Airavat: Security and Privacy for MapReduce

Vitaly Shmatikov
University of Texas at Austin

Airavat is a new, MapReduce-based system which provides strong security and privacy guarantees for large-scale distributed computations on sensitive data. Airavat combines mandatory access control with differential privacy. It enables users without security expertise to perform computations on the data, but confines these computations, preventing information leakage beyond the data providers' policies.

Back to Statistical and Learning-Theoretic Challenges in Data Privacy