基于交叉位图的多维流分类算法  

Algorithm of multi-dimensional flow classification based on intersecting bitmap

作　　者：胡茂福[1] 侯整风[1] 韩江洪[1] 何玲 

机构地区：[1]合肥工业大学计算机与信息学院,合肥230009 [2]深圳金山信息安全技术有限公司,广东深圳518057

出　　处：《计算机应用研究》2010年第8期3058-3060,3063,共4页Application Research of Computers

基　　金：安徽省自然科学基金资助项目(090412051);广东省教育部产学研结合项目(2008B090500240)

摘　　要：聚合位向量算法(ABV)是一种快速的流分类算法,但由于空间占用量大,并不适合大规模规则库。基于聚合位向量算法,提出一种新的改进算法,在不影响时间效率的基础上,通过改变算法中位图的存储方式,将聚合位图与位图交叉存储,忽略位图中全为0的部分,极大地减少了空间开销。最后,在仿真环境中对算法进行评测表明,该算法在大规模规则库中具有良好的时空效率。Aggregated bit vector is a fast scheme for flow classification,but it does not suit a large rule database because of the spending on space. By altering the way of storage,this paper presented a new algorithm. The new algorithm reduced the complexity of the storage by intersecting storage with aggregated bitmap and neglecting the part of zero,with maintaining the efficiency of time. In the end,realized the algorithm in an emulated environment and analyzed the scheme has good efficiency of time and space in a large rule database.

关 键 词：聚合位向量 流分类 位图 交叉存储 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]