Hardware-based implementation of the common approximate substring algorithm

Abstract

An implementation of an algorithm for string matching, commonly used in DNA string analysis, using configurable technology is proposed. The design of the circuit allows for pipelining to provide a performance increase. The proposal is unique in that we suggest a design that is specific to certain parameters of the problem, but may be reused for any particular instance of the problem that matches these parameters. The use of a field programmable gate array allows the implementation to be instance specific, thus ensuring maximal usage of the hardware. Analysis and preliminary results based on a prototype implementation are presented.