442 0

ASIC에서 진성난수 생성기의 구현 및 난수성 시험

Title
ASIC에서 진성난수 생성기의 구현 및 난수성 시험
Other Titles
Implementation and Randomness test of True Random Number Generator in ASIC
Author
김영준
Alternative Author(s)
Kim, Young-Jun
Advisor(s)
김동규
Issue Date
2013-02
Publisher
한양대학교
Degree
Master
Abstract
보안 시스템에서 난수는 시스템의 안전성과 밀접한 관계가 있는 중요한 요소이다. 따라서 보안 시스템에서 사용하는 난수가 예측가능하다면 보안 시스템의 안전성은 보장 받지 못할 것이다. 의사난수 생성기(PRNG, Pseudo Random Number Generator)의 경우 초기 값인 시드 값의 엔트로피가 높아야 안전성을 보장 받을 수 있다. 하지만 결정적인 시스템에서는 완전한 랜덤 소스를 발생하는 것이 불가능하기 때문에 의사난수 생성기는 안전하다고 보기 힘들다. 그에 반해 진성난수 생성기(TRNG, True Random Number Generator)는 비예측적인 물리소스를 사용하여 거의 완전한 난수를 생성한다. 하지만 진성난수 생성기는 하드웨어로 구현되어 유연성이 낮기 때문에 난수 생성기의 성능과 난수성이 보장되어야 한다. 진성난수 생성기는 물리소스의 구현 방법에 따라 아날로그 구현과 디지털 구현으로 나눌 수 있다. 물리소스를 아날로그로 구현할 경우 엔트로피가 높지만 디지털 구현에 비해 처리량도 적고 마이크로프로세서 같은 디지털회로들과 통합하기 힘들다. 따라서 순수 디지털 환경에서 작업하는 디자이너들은 디지털 구현의 난수 생성기를 선호한다. 초경량 진성난수 생성기(Ultra-lightweight TRNG)는 순수 디지털로 구현 가능한 난수 생성기 중 하나로 적은 하드웨어 리소스를 사용한다. FPGA에서 구현되어 난수성이 테스트 되었지만 아직 ASIC에서는 구현된 적이 없다. 본 논문에서는 초경량 진성난수 생성기를 FPGA 및 ASIC으로 구현하고 난수성을 테스트하였다. 성능을 객관적으로 분석하기 위해 현재 많이 사용되고 초경량 진성난수 생성기와 같은 물리소스를 사용하는 히타치 진성넘버 생성기(Hitachi TRNG)를 FPGA로 구현하고 난수성을 테스트하였다. 같은 환경에서 같은 리소스로 구현하였을 경우 초경량 진성난수 생성기가 히타치 진성난수 생성기에 비해 높은 난수성을 보였다. ASIC으로 구현한 초경량 진성난수 생성기의 경우 FPGA로 구현했을 때에 비해 샘플링 속도가 1/10으로 떨어졌지만 난수성이 보장되었다.|Random number is an important factor which closely related to the safety in security systems. Therefore, if random number used in the security system is predictable, its safety will not be guaranteed. Pseudo-random number generators (PRNGs) guarantee safety of system when the initial value, the seed value, have a high entropy. But PRNGs are not safe, because it is impossible to generates random source in deterministic systems. Whereas true random number generators (TRNGs) employ unpredictable physical source, and generate the real random number. But TRNGs implemented by hardware guarantee their performance and randomness because of the less flexibility. TRNGs can be divide by implementation method of the physical source, Analog and digital. If physical source is implemented by analog, TRNGs will have high entropy. But analog TRNGs are hard to integrate with digital circuits such as microprocessors and its throughput is lower than digital ones. So digital TRNGs are attractive circuits for designers who work on purely digital environment. Ultra-lightweight TRNG is one of the digital TRNGs with low hardware resource. This TRNG are implemented in FPGA and tested randomness, but not yet in ASIC. In this paper, we implemented ultra-lightweight TRNG in FPGA and ASIC, and tested the performance and randomness. For objective analysis, we also implemented hitachi TRNG, which well used in public and same physical source of ultra-lightweight TRNG, in FPGA and tested randomness. Ultra-lightweight TRNG is more randomness than hitachi ones in same environment and resource. TRNG implemented in ASIC is ten times slow throughput rates than FPGA ones. But its performance guaranteed randomness.; Random number is an important factor which closely related to the safety in security systems. Therefore, if random number used in the security system is predictable, its safety will not be guaranteed. Pseudo-random number generators (PRNGs) guarantee safety of system when the initial value, the seed value, have a high entropy. But PRNGs are not safe, because it is impossible to generates random source in deterministic systems. Whereas true random number generators (TRNGs) employ unpredictable physical source, and generate the real random number. But TRNGs implemented by hardware guarantee their performance and randomness because of the less flexibility. TRNGs can be divide by implementation method of the physical source, Analog and digital. If physical source is implemented by analog, TRNGs will have high entropy. But analog TRNGs are hard to integrate with digital circuits such as microprocessors and its throughput is lower than digital ones. So digital TRNGs are attractive circuits for designers who work on purely digital environment. Ultra-lightweight TRNG is one of the digital TRNGs with low hardware resource. This TRNG are implemented in FPGA and tested randomness, but not yet in ASIC. In this paper, we implemented ultra-lightweight TRNG in FPGA and ASIC, and tested the performance and randomness. For objective analysis, we also implemented hitachi TRNG, which well used in public and same physical source of ultra-lightweight TRNG, in FPGA and tested randomness. Ultra-lightweight TRNG is more randomness than hitachi ones in same environment and resource. TRNG implemented in ASIC is ten times slow throughput rates than FPGA ones. But its performance guaranteed randomness.
URI
https://repository.hanyang.ac.kr/handle/20.500.11754/133503http://hanyang.dcollection.net/common/orgView/200000420796
Appears in Collections:
GRADUATE SCHOOL[S](대학원) > ELECTRONICS AND COMPUTER ENGINEERING(전자컴퓨터통신공학과) > Theses (Master)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE