In view of current multiple regular expressions matching algorithms consume too large system resource

but provide low throughput

a new regular expression matching algorithm for network context is proposed through analyzing typical regular expression matching algorithms. The algorithm separates expressions that have interaction and could cause exponential increase in the number of states

and assigns them into different groups. Then a DFA is constructed for each group

and these DFAs are placed on the splay tree to schedule by their matching probabilities. Simulation results show that the proposed algorithm not only could save more memory space

but also could provide about three times higher throughput.