随着 Internet 的飞速发展,网络规模不断扩大,结构日趋复杂,网络带宽、用户数量和网络业务不断增长。如何有效地管理网络资源,避免和控制网络拥塞,保证网络服务质量是通信网络研究的重要课题。主动队列管理(AQM)算法在网络中间节点中对拥塞进行早期检测。通过将路由器中的队列控制在较小的数值,减小数据包排队时延和抖动的同时,可文持较高的链路利用率。 通过对主动队列管理算法中的随机指数标记算法(REM)进行了深入研究,REM 算法利用队列长度的偏差和输入输出速率的偏差计算出一个链路价格,依据这个链路价格来计算丢弃率。从控制理论的角度来说,速率是队列长度的微分,因此相比于仅仅利用队列长度进行拥塞控制的算法,REM 算法对拥塞的发生有更好的早期判断能力。本论文在 Network Simulator 平台下对算法进行了仿真,并且与随机早期检测算法(RED)进行对比,将丢包率,吞吐量,端到端的时延等性能指标进行了分析。10035
最后对全文所做的研究工作进行了总结,并指出了有待进一步研究的问题。
关键词 拥塞控制 主动队列管理 随机指数标记算法
Title Implementation of the control network congestion
algorithm REM in NS2
Abstract
With the rapid development of Internet, the scale of network tends to widen,
and its structure becomes more complex. Also the bandwidth, numbers of
users as well as types of network transactions grow fastly. How to
efficiently manage the resource of network, avoid and control network
congestion, and guarantee the quality of service have been a hot research
issue in communication networks.
Conducted in-depth study of active queue management algorithm in the random
exponential marking algorithm .REM algorithm using the queue length
deviation and input / output rate deviation calculated a link price,
according to this link price to calculate the loss rate. From the control
theory point of view, rate of queue length differential, so compared to
using only the length of queue congestion control algorithm, REM algorithm
for congestion has better early judgement. Based on the Network Simulator
platform under simulated, and compared with the random early detection
( RED) were compared, the packet loss rate, throughput, end-to-end delay
and other performance indicators are analyzed.
Finally, we draw our conclusions and propose some further research
directions.