TY - GEN
T1 - High-speed packet classification using binary search on length
AU - Lim, Hyesook
AU - Mun, Ju Hyoung
PY - 2007
Y1 - 2007
N2 - Packet classification is one of the major challenges for next generation routers since it involves complicated multi-dimensional search as well as it should be performed in wire-speed for all incoming packets. Area-based quad-trie is an excellent algorithm in the sense that it constructs a two-dimensional trie using source and destination prefix fields for packet classification. However, it does not achieve good search performance since search is linearly performed for prefix length. In this paper, we propose a new packet classification algorithm which applies binary search on prefix length to the area-based quad-trie. In order to avoid the pre-computation required in the binary search on length, the proposed algorithm constructs multiple disjoint tries depending on relative levels in rule hierarchy. We also propose two new optimization techniques considering rule priorities. For different types of rule sets having about 5000 rules, performance evaluation result shows that the average number of memory accesses is 18 to 67 and the memory consumption is 22 to 41 bytes per rule.
AB - Packet classification is one of the major challenges for next generation routers since it involves complicated multi-dimensional search as well as it should be performed in wire-speed for all incoming packets. Area-based quad-trie is an excellent algorithm in the sense that it constructs a two-dimensional trie using source and destination prefix fields for packet classification. However, it does not achieve good search performance since search is linearly performed for prefix length. In this paper, we propose a new packet classification algorithm which applies binary search on prefix length to the area-based quad-trie. In order to avoid the pre-computation required in the binary search on length, the proposed algorithm constructs multiple disjoint tries depending on relative levels in rule hierarchy. We also propose two new optimization techniques considering rule priorities. For different types of rule sets having about 5000 rules, performance evaluation result shows that the average number of memory accesses is 18 to 67 and the memory consumption is 22 to 41 bytes per rule.
KW - area-based quad-trie
KW - best matching rule
KW - binary search on length
KW - hashing
KW - packet classification
UR - http://www.scopus.com/inward/record.url?scp=58049131273&partnerID=8YFLogxK
U2 - 10.1145/1323548.1323572
DO - 10.1145/1323548.1323572
M3 - Conference contribution
AN - SCOPUS:58049131273
SN - 9781595939456
T3 - ANCS'07 - Proceedings of the 2007 ACM Symposium on Architecture for Networking and Communications
SP - 137
EP - 144
BT - ANCS'07 - Proceedings of the 2007 ACM Symposium on Architecture for Networking and Communications
T2 - 3rd ACM/IEEE Symposium on Architectures for Networking and Communications Systems, ANCS 2007
Y2 - 3 December 2007 through 4 December 2007
ER -