基于单服务器的模指数安全外包计算方案

摘要

目前,在离散对数密码协议中,模指数外包计算方案大部分都是针对素数的,很少有有关合数的研究成果.并且大多数模指数外包计算方案都是基于双服务器的,该类方案通常要求两个服务器中至少有一个是诚实的,但是在实际环境中可能并不存在完全可信的云服务器.基于单个不可信服务器模型提出了一个新的复合模指数安全外包计算方案.该方案采用新的数学分割方式,能够保证底数和指数的隐私性.与已有方案相比,该方案的外包计算结果可验证率也有很大程度的提高,用户能够以119/120的概率检测出错误结果.%At present,in discrete-log based cryptographic protocols,most of the computational models of modular exponentiation are for primes,while less work has been done for composite.What's more,most schemes are based on two servers,in which it requires at least one server to be honest.However,there may not be a fully trusted cloud server in the actual environment.Then a new secure method for outsourcing exponentiation modular a composite which based on a single server was proposed.The scheme used a new mathematical division method,it could ensure the privacy of the base and exponentiation.Compared with the existing schemes,the checkability of our scheme can be greatly improved.The user can detect the error result with the probability of 119/120.