Data Loss and Reconstruction of Location Differential Privacy Protection Based on Edge Computing
Data Loss and Reconstruction of Location Differential Privacy Protection Based on Edge Computing
复制标题
基于边缘计算的位置差分隐私保护的数据丢失与重构
DOI：
10.1109/access.2019.2922293
复制
发表时间：
2019-06
期刊：
影响因子：
3.9
通讯作者：
Chen Xuebin
中科院分区：
文献类型：
--
作者：
Jing Weipeng;Miao Qiucheng;Song Houbing;Chen Xuebin
随着物联网的发展，网络传输带来的延迟导致数据处理效率低下。边缘计算的出现可以有效降低数据传输的延迟，提高数据处理能力。但与此同时，物联网面临重要挑战，边缘计算使用大量分布式设备，难以进行集中控制。当边缘节点受到攻击时，攻击者可以继续入侵其连接的节点，从而挖掘和窃取用户的隐私数据并造成损失。一旦边缘层通信链路受到攻击或意外中断，用户的隐私信息就有可能被泄露。针对这些问题，本文提出利用差分隐私来保护用户隐私。首先，根据边缘计算的三层通信链路结构，提出数据查询模型；主要功能是捕获结构信息和数据中心连接权重，查询边缘节点与客户端的连接关系。其次，将边缘节点视为中心服务器，利用差分隐私理论实现位置隐私的保护。最后，为了减少定位保护过程中造成的数据丢失，采用线性规划实现最优定位模糊矩阵的选择，并采用数据丢失和重构方法来最小化数据的不确定性。与现有的差分隐私方法相比，本文的方法能够实现更好的隐私保护，并能有效减少数据丢失。
With the development of the Internet of Things (IoT), the delay caused by network transmission has led to low data processing efficiency. The emergence of edge computing can effectively reduce the delay of data transmission and improve data processing capacity. However, at the same time, the IoT faces important challenges, and edge computing uses a large number of distributed devices, making it difficult to perform centralized control. When an edge node is attacked, the attacker can continue to invade its connected nodes, thereby mining and stealing a user’s private data and causing losses. Once the edge layer communication link is attacked or accidentally interrupted, the user’s private information is likely to be leaked. To solve these problems, this paper proposes to protect user privacy by using differential privacy. First, according to the three-layer communication link structure of edge computing, a data query model is proposed; the main function is to capture the structure information and the data center connection weight and to query the connection relationship between the edge node and the client. Second, the edge node is regarded as the central server, and the differential privacy theory is used to realize the protection of location privacy. Finally, to reduce the data loss caused in the process of location protection, linear programming is adopted to realize the selection of the optimal location fuzzy matrix, and data loss and reconstruction methods are used to minimize the data uncertainty. In comparison to the existing differential privacy method, the method in this paper can achieve better privacy protection and can effectively reduce data loss.