抗泄露的密码协议研究

In traditional cryptographic protocols, the absolute privacy of the inputs of all parties involved is a basic assumption. In other words, no information about secret key of parties involved is leaked. However, due to a large class of attack in the practical application, such as side channel attacks, traditional security assumptions are too idealistic.. Based on different types of adversarial attack, this project studies, in the presence of information leakage, the theories and methods of cryptographic protocols which are closer to reality. Its concrete content is as follows: .1. to give valid input-preprocessing to leakage-resilient cryptographic protocols based on some primitives such as multisource-extractor, homomorphic encryption schemes, hash proof systems, key homomorphic pseudorandom functions;.2. to develop leakage-resilient cryptographic protocols with UC-security by virtue of transition probability, conditional entropy, Markov chain and hidden Markov chain;.3. to investigate leakage-resilient computations of cryptographic primitives using some primitives such as homomorphic encryption schemes;.4. to propose cryptographic protocols against continual leakage using two-source extractor and the methods in signal processing . .On that basis, we will put forward the implementation approaches on leakage-resilient general cryptographic protocols ..The fruits of the proposed research will promote and enrich the traditional cryptographic protocols. The research of this project is of great importance both in theory and practice.

传统上的密码协议假定诚实参与方的内部状态对攻击者来说是完全保密的，但实际中攻击者可能会通过各种边信道攻击（如时间攻击、电源耗损、冷启动攻击、及频谱分析等），获得诚实参与方的内部状态，这种攻击称为泄露攻击。.本项目在有泄露攻击的场景中，基于不同的敌手攻击模型，研究抗泄露安全的密码协议的理论和方法，设计更接近于现实情况的、高效的安全密码协议。具体内容包括：（1）利用多源提取器、同构加密方案、哈希证明系统、密钥同构的伪随机函数，给出抗泄露密码协议高效预处理方法；（2）利用转移概率、条件熵、马尔科夫链及隐含马尔科夫链，给出UC安全的抗泄露密码协议的实现方案；（3）利用同构加密方案，研究抗泄露的安全计算协议；（4）利用两源提取器及信号处理的方法，研究抗连续泄露的密码协议。在此基础上，设计抗泄露的一般密码协议的实现方案。.项目的研究将推广和丰富传统上的密码协议，选题具有重要的理论意义和实用价值。

针对现有密码学基础机制的抗泄露性研究中所存在的问题，本项目研究了公钥加密机制、基于身份的加密机制、无证书公钥加密机制等几类密码学基础原语的抗（连续）泄露性。基于经典的静态假设，对相应构造的安全性做了证明。同时确保泄露参数是一个独立于待加密消息空间的固定常数，即消除泄露参数与待加密消息间的相互制约关系，并且密文中的所有元素对于任何敌手而言都是完全随机的，使得任意敌手都无法从相应给出的密文中获知关于秘密钥的相关泄露信息。此外，在保持秘密钥长度不变的情况下，尽量增加方案所能容忍的泄露量，以提高相应构造的泄露率。项目执行期间共获省科学技术奖三等奖1项、国家网络安全优秀教材奖1项，出版教材1部、学术专著1部，发表科研论文55篇，申请专利9项（其中授权3项），培养博士后1名，博士研究生2名，硕士研究生10名，国内访问学者3名。