摘要
在差错信道中传输IP分组时,信道误码容易导致数据包丢失。在无反馈信道的情况下,无法使用反馈重发对数据包进行恢复。因此,该文研究了分组协议中报头字段的差错恢复问题。分析了TCP/UDP/IP等协议中的关键字段,将用于区分传输链路中各路会话的协议字段统称为复用ID。建立了复用ID估计的理论模型,分析了已有文献中基于校验方法对活动复用ID进行检测的缺点,提出利用TCP序列号间关联性的全新检测方法,并针对五元组(源/目的 IP地址,协议号,源/目的端口)提出了估计算法。实验结果表明该方法大大降低了五元组的差错概率,提高了最终的信源还原效果。最后对实验结果进行了分析。
When transmitting Internet protocol (IP) packets over error-prone channels, bit errors can cause packet losses. Non-feedback communications can not then use auto repeat request (ARQ) for packet recovery. This study analyzes the problem of header recovery for the packet protocol and key fields in the transmission control protocol (TCP), user datagram protocol (UDP) and Internet protocol (IP) that act as a multiplex identity (ID) to distinguish different sessions in the transmission link. This paper gives a mathematical model for multiplex ID estimation and analyzes shortcomings in current detection method based checksums for active multiplex ID methods. A detection method is given using the correlations between TCP sequence numbers with an algorithm to estimate the five-tuple containing the source and destination IP addresses, protocol number, and source and destination ports. Tests show that the algorithm significantly enhances the five-tuple error correction ability and improves source restoration.
出处
《清华大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2013年第11期1569-1573,共5页
Journal of Tsinghua University(Science and Technology)
基金
电子信息系统复杂电磁环境效应国家重点实验室开放课题