Please note: This seminar will take place in DC 1304 and online.
Sujaya Maiyya, Assistant Professor
Cheriton School of Computer Science
Over 94% of today’s enterprises, including 83% of healthcare organizations, have embraced cloud services for their infrastructure and data storage needs. They opt for third-party cloud providers due to the cost and resource savings compared to maintaining on-premise storage and computing systems. However, entrusting plaintext application data to these providers can expose sensitive information to potentially untrustworthy entities. While data encryption is the initial step in ensuring data privacy, a growing array of attacks exploit user access patterns on encrypted data to reveal the plaintext content.
Oblivious databases are storage systems that conceal user access patterns to safeguard against such threats. However, most existing oblivious databases face issues such as single points of failure and high performance overheads. In this talk, I will introduce two solutions addressing these challenges:
- QuORAM: the first replicated oblivious RAM-based datastore that offers both fault tolerance and high availability while efficiently safeguarding access patterns; and
- Waffle: an oblivious datastore that grants the flexibility of tunable privacy, enabling intelligent trade-offs between privacy and performance.