CubeCuts: A novel cutting scheme for packet classification

Yeim-Kuan Chang, Yu Hsiang Wang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Citations (Scopus)

Abstract

Packet Classification is one of the most important services provided by Internet routers nowadays. Decision-tree based schemes, such as HiCuts and Hyper Cuts, are the most well-known algorithms. These schemes separate search space into many equal-sized sub-spaces. But both schemes have the same rule replication problem, which might cause large memory overhead. Although another decision-tree based solution, Hyper splits, was proposed to cut space according to end-points for reducing the memory requirement, we still observe that its rule replication problem doesnt be solved well and the memory requirement can still be improved. In this paper, we propose a scheme called Cube Cuts to build a binary decision tree that does not generate any duplicated rule. By using the hybrid scheme that combines the Cube Cuts and constrained HiCuts, we can have a memory-efficient data structure such that the entire rule table of up to10K rules can be fit into the on-chip memory of FPGA device. Our design is very suitable to be implemented with parallelism and pipeline. The experimental results show that the rule replication ratio is very low in all rule tables (ACL, FW, and IPC). The proposed parallel and pipelined architecture based on the hybrid scheme can achieve a throughput of 118 Gbps, which is enough to deal with the Internet traffic that is growing rapidly in recent years.

Original languageEnglish
Title of host publicationProceedings - 26th IEEE International Conference on Advanced Information Networking and Applications Workshops, WAINA 2012
Pages274-279
Number of pages6
DOIs
Publication statusPublished - 2012 May 14
Event26th IEEE International Conference on Advanced Information Networking and Applications Workshops, WAINA 2012 - Fukuoka, Japan
Duration: 2012 Mar 262012 Mar 29

Publication series

NameProceedings - 26th IEEE International Conference on Advanced Information Networking and Applications Workshops, WAINA 2012

Other

Other26th IEEE International Conference on Advanced Information Networking and Applications Workshops, WAINA 2012
CountryJapan
CityFukuoka
Period12-03-2612-03-29

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Information Systems

Fingerprint Dive into the research topics of 'CubeCuts: A novel cutting scheme for packet classification'. Together they form a unique fingerprint.

Cite this