Privacy-preserving matchmaking in social networks

• المؤلفون الرئيسيون: Ho, Alexander Joong Khuan, Lim, Collin Tian Jun, Wong, Vincent Yong Sheng
• مؤلفون آخرون: Lam Kwok Yan
• التنسيق: Final Year Project
• اللغة: English
• منشور في: Nanyang Technological University 2021
• الموضوعات: Engineering::Computer science and engineering::Data::Data encryption; Engineering::Computer science and engineering::Software::Software engineering
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/151907

The study of privacy preserving matchmaking is a heavily researched topic in the literature, made further relevant due to the exponential growth of smartphones, mobile applications, and online social networks. This capstone project aims to tackle the research field of private set intersections, which studies how to enhance the security of sharing information between two parties in a network. Efficient Outsourced Private Set Intersection (EO-PSI) is a state of the art privacy-enhancing protocol used to identify the common attributes, or set intersection, between two different parties in a network, while delegating the storage of all parties’ attributes onto a cloud server. In this paper, we implemented an improvement to the existing EO-PSI protocol, in a bid to enhance its security and computation efficiency to obtain private set intersections between two parties, as well as benchmarked this improved protocol’s computation and communication performance with the original EO-PSI protocol, as well as other protocols in the literature, such as the Outsourced Private Set Intersection (O-PSI) and the Catalic PSI Cardinality Protocol. We then use these results to implement the improved protocol into a client-friendly full-stack web application, with the help of Amazon Web Services (AWS) Elastic Beanstalk (EB) and Relational Database System (RDS). This web application can thus be used to extrapolate the PSI protocol into countless real-life applications, such as allowing two mutually distrusting app companies to obtain mutually beneficial common attributes between each other.