Researcher, Bell Labs
Mar. 3rd, 2017, 11am-12pm, DBH 6011
Enabling data sharing with secure computation
Today, information is one of the most valuable goods. It is particularly impactful when it is versatile, structured, searchable, and available on-demand. At the same time, collection, storage, correlation and sharing of information, especially across ownership domains, may break business models, pose significant privacy, security and compliance risks, and is often illegal.
I will introduce secure multi-party computation (MPC), the area of cryptography that reconciles the fundamental conflict between data utility and privacy. Often described as “computation under encryption”, MPC allows evaluating arbitrary functions on private inputs, while guaranteeing that each party learns nothing beyond its intended output.
I will sketch the state of the art of practical MPC, focusing on several impactful techniques. As an important data sharing enabler and an MPC application, I will present Blind Seer, a database management system implementing access control, data protection, and, importantly, provably hiding the SQL query from the server. I will discuss promising research directions that could define how sensitive data is stored and processed.
Vlad Kolesnikov is a cryptography and security researcher at Bell Labs, which he joined in 2006 after receiving his Ph.D. at the University of Toronto. His main current research interest is improving and applying secure computation and crypto techniques in practice. He has authored papers on garbled circuit, homomorphic encryption, related techniques and applications. He is interested in database security and privacy, key exchange and channel security. Dr. Kolesnikov has been involved in the design and analysis of Smart Grid networks, Storage Area Networks, wireless and biometric authentication, and other secure systems. His work has been supported by IARPA and ONR grants.