云环境下保护隐私的最短距离计算方法研究  

Privacy-preserving shortest distance computing in cloud environment

作　　者：张莹光[1] 苏森[1] 陈维峰[1] 杨放春[1] 

机构地区：[1]北京邮电大学网络与交换技术国家重点实验室,北京100876

出　　处：《华中科技大学学报（自然科学版）》2013年第S2期77-80,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家重点基础研究发展计划资助项目(2011CB302704);国家自然科学基金创新研究群体资助项目(61121061);北京市教委共建项目;北京邮电大学青年科学研究和创新计划资助项目(2013RC1101)

摘　　要：基于同态加密算法提出了一种保护边权值信息隐私安全的节点间最短距离计算方法.该方法对边的权值信息进行加密,然后将加密后的数据外包到云平台中.利用广度优先原理,云服务器能够直接针对加密的图状数据从初始节点向外扩展,迭代地计算初始节点到其他节点的最短距离,从而计算出任意两节点之间的最短距离.最短距离计算由客户端和云服务器端交互合作完成,主要的计算由服务器端完成,客户端只进行简单的解密和判断操作.本方法能够使得低计算能力客户端也可以使用云计算执行计算密集型的最短距离计算任务,并保证边权值、计算的中间结果以及最终结果等隐私信息不会被泄露.理论分析和模拟实验证明了本方法的安全性和正确性.A shortest distance computing algorithm was proposed to protect the edge weight privacy based on homomorphic encryption.The algorithm encrypted the edge weight information and outsourced the encrypted data into cloud.Following the breadth-first principle,cloud servers could iteratively compute the shortest distance from initial vertex extending to other vertices directly on the encrypted data,getting the shortest distance between two arbitrary vertices on the encrypted graphs. The computation was completed by clients and cloud servers interactively:the major computation was performed by cloud servers;clients only executed simple decryption and judgement operations.This algorithm could make computation constrained clients use cloud to perform computation-intensive shortest distance computing.The edge weight information,intermediate result and final result for the computation could not be leaked during the computing procedure.The analysis and experiments demonstrated its security and validity.

关 键 词：最短距离计算 云计算 加密的图状数据 隐私保护 同态加密 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]