一种高效的OpenFlow流表存储与查找实现方法  

An efficient implementation of storage and lookup for flow tables in OpenFlow

作　　者：鄂跃鹏[1] 陈智[1] 葛敬国[1] 吴玉磊[1] 

机构地区：[1]中国科学院信息工程研究所信息安全国家重点实验室,北京100093

出　　处：《中国科学：信息科学》2015年第10期1280-1288,共9页Scientia Sinica(Informationis)

基　　金：中国科学院战略性先导专项(批准号:XDA06010306);国家重点基础研究发展计划(973计划)(批准号:2012CB315803);国家自然科学基金青年基金(批准号:61303241);国家高技术研究发展计划(863计划)(批准号:2013AA013501)资助项目

摘　　要：目前Open Flow流表的存储与查找还面临一些挑战,如流表项匹配字段数量的不断增加、流表规模的爆炸性增长以及实现Open Flow硬件交换机的端口线速处理更加困难.针对这些问题,本文提出了一种高效的Open Flow流表存储与查找实现方法.该方法将Open Flow流表划分成多个规模更小、匹配字段更少的子流表,进一步根据字段匹配类型将子流表切分成多个字段,不同字段采用不同的算法和硬件进行存储和查找,得到字段匹配结果后再结合汇总表查找从而得到子流表的匹配结果.通过划分流表与切分子流表来压缩流表存储空间,根据子流表的查找选择和子流表内部查找优化来减少查找过程中硬件资源开销.通过大量模拟实验表明,该存储与查找方法能够压缩50%左右的存储空间,且可以有效减少流表项匹配过程中80%硬件资源的消耗.The storage and lookup of Open Flow flow tables still faces many challenges such as the increasing number of fields in flow inlets, the explosive growth of flow scales, and difficulties in achieving wired-speed processing of ports in Open Flow switches. To solve these problems, this paper presents an efficient storage and lookup method for flow tables in Open Flow. This method divides flow tables into multiple sub-flow tables of smaller scale and fewer matching fields. Based on the different types of matching fields, sub-flow tables are split into multiple slices, which can be handled by different algorithms and hardware. The matching result of sub-flow tables can be obtained using the matching results of slices in sub-flow and summary tables. By splitting the flow and sub-flow tables, this method compresses the storage space of the flow table in Open Flow switches. By optimizing of subflow table selection and lookup processes, this method reduces the cost of hardware resources when executing the lookup process. Extensive simulation experiments show that the proposed method can compress storage space by about 50% and can reduce hardware resource consumption to 80% when executing matching processes.

关 键 词：OpenFlow流表 存储与查找 子流表 子流表切分 汇总表 

分 类 号：TP333[自动化与计算机技术—计算机系统结构]