网络结点缓冲区动态占用特性分析

The Analysis of Dynamic Occupation Character Based Net-Node Buffer

针对网络结点的阻塞控制与结点缓冲区的关系问题 ,利用交叉开关网络结构及其非阻塞特性 ,采用概率论分析方法 ,结合网络流量的非线性特征 ,对通过结点缓冲区的流量以及缓冲区的动态占用情况进行了分析 ,得出了缓冲区动态占用情况的定量描述函数 ,并作了相关的证明和验证 ,为基于网络结点缓冲区占用情况的流量控制建立了一定的基础 .有利于结点缓冲区优化控制机制的建立 ,提高结点缓冲区的适应能力 ,降低因为缓冲区不适应流量而造成的阻塞概率 .

This paper is on the relationship between of traffic control of net-node and its buffer. The blocking often happens in network. One of the main causes is lacking node resources and poor management, such as the buffer in network node. We hope to know the buffer state of node by analyzing the traffic through the node buffer and further study the optimization management method of the buffer. The utilization ratio of network node resource can be raised and the probability of network block can be brought down based on our research. Adopting the analytic method of probability theory and combining the nonlinear characteristic of network traffic, we analyze the flow of net-node and the occupation of net-node buffer based on non-block characteristic of the crossbar network and its structure. Crossbar network possesses strictness non-block characteristic. We give five hypotheses to make use of the crossbar characteristic and submit the logic partitioning method to be convenient for analysis. Furthermore, we analyze the network traffic of crossbar network with storage modules. We get the probability of one module that is always accessed in periodic time and give the expression. The fixed capacity function of the dynamic occupation of the buffer is obtained. It is proved and verified. The traffic characteristic of network node is analyzed in detail. Also, we discuss the relationship of the situation of node buffer input-output and the state of node. Based on two cases where the network traffic is subordinated to homogeneous Poisson distribution or inhomogeneous one, the relative functions which express the amount of requests remained in the buffer alter along with time are given. In our future research, these functions and conclusions are the good foundations.