• ETRI Journal
• /
• v.26 no.3
• /
• pp.241-251
• /
• 2004

We propose two improved scalar multiplication methods on elliptic curves over $F_{{q}^{n}}$ $q= 2^{m}$ using Frobenius expansion. The scalar multiplication of elliptic curves defined over subfield $F_q$ can be sped up by Frobenius expansion. Previous methods are restricted to the case of a small m. However, when m is small, it is hard to find curves having good cryptographic properties. Our methods are suitable for curves defined over medium-sized fields, that is, $10{\leq}m{\leq}20$. These methods are variants of the conventional multiple-base binary (MBB) method combined with the window method. One of our methods is for a polynomial basis representation with software implementation, and the other is for a normal basis representation with hardware implementation. Our software experiment shows that it is about 10% faster than the MBB method, which also uses Frobenius expansion, and about 20% faster than the Montgomery method, which is the fastest general method in polynomial basis implementation.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.11 no.6
• /
• pp.105-113
• /
• 2001

Recently, Gallant, Lambert arid Vanstone introduced a method for speeding up the scalar multiplication on a family of elliptic curves over prime fields that have efficiently-computable endomorphisms. It really depends on decomposing an integral scalar in terms of an integer eigenvalue of the characteristic polynomial of such an endomorphism. In this paper, by using an element in the endomorphism ring of such an elliptic curve, we present an alternate method for decomposing a scalar. The proposed algorithm is more efficient than that of Gallant\`s and an upper bound on the lengths of the components is explicitly given.

• Proceedings of the Korea Institutes of Information Security and Cryptology Conference
• /
• 2003.12a
• /
• pp.412-416
• /
• 2003

This paper introduces a new joint signed expansion method for computing simultaneous scalar multiplication on an elliptic curve and a modified binary algorithm for efficient use of the new expansion method. The proposed expansion method can be also be used in cryptosystems such as RSA and EIGamal cryptosystems.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• v.9 no.1
• /
• pp.931-934
• /
• 2005

The scalar point multiplication operations is one of the most time-consuming components in elliptic curve cryptosystems. In this paper, we suggest how to induce the point-quadruple (4Q) operation by improving the double-and-add method, which has been a prevailing computing method for calculating the result of a scalar point multiplication. Induced and drived numerical expressions were evaluated and verified by a real application using C programming language. The induced algorithm can be applied to a various kind of calculations in elliptic curve operations more efficiently and by a faster implementation.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.12 no.5
• /
• pp.45-51
• /
• 2002

As Koblitz curve, the Frobenius endomorphism is know to be useful in efficient implementation of multiplication on non-supersingular elliptic cures defined on small finite fields of characteristic two. In this paper a method using the extended Frobenius endomorphism to speed up scalar multiplication is introduced. It will be shown that the proposed method is more efficient than Muller's block method in ［5］ because the number of point addition for precomputation is small but on the other hand the expansion length is almost same.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.16 no.6
• /
• pp.95-109
• /
• 2006

Elliptic Curve Cryptosystem (ECC) is suitable for resource-constrained devices such as smartcards, and sensor motes because of its short key size. This paper presents an efficient multi-scalar multiplication algorithm which is the main component of the verification procedure in Elliptic Curve Digital Signature Algorithm (ECDSA). The proposed algorithm can make use of a precomputed table of variable size and provides an optimal efficiency for that precomputed table. Furthermore, the given scalar is receded on-the-fly so that it can be merged with the main multiplication procedure. This can achieve more savings on memory than other receding algorithms. Through experiments, we have found that the optimal sizes of precomputed tables are 7 and 15 when uP+vQ is computed for u, v of 163 bits and 233 bits integers. This is shown by comparing the computation time taken by the proposed algorithm and other existing algorithms.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.9 no.1
• /
• pp.85-102
• /
• 1999

In this paper, we propose an EC-SRP(Elliptic Curve - Secure Remote Password) protocol that uses ECDLP(Elliptic Curve Discrete Logarithm Problem) instead SRP protocols’s DLP. Since EC-SRP uses ECDLP, it inherits the high performance and security those are the properties of elliptic curve. And we reduced the number of elliptic curve scalar multiplication to improve EC-SRP protocol’s performance. Also we have proved BC-SRP protocol is a secure AKC(Authenticated Key Agreement with Key Confirmation) protocol in a random oracle model.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.4
• /
• pp.548-555
• /
• 2022

This paper describes the hardware implementation of elliptic curve cryptography (ECC) used as a core operation in elliptic curve digital signature algorithm (ECDSA). The ECC processor supports eight operation modes (four point operations, four modular operations) on the NIST P-521 curve. In order to minimize computation complexity required for point scalar multiplication (PSM), the radix-4 Booth encoding scheme and modified Jacobian coordinate system were adopted, which was based on the complexity analysis for five PSM algorithms and four different coordinate systems. Modular multiplication was implemented using a modified 3-Way Toom-Cook multiplication and a modified fast reduction algorithm. The ECC processor was implemented on xczu7ev FPGA device to verify hardware operation. Hardware resources of 101,921 LUTs, 18,357 flip-flops and 101 DSP blocks were used, and it was evaluated that about 370 PSM operations per second were achieved at a maximum operation clock frequency of 45 MHz.

• Proceedings of the IEEK Conference
• /
• 2002.06b
• /
• pp.345-348
• /
• 2002

This paper presents new fast scalar multiplier of elliptic curve cryptosystem that is regarded as next generation public-key crypto processor. For fast operation of scalar multiplication a finite field multiplier is designed with LFSR type of bit serial structure and a finite field inversion operator uses extended binary euclidean algorithm for reducing one multiplying operation on point operation. Also the use of the window non-adjacent form (WNAF) method can reduce addition operation of each other different points.

• ETRI Journal
• /
• v.21 no.1
• /
• pp.28-39
• /
• 1999

Koblitz has suggested to use "anomalous" elliptic curves defined over ${\mathbb{F}}_2$, which are non-supersingular and allow or efficient multiplication of a point by and integer, For these curves, Meier and Staffelbach gave a method to find a polynomial of the Frobenius map corresponding to a given multiplier. Muller generalized their method to arbitrary non-supersingular elliptic curves defined over a small field of characteristic 2. in this paper, we propose an algorithm to speed up scalar multiplication on an elliptic curve defined over a small field. The proposed algorithm uses the same field. The proposed algorithm uses the same technique as Muller's to get an expansion by the Frobenius map, but its expansion length is half of Muller's due to the reduction step (Algorithm 1). Also, it uses a more efficient algorithm (Algorithm 3) to perform multiplication using the Frobenius expansion. Consequently, the proposed algorithm is two times faster than Muller's. Moreover, it can be applied to an elliptic curve defined over a finite field with odd characteristic and does not require any precomputation or additional memory.