Summary: OC-3072 Packet Classification Using BDDs and Pipelined SRAMs
Amit Prakash Adnan Aziz
Department of Electrical and Computer Engineering
The University of Texas at Austin
We present a solution to the problem of quickly classi-
fying packets. Our approach is based on techniques from
logic synthesis. Specifically, we express the classification
rules as Boolean logic equations, build Binary Decision Di-
agrams for these equations, and then map the BDDs to a
logic network consisting of a pipeline of static RAM banks.
We illustrate our approach by applying it to the longest pre-
fix matching for IP forwarding, and present evidence that
our scheme can perform a billion matches per second on a
CAIDA backbone forwarding table containing 60,000 pre-
fixes. We show how our approach generalizes to classifying
packets on multiple fields.