摘要
在流分类算法中,聚合位向量(ABV)算法分类速度快、并行性好,但内存开销过大;位向量折叠(AFBV)算法对ABV算法进行了改进,降低了运行时内存的消耗,但其冗余计算增加了时间开销。针对上述不足,文章提出一种改进的位向量流分类算法,该算法无需进行位向量聚合,减少了内存开销,并按规则的源/目的IP地址前缀建立分组表,根据表中分组所包含IP地址数目降序排列,使得算法具有良好的时间性能。实验结果表明,本算法在大规模规则库下具有良好的时间和空间效率。
In view of flow classification,aggregated bit vector（ABV）algorithm is faster and has good parallelism,but the memory overhead is too large.The aggregated and folded bit vector（AFBV）algorithm improves ABV algorithm with less run-time memory consumption,but its redundant computation increases time overhead.An improved flow classification algorithm based on bit vector is proposed,which does not require bit vector aggregation,reduces memory overhead,establishes prefix grouped table according to the source/destination IP address prefix of rules,and makes the descending order by the number of IP addresses contained in the groups of table,so that the algorithm can have good time performance.The experimental results show that the algorithm has good time and space efficiency for large-scale rule base.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2015年第3期331-335,共5页
Journal of Hefei University of Technology：Natural Science
基金
安徽省自然科学基金资助项目(11040606M138)
关键词
流分类
聚合位向量(ABV)算法
位向量折叠(AFBV)算法
位向量
flow classification
aggregated bit vector（ABV）algorithm
aggregated and folded bit vector（AFBV）algorithm
bit vector（BV）
