Shai Caspin will present her General Exam "Enabling Privacy for Decentralized Applications with Noise" on Thursday, May 11, 2023 at 10:00 AM in Friend 125.
Committee Members: Amit Levy (advisor), Ravi Netravali, David Walker
Abstract:
Applications often use a centralized server architecture to store user data and mediate sharing. Unfortunately, this has resulted in repeated data leaks due to incorrect access control, cyber attacks, legal action, or malicious administrators. End-to-end encryption can largely avoid this problem by hiding user data from servers altogether. Leveraging existing end-to-end encryption frameworks is difficult, since they are highly specialized for their applications, enforce only simple data models, have rigid access control schemes, and provide weak consistency guarantees.
I will present Noise, an application-agnostic platform that provides end-to-end encryption for many possible applications, across many devices, on top of a common core design. At its core, Noise establishes end-to-end encryption and strong consistency guarantees across all clients in the system, and maintains these guarantees in the face of Byzantine actors. I will discuss how to build data stores with varying consistency guarantees in this model. Additionally, I will discuss schemes to verify correct behavior in untrusted settings across distributed storage and replication.
Reading List:
Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so.