Wire Speed Packet Classication Without TCAMs A Few

  • Slides: 16
Download presentation
Wire Speed Packet Classification Without TCAMs: A Few More Registers (And A Bit of

Wire Speed Packet Classification Without TCAMs: A Few More Registers (And A Bit of Logic) Are Enough Author: Qunfeng Dong, Suman Banerjee, Jia Wang, Dheeraj Agrawal Publisher: SIGMETRICS’ 07 Presenter: Fang-Chen, Kuo 1

Motivation Cache Rules ! Evolving rules 2

Motivation Cache Rules ! Evolving rules 2

Cache Evolving rules n n Hit ratio ↑ with less cache entries. Without using

Cache Evolving rules n n Hit ratio ↑ with less cache entries. Without using TCAM 3

Architecture 4

Architecture 4

Classification Scheme (SPDD) 5

Classification Scheme (SPDD) 5

Cache Manager (Regular Hyper-Cube List) 6

Cache Manager (Regular Hyper-Cube List) 6

Rule Expanding 8

Rule Expanding 8

Hardware Design of Rule Cache 9

Hardware Design of Rule Cache 9

OPTIMIZATION (1/2) - PPDD (Pruned Packet decision Diagram) 10

OPTIMIZATION (1/2) - PPDD (Pruned Packet decision Diagram) 10

OPTIMIZATION (2/2) – Ordering Packet Header Fields 11

OPTIMIZATION (2/2) – Ordering Packet Header Fields 11

Evaluation (1/6) n n We evaluate the performance of our smart rule cache using

Evaluation (1/6) n n We evaluate the performance of our smart rule cache using 4 real traffic traces and 10 real rule sets obtained from a tier-1 ISP backbone network. The decision of rules is either permit or deny. 12

Evaluation (2/6)Memory Requirement The default ordering is: (1) source IP address; (2) destination IP

Evaluation (2/6)Memory Requirement The default ordering is: (1) source IP address; (2) destination IP address; (3) source port; (4) destination port; (5) protocol type. The best ordering is: (1) protocol type; (2) source IP address; (2) destination IP address; (3) source port; (4) destination port. 13

Evaluation (3/6) 14

Evaluation (3/6) 14

Evaluation (4/6) – Cache Miss Ratio 15

Evaluation (4/6) – Cache Miss Ratio 15

Evaluation (5/6) – Sampling Interval 16

Evaluation (5/6) – Sampling Interval 16

Evaluation (6/6) – Result for Complicated Rule Sets 17

Evaluation (6/6) – Result for Complicated Rule Sets 17