TCP拥塞控制机制浅析  

Discussion on TCP Congestion Control Mechanism

作　　者：杨彦彬[1,2] 

机构地区：[1]华南理工大学电信学院,广东广州510640 [2]广东邮电职业技术学院通信系,广东广州510630

出　　处：《通信技术》2009年第4期58-60,共3页Communications Technology

摘　　要：TCP是当今网络中主要的传输协议,它采用了慢启动、拥塞避免、快速重传、快速恢复四种算法,能满足IP网络中数据的可靠传输。但是当出现多个数据包丢失时,由于TCP采用了累计确认机制,造成系统吞吐量下降。文章介绍了一种SACK拥塞控制机制,与传统的Tahoe、Reno对比,并通过仿真实验说明了SACK是一种最好的TCP恢复机制。TCP is the major transportation protocol, which employs slow start, congestion avoidance, fast retransmit, and fast recovery algorithms, and could realize reliable data transportation in IP network. However, when there exist lots of lost data packets, the throughput is lowered because of TCP＇ s ACK mechanism. This thesis proposes a SACK congestion control mechanism, and the experimental simulation has proved that SACK is a good TCP recovery mechanism, and much better than conventional Tahoe and Reno.

关 键 词：传输控制协议 拥塞控制 Tahoe Reno SACK 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]