摘要
研究多协议标签交换(MPLS)网络中的流量工程,提出了基于标签交换路径(LSP)时延和拥塞来优化网络流量的算法.给出全网流量优化的数学模型,并根据该模型内在的层次性,设计了分布式自适应流量工程算法.为了使算法具有实时性强、收敛速度快、稳定性较好、运算和网络开销小等特点,在设计中,采用了禁忌搜索和算法参数的非线性调整来提高算法的性能.仿真分析表明,该算法能够实现对网络性能的全面优化,适合应用于骨干网的流量优化.
A distributed adaptive traffic engineering algorithm HDCA (heuristic delay and congestion based algorithm of traffic engineering) for the MPLS (multi-protocol label switching) network is proposed in this paper. This algorithm is based on the LSP's (label switching path) delay and congestion. A mathematical model of the whole network traffic optimization was established, and HDCA algorithm was designed in terms of the mathematical model's hierarchies. In this algorithm, tabu search and nonlinear changes of parameters were adopted to achieve excellent real-time performance, fast convergence speed, high stability, little running time and little network overhead. The simulation analysis shows that the HDCA algorithm fits to the traffic optimizing in the bone network.
出处
《北京邮电大学学报》
EI
CAS
CSCD
北大核心
2005年第2期63-67,共5页
Journal of Beijing University of Posts and Telecommunications
基金
博士点基金资助项目(2003001312)
国家自然科学基金资助项目(60372101)
关键词
流量工程
启发式
多协议标签交换
数学模型
Computer networks
Computer simulation
Data flow analysis
Delay control systems
Mathematical models
Optimization
