光突发交换网络中的传输控制协议性能  

TCP Performance in Optical Burst Switching Networks

作　　者：胡忠良[1] 毛期俭[1] 亓波[1] 徐昌彪[1] 隆克平[1] 

机构地区：[1]重庆邮电学院光互联网及无线信息网络研究中心,重庆市400065

出　　处：《电信快报》2004年第11期12-15,共4页Telecommunications Information

摘　　要：光突发交换(OBS)是IP over WDM核心网络采用的交换技术。在OBS网络中,送往同一边缘节点的IP分组汇聚成传输和交换的基本单元———数据突发(DB),DB丢弃会导致大量IP分组丢失,显著影响传输层的性能。文中分析OBS网络的参数对传输控制协议(TCP)吞吐量和时延的影响。仿真结果表明,DB丢包率越高,TCP性能越差。在低丢包率情况下,随着DB长度的增加,TCP吞吐量和端到端时延明显增加,高丢包率情况下则不明显;随着汇聚周期的增加,TCP吞吐量逐渐下降,端到端时延逐渐增加。Optical burst switching(OBS) is the switching technology used in IP over WDM backbone network. In OBS network, IP packets destined to a common egress router are converged into a basic unit for transmission and switching -- data bursts (DB), at the network edge and then routed as a single unit. IP packet drops produced as a result of data burst drops can significantly affect performance of the transmission layer. This paper discusses the effect of OBS parameters on TCP throughput and delay. Simulation results show that the higher the rate of DB drop the poorer the TCP performance. For low drop probabilities, with increasing DB size, TCP throughput and end to end delay increase markedly. For higher drop probabilities, this increase was not significant with increasing burst size. Increasing convergence timeout values resulted in gradual decrease of TCP throughput and gradual increase of end to end delay.

关 键 词：光突发交换 OBS 时延 传输控制协议 吞吐量 IP分组 核心网络 DB 丢包率 端到端 

分 类 号：TN929.1[电子电信—通信与信息系统]