An RNS-enabled microprocessor for public key cryptography.
Date
2010
Authors
Lim, Zhining
Editors
Advisors
Phillips, Braden Jace
Liebelt, Michael J.
Liebelt, Michael J.
Journal Title
Journal ISSN
Volume Title
Type:
Thesis
Citation
Statement of Responsibility
Conference Name
Abstract
Mobile computing platforms must be secure, fast, small, and power efficient. While public key cryptography methods provide the framework for secure communications, their algorithms tend to be computationally complex and hence need some form of hardware acceleration. This thesis investigates an alternative implementation of public key cryptology algorithms on amicroprocessor that has been augmented with hardware support for the residue number system (RNS).
The RNS promises efficient arithmetic by replacing the long integers used in public key cryptology with sets of smaller independent numbers. Doing so allows flexibility in instruction scheduling and reduces carry propagation delays. Past RNS public key cryptosystem implementations used a parallel architecture.
Although well-suited to the RNS, a parallel architecture is not necessarily power efficient nor small since it requires several copies of the RNS hardware. The microprocessor implementation in this thesis has a novel sequential architecture that has modest area and power consumption because it uses just one set of the RNS hardware.
The RNS-enhanced microprocessor, the Z-Core, is based on the small, lowpower Xtensa LX2.1 core from Tensilica, Inc. It was augmented with RNS hardware specified using the Tensilica Instruction Extension language. Unlike other cryptographic accelerators, these RNS enhancements are useful for a variety of
cryptosystems. The Z-Core was used for implementations of RSA and the elliptic curve digital signature algorithm. The RNS implementations on the Z-Core outperformed equivalent implementations that use the normal multiprecision methods.
The area and power consumption of the Z-Core were within the GSM 11.18 1.8-V SIM specifications for smart cards when running the two algorithms, hence it satisfied the requirements of mobile computing.
School/Discipline
School of Electrical and Electronic Engineering
Dissertation Note
Thesis (Ph.D.) -- University of Adelaide, School of Electrical and Electronic Engineering, 2010
Provenance
Copyright material removed from digital thesis. See print copy in University of Adelaide Library for full text.