当前位置: X-MOL 学术Concurr. Comput. Pract. Exp. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
AGFT: Adaptive entries aggregation scheme to prevent overflow in multiple flow table environment
Concurrency and Computation: Practice and Experience ( IF 1.5 ) Pub Date : 2021-07-13 , DOI: 10.1002/cpe.6491
Priyanka Nallusamy 1 , Reshmi TR 2 , Murugan Krishnan
Affiliation  

The revolutionary architecture termed Software-Defined Network provides flexible network management by detaching the control logic from the underlying data plane. The flow table resides in Ternary Content Addressable Memory imparts rules for the incoming flows with a limitation of high cost, limited storage, and consumes high power. In data center networks, when the traffic rate is high the overflow occurs due to storage limitation with high packet drop, frequent rule miss, and severe controller overhead. To overcome these challenges and to provide Quality of Service to the current network design sketch-based entry reduction scheme is proposed. It incorporates three concurrent modules integrated to function sequentially where (1) Periodical analysis in Multiple Flow Tables is performed to ensure the availability of redundant entries using the robust data mining algorithm called Term Frequency. (2) Recurrent entries are further classified and clustered using the Boyer–Moore pattern matching algorithm to facilitate the forthcoming aggregation process. (3) A compact flow table is achieved with a customized multibit trie using Huffman coding compression technique. The experimental outcomes prove that this work prevents overflow by 99.98% with 98.99% enhanced flow table space and provides a significant reduction of controller overhead than the existing schemes.

中文翻译:

AGFT:自适应表项聚合方案,防止多流表环境溢出

称为软件定义网络的革命性架构通过将控制逻辑与底层数据平面分离来提供灵活的网络管理。流表驻留在三元内容可寻址存储器中,为传入流传递规则,具有高成本、有限存储和高功耗的限制。在数据中心网络中,当流量率很高时,由于存储限制、丢包率高、规则缺失频繁和严重的控制器开销,会发生溢出。为了克服这些挑战并为当前的网络设计基于草图的条目减少方案提供服务质量,提出了。它合并了三个并发模块,它们按顺序运行,其中 (1) 执行多流表中的定期分析,以使用称为术语频率的稳健数据挖掘算法来确保冗余条目的可用性。(2) 使用 Boyer-Moore 模式匹配算法对循环条目进行进一步分类和聚类,以促进即将进行的聚合过程。(3) 使用霍夫曼编码压缩技术通过定制的多比特树实现紧凑的流表。实验结果证明,这项工作通过 98.99% 的增强流表空间防止了 99.98% 的溢出,并且比现有方案显着减少了控制器开销。(2) 使用 Boyer-Moore 模式匹配算法对循环条目进行进一步分类和聚类,以促进即将进行的聚合过程。(3) 使用霍夫曼编码压缩技术通过定制的多比特树实现紧凑的流表。实验结果证明,这项工作通过 98.99% 的增强流表空间防止了 99.98% 的溢出,并且比现有方案显着减少了控制器开销。(2) 使用 Boyer-Moore 模式匹配算法对循环条目进行进一步分类和聚类,以促进即将进行的聚合过程。(3) 使用霍夫曼编码压缩技术通过定制的多比特树实现紧凑的流表。实验结果证明,这项工作通过 98.99% 的增强流表空间防止了 99.98% 的溢出,并且比现有方案显着减少了控制器开销。
更新日期:2021-07-13
down
wechat
bug