基于安全多方计算的隐私保护图查询

Privacy-Preserving Graph Query Based on Secure Multi-Party Computation

【目的】在互联网时代,图数据凭借着其丰富语义和结构信息,在众多的领域中发挥着独特的作用。同时,越来越多的公司选择使用“云服务”作为基础设施平台,个人敏感数据的保护问题愈发受到人们的关注。这为隐私保护的图计算带来了严峻的挑战。【方法】本文针对图计算中至关重要的子图匹配问题,首次提出了基于安全多方计算的图查询保护策略,将隐私保护图查询问题转化为关系表的安全连接问题,并根据图数据的特性对安全连接子协议进行改进。【结果】相比于之前的隐私保护图查询工作,本文协议不仅提供了更低的计算和通讯开销,并且具有更高的安全保障性和可信度。

[Objective]In the era of the Internet,graph data,with its rich semantics and structural information,plays a unique role in numerous fields.At the same time,more and more companies are choosing to use"cloud services"as their infrastructure platform,and the protection of personal sensitive data is increasingly attracting people's attention.This poses severe challenges to privacy-preserving graph computing.[Methods]This paper focuses on the crucial subgraph matching problem in graph computing and proposes,for the first time,a privacy-preserving graph query strategy based on secure multi-party computation.The privacy-preserving graph query problem is transformed into a secure join problem for relational tables,and the secure join sub-protocol is improved according to the characteristics of graph data.[Results]Compared with previous works on privacy-preserving graph query,our protocol not only provides lower computation and communication overhead,but also has higher security and credibility.