GC-Lite: Hiding Software, Data & Computed Values Using Lightweight Primitives
Primary Investigator:
Mike Atallah
Shoaib Khan, Mikhail Atallah, Qutaibah Malluhi
Abstract
We present a computationally lightweight protocol that enables execution of a confidential software program $\mathcal{P}$ owned by Alice, on confidential data $\mathcal{D}$ owned by Bob, without revealing anything about $\mathcal{P}$ or $\mathcal{D}$, in an honest-but-curious adversarial model. Depending on the requirements, our protocol can be set up such that the output may be learned only by Alice, Bob, both or neither(in which case an agreed upon third party would learn it). The protocol is run by two parties which can be Alice and Bob, or alternatively, they could be two commodity cloud servers (in which case neither Alice nor Bob participates in the protocol's execution - they merely initialize the two cloud servers, then go offline). Our approach provides a new operating point on the usability/performance/security tradeoff curve for this kind of oblivious co-execution framework, one that fares much better than others from the point of view of programmability, and is competitive when it comes to performance without sacrificing security. We implemented and evaluated a prototype that is made available on Github. We report our experimental findings that demonstrate the viability of our approach.