Title    name Lastest   |   Most hits   년도선택 최근 50건20102009200820072006200520042003200220012000     2010 International Workshop on Mathematical Cryptology HCI Security: 1. Human Knowledge-Based Authentication Kwon, Taekyoung (Sejong University) 2010-08-28   사용자 인증을 위한 편리한 방법으로서 사용자의 지식에 기반한 인증 기법이 널리 사용되고 있다. 하지만 이것은 사용자 인터페이스를 통한 입력이 필요하며, 특히 입력 시점에 shoulder-surfing 공격을 통해서 사용자의 입력값 또는 비밀값이 쉽게 유출될 수 있는 위험이 있다. 이와 같은 문제를 해결하기 위해서 그동안 많은 기법이 제안되었지만 HCI를 고려한 안전성과 사용성 측면에서 폭넓은 사용이 어렵다. 이러한 주제를 바탕으로 HCI 보안의 어려움과 중요성에 대해서 알아본다.   2010 International Workshop on Mathematical Cryptology Traffic measurement with virtual vector and anomaly detection Yoon, MyungKeun (Kookmin University) 2010-08-28   고속 네트워크 환경에서 정확한 트래픽 측정은 중요하지만 어려운 문제이며, 작은 SRAM내에 상주할 수 있는 컴팩트한 자료구조 설계가 핵심기술 중 하나이다. 정확한 측정결과는 네트워크 보안에 직접 응용될 수 있는데 heavy hitter와 heavy distinct hitter를 탐지하는 경우가 이에 해당한다. Heavy hitter란 측정기간 동안 비정상적으로 많이 등장하는 트래픽을 말하며, heavy distinct hitter란 비정상적으로 많은 서로 다른 목적지로 접근을 시도하는 출발지를 의미한다. 우리가 제안했던 가상벡터 자료구조는 heavy distinct hitter를 탐지하는데 있어서 기존 알고리즘보다 빠르면서도 메모리 사용량을 줄이는 장점을 가진다. 최근에는 가상벡터가 heavy hitter 탐지에도 매우 효과적으로 사용될 수 있음이 확인되었으며 이번 워크숍을 통해 관련된 연구내용을 소개한다.   2010 International Workshop on Mathematical Cryptology Order Preserving Symmetric Encryption Lee, Younho (Yeungnam University) 2010-08-28   SIGMOD 04에서 Agrawal에 제안했던 순서 보존 대칭 암호 (Order Preserving Symmetric Encryption: OPE)에 대하여 현대 암호학적 측면에서의 연구를 시작한다. 첫번째로 순서 보존 대칭 암호에서는 표준 보안 모델의 하나인 선택적 평문 공격 (IND-CPA) 에 대한 구별 불가능성을 이룰 수 없음을 증명한다. 이를 대신하여, 우리는 의사 난수 함수 (PRF)에 기반한 보안 모델을 제안한다. 본 모델에서는 순서 보존 제한 조건하에서 "가능한 한 가장 무작위로(as-random-as-possible)" 라는 개념을 근거로 한다. 우리는 효율적인 OPE 방법을 설계하고, 제안 방법이 블럭암호의 의사난수성에 기반하여 제안 보안 모델에서 안전함을 증명한다. 제안 구현 방법은 순서 보존 난수 함수와 초기하 분포 (hypergeometric distribution) 와의 관계를 바탕으로한다. 즉, 효율적인 초기하 분포 추출(sampling) 알고리즘을 black-box 로 이용한다.   2010 International Workshop on Mathematical Cryptology Pairing-based Non-interactive Zero-Knowledge Proofs Jens Groth (University College London) 2010-08-28    Non-interactive zero-knowledge proofs enable a prover to convince a verifier of the truth of a statement without revealing any other information. They are useful in contexts where verifiability and privacy are needed at the same time and have many applications, for instance in the context of anonymous signatures. Pairing-based cryptography has provided new insights in the area of non-interactive zero-knowledge proofs. We will see how pairing-based non-interactive proofs can be constructed in a simple fashion and how to obtain perfect and everlasting privacy. We will also see how pairing-based techniques have improved the efficiency of non-interactive zero-knowledge proofs and made them useful in practical application.   2010 International Workshop on Mathematical Cryptology Sharing Sensitive Information with Privacy Gene Tsudik (University of California) 2010-08-28    Modern society is increasingly dependent on, and fearful of, the availability of electronic information. There are numerous realistic scenarios where sensitive data must be (sometimes reluctantly or suspiciously) shared between two or more entities without mutual trust. As often happens, the research community has foreseen the need for mechanisms to enable limited (privacy-preserving) sharing of sensitive information and a number of effective (if not always efficient) solutions have been proposed. Among them, Private Set Intersection techniques are particularly appealing whenever two parties wish to compute an intersection of their respective sets of items without revealing to each other any other information. This talk motivates the need for Private Set Intersection (PSI) techniques with various features and privacy properties and illustrates several concrete Private Set Intersection protocols that offer appreciably better efficiency than prior work. We show how to integrate these protocols into a real database system supporting privacy-preserving querying. We also demonstrate their practicality via experimental results obtained from a prototype implementation and discuss a number of systems issues encountered in developing a toolkit that provides various flavors of PSI.   2010 International Workshop on Mathematical Cryptology Effcient Verifiably Encrypted Signatures Without Random Oracles Lee, Juhee (Ewha Womans University) 2010-08-28   A verifiably encrypted signature can convince the verifier that a given ciphertext is the encryption of the signature on a given message, but only the party whose key was used to encrypt it can recover the signature. Recently, Yang et al. proposed a verifiably encrypted signature scheme and claimed that their scheme was secure without random oracle model. Although their verifiably encrypted signature scheme is more efficient and provably secure in standard model, their security proof of opacity is awed. To overcome this weakness, we revisited this scheme and as a result, our verifiably encrypted signature scheme is more efficient and provably secure in the standard model under the q-SDH assumption.   2010 International Workshop on Mathematical Cryptology Practical and Robust Multi-Party Private Set Intersection Seo, Jaehong (Seoul National University) 2010-08-28   One challenge in privacy-preserving multiparty set operations is to design a protocol with linear complexities for communication and computation. This paper describes a new approach that enables to gain quasi-linear complexities in the input size. We outline technical alternatives to represent user’s dataset using polynomial and to carry out polynomial arithmetics, and show how they help improving practical efficiency, especially, in polynomial arithmetics and zero-knowledge proof. The complexities of our protocols are O(n2k) transmission overhead and O((n+log k)2k+(n+(log k)2)k ) computation overhead for k size dataset, n players, and plaintext domain size of an encryption. Based on these building blocks, we construct a multiparty set intersection (MPSI) protocol and then prove that our construction is secure in the malicious model. Our protocol is robust even when corrupted players leave during the process of the protocol.   2010 International Workshop on Mathematical Cryptology On the Generic Group Model in Pairing-Based Cryptography Hovav Shacham (University of California) 2010-08-27    The so-called generic group model, due to Nechaev, to Shoup, and to Maurer, abstracts away the details of the group operation and of element representation. For many problems of cryptographic interest, it is possible to prove lower bounds in the generic group model on the number of operations performed by algorithm solving the problem; the implication is that any more-efficient algorithm must be non-generic, i.e., make use of the details of the group in which the problem is posed. Since Boneh and Boyen's 2004 paper, the generic group model has been extensively used to gain confidence in hardness assumptions for pairing-based cryptography by demonstrating lower bounds for the associated problems in a generic group equipped with a bilinear map. This use of the generic group model has coincided with growth in the number and baroqueness of hardness assumptions introduced and used in the literature. We recall the generic group model and survey its use in pairing-based cryptography. We then consider some limitations of the generic group model, the most important of which is that the groups used for pairing-based cryptography have substantial special structure that could be exploited by non-generic algorithms. We suggest that the following problem is worth studying: Give a (natural) algebraic problem that is hard in the generic group model but not in (one of) the groups used for pairing-based cryptography.   2010 International Workshop on Mathematical Cryptology A New Approach on Pairings and Functional Encryption Tatsuaki Okamoto (NTT) 2010-08-27    If a cryptographic system is highly functional, it is not easy to design the system and analyze its security. It is a natural strategy to design and analize such a complicated system in a modular way.This talk introduces a type of modular way of design and analysis, hierarchical design and security proof. In our approach, the underlying primitive is bilinear pairing groups, and a higher level of concept, dual pairing vector spaces (DPVS), is constructed on the primitive, where a highly functional crypto system is directly designed over the higher level of concept. In addition, the security proof can be executed in a hierarchical manner.   2010 International Workshop on Mathematical Cryptology A Provable-Security Treatment of Symmetric Encryption Phillip Rogaway (University of California) 2010-08-27    Once a tool only for making quite theoretical claims in cryptography, reductions have become the central tool for the design and analysis of practical, high-assurance cryptographic schemes for intermediate-level primitives. This talk will develop the reduction-based treatment of symmetric (that is, shared-key) encryption. We will explore formalizations for the problem and then means for provably achieving symmetric-encryption goals, assuming the existence of a secure blockcipher

Information Center for Mathematical Sciences KAIST 335 Gwahangno (373-1 Guseong-dong), Yuseong-gu, Daejeon 305-701, Republic of Korea TEL +82-42-869-8195~6, FAX +82-42-869-5722, e-mail : mathnet@mathnet.or.kr Copyright (c) 2009 by ICMS. All Rights Reserved