k-min问题安全多方计算方案及应用

安全多方计算(MPC)是密码学的一个重要研究方向.保密计算第k小元素是一个重要的MPC问题(简称k-min问题).k-min值MPC协议在保密的投票选举,保密的招投标以及保密的数据统计分析等方面具有广泛应用.目前k-min问题的MPC解决方案大都需要多次调用保密求和协议以及比较协议,协议效率较低.也有一些协议基于移动网络通信应用设计,无法解决MPC应用问题.本文提出新的编码方式,以此为基础并结合Lifted ElGamal门限密码系统设计了简单高效的k-min值MPC协议,应用模拟范例严格证明了协议的安全性,并利用实验证明了方案的可行性.以k-min协议为基础进一步设计了多方成绩保密统计与排序协议.理论分析和实验测试表明本文协议是安全且简单高效的.

基于区块链与安全计算的电能质量扰动分析方法

电力系统中的非线性特性极易引起电能质量扰动,破坏电网运行的稳定性。现有电能质量扰动分析方法大多将采集到的电信号传输到中心服务器进行统一存储,提取统计特征并利用机器学习构建模型。然而,在真实环境中存在隐私保护弱、设备环境复杂、模型过度依赖人工经验等问题。为此,提出了一种基于区块链与安全计算的电能质量扰动分析方法。首先,构建基于智能合约与联邦学习的私有链来保护数据隐私,通过无证书加密保障设备的身份可信,其次,利用基于Paillier密码体制的模型参数同态加密保护深度学习过程中的梯度安全,并建立基于变分模态分解与长短期记忆网络的电能质量扰动异常分析模型,以弥补传统统计特征建模的覆盖率与准确率不足。在真实搭建的微电网下的实验结果表明,该方法能够兼顾隐私性、可用性、安全性与精确性。

An efficient MAC scheme for secure network coding with probabilistic detection

Network coding is vulnerable to pollution at- tacks, which prevent receivers from recovering the source message correctly. Most existing schemes against pollution attacks either bring significant redundancy to the original message or require a high computational complexity to ver- ify received blocks. In this paper, we propose an efficient scheme against pollution attacks based on probabilistic key pre-distribution and homomorphic message authentication codes （MACs）. In our scheme, each block is attached with a small number of MACs and each node can use these MACs to verify the integrity of the corresponding block with a high probability. Compared to previous schemes, our scheme still leverages a small number of keys to generate MACs for each block, but more than doubles the detection probability. Mean- while, our scheme is able to efficiently restrict pollution prop- agation within a small number of hops. Experimental results show that our scheme is more efficient in verification than existing ones based on public-key cryptography.