论文部分内容阅读
包分类算法的性能直接影响数据包的收发速度,决定了网络的时延和吞吐量。防火墙中使用分类算法进行过滤规则的匹配查找,能有效降低规则匹配搜索时间,极大地提升防火墙的性能。递归流分类(RFC,Recursive Flow Classification)算法查找速度快,但预处理时间长,存储开销大。现在RFC算法的基础上,结合哈希树算法对数据包各字段分开处理。将两种算法结合,综合考虑了空间和时间性能,不仅减少了存储开销,而且能保持相对快的查找速度。
The performance of the packet classification algorithm directly affects the sending and receiving speed of the data packet, which determines the network delay and throughput. The classification algorithm in the firewall is used to search the matching rules of the filtering rules, which can effectively reduce the time that the rules match the search and greatly improve the performance of the firewall. The Recursive Flow Classification (RFC) algorithm finds fast, but the preprocessing time is long and the storage cost is large. Now based on the RFC algorithm, combined with the hash tree algorithm to deal with each field separately. Combining the two algorithms, taking into account both space and time performance, not only reduces storage overhead but also maintains relatively fast lookup speed.