摘要
针对传统RED算法数据包丢弃概率计算方法的不足,提出一种RED改进算法—分段随机早期检测(Partitioned Random Early Detection,PRED)算法。为了更好地调控网络拥塞,算法通过调整数据包丢弃概率计算函数,以两种不同趋势曲线分段增长方式代替原RED算法包丢弃概率与平均队列长度之间单一的线性关系。系列仿真实验结果验证了改进算法的有效性,一定程度地提高了网络性能。
Aimed at the deficiencies of the traditional RED packets discarded probability calculation method, an improved RED algorithm(Partitioned Random Early Detection, PRED) is proposed in this paper. In order to control network congestion, new algorithm replaces the linear growth relationship between RED algorithm packet discard probability and the average queue length with the partition "growth of two different trend curves by regulating the calculation function of packet discard probability. Results of series NS2 simulation validate the efficiency of the improved algorithm and enhance the network performance to some extent.
出处
《计算机工程与应用》
CSCD
2013年第16期92-96,共5页
Computer Engineering and Applications
基金
国家社会科学基金项目(No.CCA110109)
福建省教育厅科技项目(No.JB11205)
江西省教育厅科研项目(No.GJJ13228)
江西省光电子与通信重点实验室项目(No.2011010)
江西省分布计算工程技术研究中心项目(No.2012006)
宁德师范学院科研资助项目(No.2011305)
关键词
网络拥塞
分段随机早期检测
数据包丢弃概率
NS2
network congestion
Partitioned Random Early Detection(PRED)
packet discard probability
NS2