摘要
IPv6协议规定路由器不再对IP分组进行分段,因此端到端的PMTU探测成为IPv6应用中的关键环节,研究高性能的PMTU探测算法是当前下一代互联网领域中的研究热点.论文针对目前分层探测算法存在等概率探测点带来的发送探测包较多的问题,提出了一种基于经验值的分层探测算法,并从探测精度、总花费时间和总发包数等方面分析了该算法的测试性能,在CERNET2环境下,对比测试了一些常见PMTU算法的性能.实验结果表明,该算法的测试精度、测试耗时和发送探测分组数等性能方面都优于其他探测算法.
In this paper, we suggest some parameters to evaluate the performance of various PMTU discovery methods, such as the precision,the consuming time,and the number of packages to be sent,etc. And we also suggest a new efficient and agile method called Hierarchical-Test based Empiristic-data to detect the PMTU. Then we elaborately discuss the quantitative analyses method of the algorithm.Meanwhile, we have realized the algorithm under Linux operating system environment,and have done some real-world tests based on 58 websites given stochastically within the Education and Research Network of China, Finally, we further analyze the performance of the method through these data that we have attained by experiment.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2007年第10期1865-1869,共5页
Acta Electronica Sinica
基金
国家863高技术研究发展计划(No.2006AA01Z444)
国家自然科学基金(No.60773140)