云计算下多源异构大数据跨源调度算法

Multi Source Heterogeneous Data Scheduling Algorithm under Cloud Computing

针对云计算下的多源异构大数据,需要耗费大量时间对其进行传输,当前调度算法大多通过启发式算法实现大数据调度,加速比和吞吐量较低,负载均衡性不佳。为此,提出一种新的云计算下多源异构大数据跨源调度算法,为了降低计算开销,在调度前首先进行预取操作。在此基础上,对全部变量进行更新处理,将所有待调度多源异构子流的质量看作子流权重进行排列,从传输窗口中还存在剩余待调度多源异构子流中选择质量最佳的子流传输数据,直至全部待调度多源异构子流均被处理过。实验结果表明,所提算法加速比和吞吐量高,且负载均衡性好。

Aiming at the multi-source heterogeneous big data cloud computing,a lot of time need be spend to transmit the current scheduling algorithm mostly through the heuristic algorithm to achieve large data scheduling,lower than the acceleration and throughput,load balancing of the poor. For this reason,a new scheduling algorithm was proposed for multi source and heterogeneous data based on cloud computing. On the basis of all the variables are updated,all quality to be scheduled for multi-source heterogeneous sub streams as sub stream weights are arranged from the transmission window are still remaining to be heterogeneous in scheduling sub optimal sub quality streaming data,until all the pending transfer degree of multi-source heterogeneous sub streams are processed. The experimental results show that the proposed algorithm has high speedup and throughput,and good load balance.