基于大型混合网络的突变流量控制模型仿真分析

Mutation Flow Control Model Simulation Analysis Based on the Large Hybrid Network

当大型混合网络受到异常攻击或大规模登录时,流量会发生短时间的巨幅变化,使得单个节点面临瘫痪的风险,导致传统基于单个节点的大型混合网络突变流量控制模型,由于不能适应流量大幅度变化,无法有效实现突变流量控制。提出一种基于自适应PD流量控制算法的大型混合网络突变流量控制模型,将大型混合网络流量统计时间划分成几个周期,获取周期个数的估计值,求出大型混合网络操作行为数据序列中的各统计周期中数据的离均差平方、与组间离均差平方和以及各周期中的方差比值,对流量成分模型进行塑造,通过简单PD控制算法对突变流量进行初步控制,引入延时环节,获取各闭环极点的位置,求出PD控制器参数求出交换节点瓶颈链接个数的估测值,从而实现突变流量控制。仿真实验结果表明,所提方法在控制突变流量方面具有很高的优越性及高效性。

When a large hybrid network anomaly attacks or massive logs in, flow can produce the huge change in a short period of time, make the paralyzed risk faced by a single node.Lead to traditional based on big mutation hybrid network flow control model of a single node, due to the large variation can not adapt to flow, flow control, cannot effectively mutations presents a flow based on adaptive PD control algorithm of large mutation hybrid network flow control model, mix a large network traffic statistics of time into several cycles, obtaining estimates of the number of cycles, and the large hybrid network operating behavior of each cycle of data in a sequence of data from the divided difference square, alienation and group were sum of squared residuals and variance ratio of each cycle, to shape the traffic composition model, through a simple PD control algorithm to control the initial mutation flow, introducing delay link, obtain the closed-loop pole position, and the parameters of PD controller and the estimated value of bottleneck link exchange node number, so as to realize mutation flow control.The simulation results show that the proposed method has the advantages of high in mutation flow control and high efficiency.