Assistant Professor Johes Bater and his team seek to eliminate data breaches by building secure and private database management systems that balance the trade-offs among privacy, performance, and query result accuracy.
His core research projects include:
- Private Data Sharing
- Secure Databases in the Untrusted Cloud
- Data-Adaptive Privacy for Confidential Computing
Private Data Sharing
To securely share private data among collaborators, where participating organizations pool their data to answer SQL queries without revealing private information to each other, privacy-preserving federated databases are needed. Our research looks at how to build these systems using cryptographic primitives to ensure privacy and security, while still providing the necessary speed and accuracy for real-world applications.
Secure Databases in the Untrusted Cloud
Current data pipelines largely rely on third party cloud services to host and query proprietary data that is continually uploaded from users. To ensure data protections and prevent data breaches, extensive cryptographic protocols and privacy protections are needed. Our research tackles how to build privacy-preserving cloud databases systems that maintains system guarantees in the face of large, ever-increasing data sources.
Data-Adaptive Privacy for Confidential Computing
Decades of work in database systems research has focused on adapting systems to changing data through data-dependent optimizations. However, these are all data-dependent processes, which means that they cannot be used without creating side channels that potentially reveal sensitive information. Our research seeks to build data-adaptive processes that can be optimized to improve performance based on the incoming data while still maintaining required privacy and security guarantees.