干扰状态下大规模网络数据流的分发协调技术研究

摘要

Large-scale network data flow interference conditions possesses the characteristics of real-time,volatility,sudden and current data stream distributed coordination method cannot ensure that data flow smooth transmission and real-time distributed coordination.For this,a new interference methods coordination state of large-scale network data flow distribution and flow of data through the network error between the predicted value and actual value interference filter out interference condition data,ensure the rest of the data flow of purity was put forward.Large-scale network data flow distribution coordination problem is transformed into a multi-objective optimization problem,in order to improve the large-scale network data flow distribution of coordination of integrity for the purpose to set an objective function,a kind of large-scale network data flow distribution coordinate prediction technology,through comparing the processor computing resources and future expected load to determine the data flow of distributed coordination strategy was put forward.The experimental results show that the proposed method for large-scale network data flow interference condition distributed coordination,not only is high efficiency,good realtime,and CPU resources utilization rate is very high.%干扰状态下大规模网络数据流具有实时性、波动性、突发性等特点,当前数据流分发协调方法无法保证数据流的平稳传输及分发协调的实时性.为此,提出一种新的干扰状态下大规模网络数据流分发协调方法,通过网络数据流预测值和实际值之间的误差滤除干扰状态下的干扰数据,保证剩余数据流的纯正性.将大规模网络数据流分发协调问题转换成多目标优化问题,以提高大规模网络数据流分发协调的整体性能为目的设定一个目标函数,提出一种大规模网络数据流分发协调预测技术,通过比较处理器计算资源和未来的预期负载确定数据流分发协调策略.实验结果表明,采用所提方法对干扰状态下大规模网络数据流进行分发协调,不仅效率高、实时性好,而且CPU资源利用率很高.