• Proceedings of the IEEK Conference
• /
• 2004.06a
• /
• pp.47-50
• /
• 2004

In this paper, we implement an Elliptic Curve Cryptosystem(ECC) processor for DTCP. Because DTCP(Digital Transmission Content Protection) uses GF(p), where p is a 160-bit prime integer, we design a scalar multiplier based on GF(p). The scalar multiplier consists of a modular multiplier and an adder. The multiplier uses montgomery algorithm which is implemented with CSA(Carry-save Adder) and CLA(Carry-lookahead Adder). Our new scalar multiplier has been synthesized using Samsung 0.18 um CMOS technology and the maximum operation frequency is estimated 98 MHz, with the size about 65,000 gates. The resulting performance is 29.6 kbps, that is, it takes 5.4 msec to process a 160-bit data frame. We assure that this performance is enough to be used for digital signature, encryption/decryption, and key exchanges in real time environments.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.16 no.4
• /
• pp.33-41
• /
• 2006

RSA cryptosystem is of great use in systems such as IC card, mobile system, WPKI, electronic cash, SET, SSL and so on. RSA is performed through modular exponentiation. It is well known that the Montgomery multiplier is efficient in general. The critical path delay of the Montgomery multiplier depends on an addition of three operands, the problem that is taken over carry-propagation makes big influence at an efficiency of Montgomery Multiplier. Recently, the use of the Carry Save Adder(CSA) which has no carry propagation has worked McIvor et al. proposed a couple of Montgomery multiplication for an ideal exponentiation, the one and the other are made of 3 steps and 2 steps of CSA respectively. The latter one is more efficient than the first one in terms of the time complexity. In this paper, for faster operation than the latter one we use binary signed-digit(SD) number system which has no carry-propagation. We propose a new redundant binary adder(RBA) that performs the addition between two binary SD numbers and apply to Montgomery multiplier. Instead of the binary SD addition rule using in existing RBAs, we propose a new addition rule. And, we construct and simulate to the proposed adder using gates provided from SAMSUNG STD130 $0.18{\mu}m$ 1.8V CMOS Standard Cell Library. The result is faster by a minimum 12.46% in terms of the time complexity than McIvor's 2 method and existing RBAs.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.3
• /
• pp.715-722
• /
• 2010

This paper describes a hardware design of hash function processor which implements Korean Hash Algorithm Standard HAS-160. The HAS-160 processor compresses a message with arbitrary lengths into a hash code with a fixed length of 160-bit. To achieve high-speed operation with small-area, arithmetic operation for step-operation is implemented by using a hybrid structure of 5:3 and 3:2 carry-save adders and carry-select adder. It computes a 160-bit hash code from a message block of 512 bits in 82 clock cycles, and has 312 Mbps throughput at 50 MHz@3.3-V clock frequency. The designed HAS-160 processor is verified by FPGA implementation, and it has 17,600 gates on a layout area of about $1\;mm^2$ using a 0.35-${\mu}m$ CMOS cell library.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.1C
• /
• pp.131-139
• /
• 2008

This paper presents a scalable dual-field Montgomery multiplier based on a new multi-precision carry save adder(MP-CSA), which operates in both types of finite fields GF(p) and GF($2^m$). The new MP-CSA consists of two carry save adders(CSA). Each CSA is composed of n = [w/b] carry propagation adders(CPA) for a modular multiplication with w-bit words, where b is the number of dual field adders(DFA) in a CPA. The proposed Montgomery multiplier has roughly the same timing complexity compared with the previous result, however, it has the advantage of reduced chip area requirements. In addition, the proposed circuit produces the exact modular multiplication result at the end of operation unlike the previous architecture. Furthermore, the proposed Montgomery multiplier has a high scalability in terms of w and m. Therefore, it can be used to multiplier over GF(p) and GF($2^m$) for cryptographic applications.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.4
• /
• pp.223-230
• /
• 2006

The method of implementing a modular multiplier for Montgomery multiplication by using an adder depends on a selected adder. When using a CPA, there is a carry propagation problem. When using a CSA, it needs an additional calculation for a final result. The Multiplier using a Multi-precision CSA can solve both problems simultaneously by combining a CSA and a CPA. This paper presents an improved MP-CSA which reduces hardware resources and operation time by changing a MP-CSA's carry chain structure. Consequently, the proposed multiplier is more suitable for the module of long bit multiplication and exponentiation using a modular multiplier repeatedly.

• Proceedings of the Korea Society of Information Technology Applications Conference
• /
• 2001.11a
• /
• pp.199-199
• /
• 2001

• Proceedings of the IEEK Conference
• /
• 2001.06a
• /
• pp.337-340
• /
• 2001

This paper presented a new structure of RSA cryptosystem using modified Montgomery algorithm and CSA(Carry Save Adder) tree. Montgomery algorithm was modified to a radix-4 modified Booth algorithm. By appling radix-4 modified Booth algorithm and CSA tree to modular multiplication, a clock cycle for modular multiplication has been reduced to (n＋3)/2 and carry propagation has been removed from the cell structure of modular multiplier. That is, the connection efficiency of full adders is enhanced.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.04b
• /
• pp.837-840
• /
• 2002

본 논문에서는 해쉬 함수를 바탕으로 한 메시지 인증 코드 중의 하나인 MD5 를 하드웨어로 설계하였다. MD5 는 block-chained digest 알고리즘으로 64 단계의 동일한 단계 연산 구조를 가지므로 가장 기본적인 연산 한 단계를 구현하여 반복적으로 수행하는 구조로 설계하였다. 단계 연산구조 내에서는 연속된 32bit 덧셈 연산이 이루어지는데 기존의 CLA(carry-lookahead-adder)만을 사용하여 구현한 구조 대신 본 논문에서는 CSA(carry-save-adder)와 CLA 를 혼용하였다. 덧셈연산의 결과는 순서와 상관없기 때문에 연산자의 덧셈 순서를 리스케줄링 하였으며, 이는 기존의 CLA 만을 이용한 방법과 비교하여 최장지연 경로를 15% 줄여 훨씬 빠르게 연산을 수행하고, 전체 면적도 30%를 줄일 수 있었다. 결과적으로 본 논문에서 제안하는 구조는 지금까지 나온 어떤 MD5 프로세서 보다 작고 빠른 프로세서를 구현 할 수 있을 것으로 판단된다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04a
• /
• pp.661-663
• /
• 2002

현대의 Deep-Submicron Technology(DSM)에선 배선에 관련된 문제, 예를 들어 crosstalk이나 노이즈 등이 큰 문제가 된다. 그리하여, 배선은 논리 구성요소들보다 더욱 중요한 위치를 차지하게 되었다. 우리는 이러한 배선을 고려하여 연산식을 최적화하기 위해 carry-save-adder(CSA)를 이용한 모듈 함성 알고리즘을 제시한다. 즉, 상위 단계에서 생성 된 규칙적인 배선 토폴로지를 유지하며 CSA간의 배선을 좀더 향상시키는 최적의 알고리즘을 제안한다. 우리는 우리의 이러한 방법으로 생성된 지연시간이 [1]에 가깝거나 거의 근접하는 것을 많은 testcase에서 보이며(배선을 포함하지 않은 상태에서), 그리고 그와 동시에 최종 배선의 길이가 짧고 규칙적인 구조를 갖는것을 보인다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10a
• /
• pp.538-540
• /
• 2004

본 연구는 컴퓨터 연산을 위한 하드웨어 설계에서 고성능 연산에 사용되는 케리-세이브 가산기 (Carry-save adder) 합성에 관한 연구이다. 기존의 연구에서는, 연산 합성 문제와 합성된 연산의 배치 문제를 두개의 연속된 독립된 두개의 문제로 간주하고 풀었지만, 본 연구에서는 연산 합성 과정에서 연산 배치를 고려한 통합된 방법을 제시하여 전체적인 최적화된 결과를 얻었다. 연결선 상에서의 전력 소모나 지연시간이 점점 더 중요해지는 시스템-온-칩 (system-on-chip) 설계에서 본 연구의 통합적인 설계 방법은 매우 긴요하며 앞으로 효과적으로 이용될 수 있을 것이다.