Through recent years, much research has been conducted into processing privacy policies and presenting them in ways that are easy for users to understand. However, understanding privacy policies has little utility if the website’s data processing code does not match the privacy policy. Although systems have been proposed to achieve compliance of internal software to access control policies, they assume a large trusted computing base and are not designed to provide a proof of compliance to an end user. We design Mitigator, a system to enforce compliance of a website’s source code with a privacy policy model that addresses these two drawbacks of previous work. We use trusted hardware platforms to provide a guarantee to an end user that their data is only handled by code that is compliant with the privacy policy. Such an end user only needs to trust a small module in the hardware of the remote back-end machine and related libraries but not the entire OS. We also provide a proof-of-concept implementation of Mitigator and evaluate it for its latency. We conclude that it incurs only a small overhead with respect to an unmodified system that does not provide a guarantee of privacy policy compliance to the end user.

No abstract

Differential privacy (DP) allows data analysts to query databases that contain users' sensitive information while providing a quantifiable privacy guarantee to users. Recent interactive DP systems such as APEx provide accuracy guarantees over the query responses, but fail to support a large number of queries with a limited total privacy budget, as they process incoming queries independently from past queries. We present an interactive, accuracy-aware DP query engine, CacheDP , which utilizes a differentially private cache of past responses, to answer the current workload at a lower privacy budget, while meeting strict accuracy guarantees. We integrate complex DP mechanisms with our structured cache, through novel cache-aware DP cost optimization. Our thorough evaluation illustrates that CacheDP can accurately answer various workload sequences, while lowering the privacy loss as compared to related work.