一种高效的OpenFlow流表存储与查找实现方法

发布时间：2018-10-20 08:04

【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.
[Abstract]:At present, the storage and search of Open Flow stream table still face some challenges, such as the increasing number of matching fields, the explosive growth of stream table scale and the difficulty of implementing port line speed processing of Open Flow hardware switch. In order to solve these problems, this paper proposes an efficient Open Flow stream table storage and lookup method. In this method, the Open Flow stream table is divided into several sub-stream tables with smaller scale and fewer matching fields. The sub-stream table is further divided into multiple fields according to the matching type of the fields. Different fields are stored and searched by different algorithms and hardware. The matching result of the substream table is obtained by combining the field matching result with the summary table lookup. The storage space of the stream table is compressed by dividing the stream table and the tangent molecular flow table, and the hardware resource overhead is reduced according to the search selection of the sub-stream table and the optimization of the internal search of the sub-flow table. A large number of simulation experiments show that the proposed method can compress about 50% of the storage space and reduce the consumption of 80% hardware resources in the matching process of stream table items.
【作者单位】：中国科学院信息工程研究所信息安全国家重点实验室;
【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
【分类号】：TP333

【参考文献】