CAREER: Towards Practical Systems for Trustworthy Cloud Computing

University Of Maryland, College Park, College Park MD

Investigators

Abstract

An increasing number of individuals, enterprises, and governments migrate their data and applications to the cloud. Computations delegated to the cloud may return erroneous results, outsourced files may be lost or discarded, and sensitive information may be arbitrarily accessed for advertising purposes. This project enhances the ability of cloud companies to integrate advanced protection mechanisms into their products, benefiting the online safety of cloud clients. The project also aims to improve security education via new undergraduate and graduate curriculum and outreach to high school students. This project addresses foundational problems on the security, availability, and privacy of cloud computing by using an application-driven approach and demonstrates the broad applicability of this theory in practice. The project develops low bandwidth and sub-linear computation protocols for securely recovering correct data blocks and computation results. The researchers investigate practical and expressive privacy preserving algorithms with reduced leakage profiles and low I/O complexity for advanced queries on encrypted static and dynamic data. They also develop new fundamental techniques for verifying frequently recurrent classes of delegated computations