Copyright © 2006 The Institute of Electronics, Information and Communication Engineers
Special Section on Cryptography and Information Security -- Papers -- Elliptic Curve Cryptography |
Radix-r Non-Adjacent Form and Its Application to Pairing-Based Cryptosystem*
1 The author is with the School of Systems Information Science, Future University-Hakodate, Hakodate-shi, 041-8655 Japan. E-mail: takagi{at}fun.ac.jp, 2 The author is with the School of Electrical and Computer Engineering, State University of Campinas, Caixa Postal 6101, Brazil. E-mail: davidjr{at}dca.fee.unicamp.br, 3 The authors are with the Laboratory of Cryptography and Information Security (LCIS), Department of Computer Science and Information Engineering, National Central University, Taiwan 320, R.O.C. E-mail: yensm{at}csie.ncu.edu.tw, E-mail: wubq{at}csie.ncu.edu.tw
Recently, the radix-3 representation of integers is used for the efficient implementation of pairing based cryptosystems. In this paper, we propose non-adjacent form of radix-r representation (rNAF) and efficient algorithms for generating rNAF. The number of non-trivial digits is (r 2)(r + 1)/2 and its average density of non-zero digit is asymptotically (r 1)/(2r 1). For r = 3, the non-trivial digits are {± 2, ± 4} and the non-zero density is 0.4. We then investigate the width-w version of rNAF for the general radix-r representation, which is a natural extension of the width-w NAF. Finally we compare the proposed algorithms with the generalized NAF (gNAF) discussed by Joye and Yen. The proposed scheme requires a larger table but its non-zero density is smaller even for large radix. We explain that gNAF is a simple degeneration of rNAFwe can consider that rNAF is a canonical form for the radix-r representation. Therefore, rNAF is a good alternative to gNAF.
Key Words: non-adjacent form, radix-r representation, signed window method, elliptic curve cryptosystem, pairing based cryptosystem
Manuscript received March 7, 2005. Manuscript revised June 24, 2005. Final manuscript received September 5, 2005.
* The preliminary version of this paper was presented at the 7th Information Security Conference (ISC 2004), held in Palo Alto.