheduling strategy of big data platform has low scheduling efficiency,low resource utilization and frequent scheduling of high-quality virtual resources.In order to improve it,this study applies the improved shuffled frog leaping algorithm to the load balancing scheduling of cloud workflow on big data platform,and studies the optimization of initial population quality by time greedy initialization method instead of traditional random method.In addition,a load aware balanced scheduling optimization method is proposed based on the dynamic scheduling to be the visual threshold.The research results show that the improved shuffled frog leaping algorithm is better than the traditional static algorithm in terms of search efficiency,load balancing degree and workflow completion time.