Outsourced data is ubiquitous and raises the privacy model in which a client must interact and store data on a potentially untrusted server. Databases may be encrypted but usage in the form of access patterns still may leak information about both data and queries. This motivates us to examine techniques to conceal access patterns from an untrusted server. Doubly-Efficient Private Information Retrieval (DEPIR) is one technique that allows a single server to perform an read with overhead (both bandwidth and server computation) that is sub-linear in the database size.
In this thesis proposal, I will discuss three different applications of DEPIR for outsourced data use. Private Anonymous Data Access (PANDA) uses techniques from DEPIR to achieve a multi-client private and anonymous data access. Rewindable ORAM (R-ORAM) leverages DEPIR to provide a variant of ORAM used in our Fully Homomorphic Encryption for RAM construction. Finally, I will explore DEPIR's applications to Distributed ORAM (DORAM) in the secure computation model.
Document| Committee Member | Webpage | Justification |
|---|---|---|
| Daniel Wichs | link | Advisor |
| abhi shelat | link | Abhi is an expert in cryptography, with a focus on applied aspects |
| Jon Ullman | link | Jon is an expert on privacy, especially in the distributed setting, which is closely related to some of this work. |
| Ran Canetti | link | Ran is an expert on the theory of cryptography. His prior work on doubly-efficient private information retrieval is the basis of much of Ariel's proposed research. |