DocumentCode :
71758
Title :
An Efficient and Scalable Semiconductor Architecture for Parallel Automata Processing
Author :
Dlugosch, Paul ; Brown, Dean ; Glendenning, Paul ; Leventhal, Michael ; Noyes, Harold
Author_Institution :
Micron Technol., DRAM Solutions Group, Boise, ID, USA
Volume :
25
Issue :
12
fYear :
2014
fDate :
Dec. 2014
Firstpage :
3088
Lastpage :
3098
Abstract :
We present the design and development of the automata processor, a massively parallel non-von Neumann semiconductor architecture that is purpose-built for automata processing. This architecture can directly implement non-deterministic finite automata in hardware and can be used to implement complex regular expressions, as well as other types of automata which cannot be expressed as regular expressions. We demonstrate that this architecture exceeds the capabilities of high-performance FPGA-based implementations of regular expression processors. We report on the development of an XML-based language for describing automata for easy compilation targeted to the hardware. The automata processor can be effectively utilized in a diverse array of applications driven by pattern matching, such as cyber security and computational biology.
Keywords :
XML; field programmable gate arrays; finite automata; parallel processing; pattern matching; XML-based language; automata processor; complex regular expressions; computational biology; cyber security; hardware; high performance FPGA; nondeterministic finite automata; parallel automata processing; parallel non-von Neumann semiconductor architecture; pattern matching; regular expression processors; scalable semiconductor architecture; Arrays; Automata; Complexity theory; Hardware; Radiation detectors; Routing; Automata; accelerator architectures; hardware; high performance computing; parallel architectures; reconfigurable architectures;
fLanguage :
English
Journal_Title :
Parallel and Distributed Systems, IEEE Transactions on
Publisher :
ieee
ISSN :
1045-9219
Type :
jour
DOI :
10.1109/TPDS.2014.8
Filename :
6719386
Link To Document :
بازگشت