论文部分内容阅读
针对高速网络环境下连接记录管理的性能需求,提出了一种改进的高效哈希表PRH-MTF(伪随机哈希-移至最前).首先在定义输入关键字即连接标识符的基础上,通过选择适当的运算符,设计了高效鲁棒的哈希函数PRH.为有效解决哈希冲突,根据网络数据流局部性特点,应用MTF启发法,改进了传统的链式冲突解决方法.以分组火车模型作为数据包到达模式,分析了PRH-MTF哈希表的算法复杂度,推导出了平均查找长度.最后通过实际高速网络数据流和模拟攻击的方式,对PRH-MTF哈希表进行了实验评估.实验结果表明,PRH-MTF哈希表在查找性能和抗攻击能力等方面均优于传统的简单排序哈希表.
According to the performance requirements of connection record management in high-speed network environment, an improved high-efficiency hash table PRH-MTF (Pseudo Random Hash - moved to the top) is proposed.Firstly, on the basis of defining the input keyword, In order to solve the hash conflicts efficiently, the paper proposes a method to solve the traditional chain conflict resolution by using MTF heuristics according to the local characteristics of the network data stream, by selecting the appropriate operators. Train model as the packet arrival pattern, the algorithm complexity of the PRH-MTF hash table is analyzed and the average search length is deduced.Finally, the PRH-MTF hash table is made through the actual high-speed network data flow and simulation attacks The experimental results show that the PRH-MTF hash table is superior to the traditional simple sort hash table in search performance and anti-attack ability.