Earticle

현대의 인공지능은 사회를 구성하는 필수적인 기술로 여겨지고 있다. 특히, 인공지능에서 프라이버시 침해 문제는 현대 사 회에서 심각한 문제로 자리 잡고 있다. 개인정보보호를 위해 2019년 MIT에서 제안된 분할 학습은 연합 학습의 기술 중 하나 로 개인정보보호 효과를 지닌다. 본 연구에서는 데이터를 안전하게 관리하기 위해 알려진 차분 프라이버시를 이용하여 안전하 고 정확한 분할 학습 모델을 연구한다. 또한, SVHN과 GTSRB 데이터 세트를 15가지의 차등적인 차분 프라이버시를 적용한 분할 학습 모델에 학습시키고 학습이 안정적으로 되는지를 확인한다. 최종적으로, 학습 데이터 추출 공격을 진행하여, 공격을 예방하는 차분 프라이버시 예산을 MSE를 통해 정량적으로 도출한다.

Recently, artificial intelligence is regarded as an essential technology in our society. In particular, the invasion of privac y in artificial intelligence has become a serious problem in modern society. Split learning, proposed at MIT in 2019 for pri vacy protection, is a type of federated learning technique that does not share any raw data. In this study, we studied a sa fe and accurate segmentation learning model using known differential privacy to safely manage data. In addition, we traine d SVHN and GTSRB on a split learning model to which 15 different types of differential privacy are applied, and checked whether the learning is stable. By conducting a learning data extraction attack, a differential privacy budget that prevents attacks is quantitatively derived through MSE.

니콜라 테슬라에 의해 항공기의 무선제어 가능성이 제시되면서 출현한 무인항공기는 제 1, 2차 세계대전을 통해 항공력의 급속한 발전과 함께 군사, 방산용으로 사용하게 되었다. 2000년대, 무인항공기의 분야가 촬영, 배송, 통신 등 민간분야까지 확대 됨에 따라 여러 서비스와 융합되어 활용되고 있다. 하지만, 최근 무인항공기 시스템에서의 통신이나 무인항공기 자체의 보안 취 약점을 이용하여 GPS 스푸핑, 전파 교란 공격 등을 시도하는 보안사고가 발생하고 있다. 이에, 안전한 무인항공기의 도입을 위 하여 국내에서는 자체 무인항공기 검증 제도인 감항 인증 제도가 마련되었다. 그러나 감항 인증 제도는 무인항공기의 보안성보 다는 시험 비행, 설계 및 물리적 구조의 안전성과 인증하는 쪽에 초점이 맞추어져 있다. 보안성 높은 안전한 무인항공기의 도입 을 위해 본 논문에서는 무인항공기 시스템 모델을 제안하고 데이터 흐름도를 작성하였다. 작성한 데이터 흐름도를 바탕으로 무 인항공기 시스템에서의 위협을 도출하였고, 도출한 위협을 방지할 수 있는 보안기능요구사항을 개발하였다. 제안한 보안기능요 구사항을 통해 향후 무인항공기의 안전한 도입을 위한 앞으로의 평가, 검증 기술의 발전 방향을 제시한다

With the possibility of wireless control of the aircraft by Nicola Tesla, Unmanned Aerial Vehicle(UAV) was mainly used for military an d defense purposes with the rapid development through World War I and II. As civilian applications of unmanned aerial vehicles have exp anded, they have been used with various services, and attempts have been made to control various environmental changes and risk factor s of unmanned aerial vehicles. However, GPS spoofing, Jamming attack and security accidents are occurring due to the communication in the unmaned aerial vehicle system or the security vulnerability of the unmanned aerial vehicle itself. In order to secure introduction of Un manned aerial vehicle, South Korea has established Unmanned Aerial Vehicle verification system called Airworthiness Certification. Howe ver, the existing cerfication system is more focused on test flight, design and structure's safety and reliability. In this paper, we propose a unmanned aerial vehicle system model and propose security functional requirements on unmanned aerial vehicle system in the correspon ding system model for secure-introduction of Unmanned Aerial Vehicle. We suggest the development direction of verification technology. From this proposal, future development directions of evaluation and verification technology of Unmanned Aerial Vehicle will be presented.

블록체인은 중앙신뢰 기관의 개입 없이 분산 컴퓨팅 환경에서 데이터를 관리하는 기술이다. 블록체인의 보안성, 효율 성, 응용성으로 인하여 현재 금융 분야뿐만 아니라 제조, 문화, 공공 등 다양한 분야에서 블록체인 기술이 활용되고 있 다. 그동안 블록체인에서 공격자는 51% 이상의 해시 파워를 갖출 수 없다고 여겨졌지만 최근 이에 대한 공격과 피해사 례가 발생하고 있으며, 이기적인 채굴자 공격을 포함한 대용량 컴퓨팅 능력을 갖춘 공격의 빈도가 증가하고 있다. 또한, 일반 컴퓨터와 차원이 다른 성능을 발휘하는 양자컴퓨터의 발전은 블록체인의 새로운 위협이 되고 있다. 본 논문에서는 블록체인 특징과 합의 알고리즘에 대해 소개하고 컴퓨팅 연산력을 이용한 블록체인 공격기법을 설명한다. 그리고, 대용 량 컴퓨팅 환경 구축방법과 양자 컴퓨터를 사용하는 공격 알고리즘이 블록체인 보안성에 미치는 영향을 분석한다. 마지 막으로, 블록체인의 보안성을 향상하기 위한 대용량 컴퓨팅 공격 보호 기술 및 앞으로의 발전 방향을 제시한다.

The blockchain is a technique for managing transaction data in distributed computing manner without the involvement of central trust authority. The blockchain has been used in various area such as manufacturing, culture, and public as well as finance because of its advantage of the security, efficiency and applicability. In the blockchain, it was considered safe a gainst 51% attack because the adversary could not have more than 50% hash power. However, there have been cases cau sed by large-scale computing attacks such as 51% and selfish mining attack, and the frequency of these attacks is increas ing. In addition, since the development of quantum computers can hold exponentially more information than their classical computer, it faces a new type of threat using quantum algorithms. In this paper, we perform the security analysis of block chain attacks composing the large computing capabilities including quantum computing attacks. Finally, we suggest the te chnologies and future direction of the blockchain development in order to be safe against large-scale computing attacks.

스마트폰 사용이 보편화되면서 자연스럽게 서로간의 소통이 메신저를 통하여 이뤄지게 되었다. 하지만 서로간의 대화공간 이 범죄를 공모하는 공간으로도 활용되고 있는 실정이다. 이에 따른 범죄와 관련된 증거들이 스마트폰에 저장된다. 스마트폰 의 특성상 저장정보의 삭제를 손쉽게 할 수 있기 때문에 증거를 신속히 확보 하는 것이 중요하다. 따라서 본 논문에서는 국· 내외에서 대표적으로 사용 중인 모바일 메신저에 대하여 데이터 파일의 아키텍처 분석을 수행하였다. 사용자가 메시지 삭제를 위해 이용할 수 있는 방법에 대한 시나리오를 설정하고 실험을 통해 메시지 복원가능 여부에 대하여 비교 분석한다.

As the use of smartphones become more common, the communication via instant messenger becomes natural. However, it is important to secure the relevant information promptly since the chat room between participants can be used as a space for a criminal conspiracy, and crime-related contents can be stored and deleted easily on smartphones. Therefore, this study aims to identify the available data and to use it as proof by comparing and analyzing the instant messengers with high usage rate.

Key management in multicast dynamic groups, where users can leave or join at their ease is one of the most crucial and essential part of secure communication. Because communication bandwidth, storage memory, and computational power are limited resources, group key management schemes for scalable secure multicast communications have to consider these factors however, most of proposed key management schemes focus on guarantying secure distribution of keys. Even though security assurance is critical, reducing the number of update messages, number of stored keys and computational load also needs equivalent attention. Scalability, security, computation, and communication complexities are some of the tradeoffs. Here, we propose an efficient secure scheme in such a way that solves these problems and that can be implemented easily. In our scheme, we applied very simple mathematical concept with modulo and XOR operations in order to make the system simple for implementation and management. As a novel feature, the scheme is scalable with key update message of order O(1) - logM N, storage overhead of order O(1) and computational load of order O(log(N/M))- O(1).

Anonymous user authentication schemes are essential to guarantee personal privacy in global mobility networks. Recently, Chung, Lee and Won proposed an improved authentication scheme with anonymity which remedies security weaknesses showed by Youn, Park and Lim. However, their improved scheme does not provide unlinkability. In this paper, we propose an anonymous authentication scheme which ensures user unlinkability. It is impossible for attackers to know particular sessions, that have already occurred several times, are originated from one same user. Moreover, we decrease the rounds of transmission among involved entities while consuming same computational power compared with previous schemes. Therefore, the proposed scheme is lightweight and suitable for the mobility networks.

Development of hacking techniques demands more and more network security. For this reason, Major facilities as well as government agencies divide the Protected Network from Internet Network Physically. However, if Internal/external network is divided, file transfer and work efficiency is reduced. To solve this problem and to transfer data between the Internet Network and protected Network, Unidirectional Security Gateway System was born. This paper analyzes unidirectional Security Gateway and suggests a protection profile based on the CC V3.1 to help develop and evaluation of unidirectional Security Gateway.

Home network service is recently installed in each home with the popularity of ubiquitous environment. There are being developed systems that efficiently controlled and handling home entry, electricity product, lamp and boiler at the remote environment as well as at home. The home network has the convenience, but because of there may be the danger to be divested of the authority handling the home network. Therefore, it is ensured the authority handling the home network. This paper proposes the scheme for differential user authentica-tion based on client in home network at a remote place. The scheme supports the convenience and strengthens the security by using the unique value of the registered client. In case of us-ing the unregistered client, it protects the authority accessing the home network via the meth-od strengthening the user authentication by using the security card.

Since 2004, several user authentication schemes purpose to provide user anonymity in wireless communication have been announced, however, many of them failed to provide user anonymity, actually. In 2011, Li and Lee proposed a secure user authentication scheme for wireless communications. Differently from the previous researches, Li and Lee claimed two more session keys to enhance entire security of the scheme. However, to fulfil their claim, Li and Lee adopted Diffie-Hellman key agreement method in their scheme, and it required plenty of resources. Thus, in this paper, we point out inefficiency of Li and Lee’s scheme and propose a more efficient user authentication scheme for wireless communications.

Since the Internet is a widely used tool in many business areas these days, a large amount of contents are offered through the Web. Almost all of them, however, are being provided to users without any protection. Anyone can copy and reuse the contents without permission by using features of the Web browser and even use them for commercial purposes. It is time to consider having countermeasures to protect the Web contents from illegal use and leakage. Therefore, this paper proposes security requirement of Web contents protection system by modeling a threat and applying a security requirements engineering methodology base on Common Criteria V3.1. It can be used as reference in the case of introduction or evaluation of the system.

A dynamic ID-based user authentication scheme is designed to protect leakage of a user’s partial information from intruders while enabling authenticated users to be granted access to the network service. In 2012, Wen and Li proposed a dynamic ID-based remote user authentication scheme with key agreement and claimed that their scheme resisted impersonation attacks and avoided leakage of partial information. However, Kim, et al., described that Wen and Li’s scheme could leak some key information to an adversary and is vulnerable to a man-in-the-middle attack launched by any adversary. In this paper, we show how to solve the vulnerabilities in Wen and Li’s scheme.

Remote user authentication scheme is a kind of way to authenticate the communication parties who transmit messages through an insecure channel. Researchers in this area have proposed some approaches during the last couple of decades. Unfortunately, most of them are proved to be insecure against various attacks. In 2009, Kim and Chung improved Yoon and Yoo’s scheme, and claimed that their scheme can prevent masquerading attack as well as resist to other malicious attacks. However, we found that Kim and Chung’s scheme is still not secure enough, especially in preventing off-line password guessing attack. In this study, we proposed a more secure and practical remote user authentication scheme to resolve all of the aforementioned security vulnerabilities while preserving the merits of Kim-Chung’s scheme.

Protocols for password-based authenticated key exchange (PAKE) enable two or more parties communicating over a public network to build a secure communication channel using their easy-to-remember passwords. However, off-line dictionary attacks have always been a major security concern in designing such password-based protocols. Compared with the two-party setting, the concern is significantly increased in the three-party setting where insider attacks may be mounted. In this paper, we identified an inherent flaw in the design of Nam et al.’s three-party PAKE protocol (IEEE Communications Letters, 13(3), 2009) and Lu and Cao’s protocol (Computers & Security, 26(1), 2007) and demonstrated that both protocols are susceptible to a previously unpublished off-line dictionary attack. We hope that by identifying this design flaw, similar structural mistakes can be avoided in future design. We conclude the paper with a simple countermeasure.

Many enterprises are adopting Mobile Device Management systems to monitor the status and control the functionalities of smart phones and tablet PCs in order to solve the security problems of confidential enterprise data being leaked whenever a device is misused or lost. However, no criteria have been established as yet to evaluate whether such Mobile Device Management systems correctly provide the basic security functions needed by enterprises and whether such functions have been securely developed. Therefore, this paper proposes security requirements of a Mobile Device Management system by modeling a threat and applying a security requirement engineering methodology based on Common Criteria.

Recently, a malware is growing rapidly and the number of malware applies various techniques to protect itself from the anti-virus solution detection. The reason of this phenomenon is that a longer resident on an infected host guarantees the more profit. As a result, these many protection techniques are applied to a malware, a representative of those is a Packing. It is not an exaggeration that most of the malware currently is distributed. In other words, a packer is widely used for a malware protection. Therefore analysts must determine whether the malware was packed or not and if the malware is packed, what packer is used, before an analysis of the malware. For these procedures, some packer detection tools were released and used. But, the detection performance is not good and there is some false positive and false negative. Therefore we propose a signature generation method that is based on an unpacking process and algorithm in this paper. And we offer the packer detection experiment result using the proposed packer detection signature generation method.

Recently, a malware is growing rapidly and the number of malware applies various techniques to protect itself from the anti-virus solution detection. The reason of this phenomenon is that a longer resident on an infected host guarantees the more profit. As a result, these many protection techniques are applied to a malware, a representative of those is a Packing. It is not an exaggeration that most of the malware currently is distributed. In other words, a packer is widely used for a malware protection. Therefore analysts must determine whether the malware was packed or not and if the malware is packed, what packer is used, before an analysis of the malware. For these procedures, some packer detection tools were released and used. But, the detection performance is not good and there is some false positive and false negative. Therefore we propose a signature generation method that is based on an unpacking process and algorithm in this paper. And we offer the packer detection experiment result using the proposed packer detection signature generation method.

Authentication of communicating entities and confidentiality of transmitted data are fundamental procedures to establish secure communications over public insecure networks. Recently, many researchers proposed a variety of authentication schemes to confirm legitimate users. Among the authentication schemes, a one-time password authentication scheme requires less computation and considers the limitations of mobile devices. The purpose of a one-time password authentication is to make it more difficult to gain unauthorized access to restricted resources. This paper discusses the security of Kuo-Lee's one-time password authentication scheme. Kuo-Lee proposed to solve the security problem based on Tsuji-Shimizu's one-time password authentication scheme. It was claimed that their proposed scheme could withstand a replay attack, a theft attack and a modification attack. Therefore, the attacker cannot successfully impersonate the user to log into the system. However, contrary to the claim, Kuo-Lee's scheme does not achieve its main security goal to authenticate communicating entities. We show that Kuo-Lee's scheme is still insecure under a modification attack, a replay attack and an impersonation attack, in which any attacker can violate the authentication goal of the scheme without intercepting any transmitted message. We also propose a scheme that resolves the security flaws found in Kuo-Lee's scheme.

그룹 키 동의 프로토콜은 일련의 그룹을 형성하는 다수의 통신 참여자들이 공개된 통신망을 통해 안전하고 효율적인 방법으로 그룹의 세션키를 설정하기 위한 목적으로 설계된다. 하지만, 기존에 제안된 그룹 키 동의 프로토콜들은 모두 상당한 양의 통신 부하를 유발하기 때문에 전송 지연이 긴 WAN 환경에는 적합하지 않다. 이러한 네트워크 환경에서는 특히 라운드 복잡도와 메시지 복잡도가 프로토콜의 수행 시간을 결정하는 핵심 요소들로서, 무엇보다 이들을 줄이는 것이 효율적인 그룹 키 동의 프로토콜의 설계를 위해 중요하다고 할 수 있다. 따라서 본 논문에서는 라운드 수와 메시지 수 측면에서 효율적인 그룹 키 동의 프로토콜을 제안하고, 이의 안전성을 소인수 분해 문제에 기반 하여 랜덤 오라클 모델에서 증명한다. 제안된 프로토콜은 완전한 전방향 안전성과 최적의 메시지 복잡도를 제공하면서도 상수 라운드만에 그룹멤버의 변경에 따른 세션키 갱신을 수행한다.

Group key agreement protocols are designed to solve the fundamental problem of securely establishing a session key among a group of parties communicating over a public channel. Although a number of protocols have been proposed to solve this problem over the years, they are not well suited for a high-delay wide area network; their communication overhead is significant in terms of the number of communication rounds or the number of exchanged messages, both of which are recognized as the dominant factors that slow down group key agreement over a networking environment with high communication latency. In this paper we present a communication-efficient group key agreement protocol and prove its security in the random oracle model under the factoring assumption. The proposed protocol provides perfect forward secrecy and requires only a constant number of communication rounds for my of group rekeying operations, while achieving optimal message complexity.