Simultaneous Finite Automata: An Efficient Data-Parallel Model for Regular Expression Matching

TL;DR

This paper introduces simultaneous finite automata (SFA), a new automaton model that enables efficient parallel regular expression matching, achieving significant speedups on multicore processors by reducing simulation overhead.

Contribution

The paper proposes SFA, a novel automaton model that simplifies parallel implementation of automata, improving efficiency over traditional methods.

Findings

Achieved over 10-times speedup on dual hexa-core CPUs.

Developed a regular expression matcher based on SFA.

Demonstrated effective parallelism with low overheads.

Abstract

Automata play important roles in wide area of computing and the growth of multicores calls for their efficient parallel implementation. Though it is known in theory that we can perform the computation of a finite automaton in parallel by simulating transitions, its implementation has a large overhead due to the simulation. In this paper we propose a new automaton called simultaneous finite automaton (SFA) for efficient parallel computation of an automaton. The key idea is to extend an automaton so that it involves the simulation of transitions. Since an SFA itself has a good property of parallelism, we can develop easily a parallel implementation without overheads. We have implemented a regular expression matcher based on SFA, and it has achieved over 10-times speedups on an environment with dual hexa-core CPUs in a typical case.