Paper 2022/894

Toward Full Accounting for Leakage Exploitation and Mitigation in Dynamic Encrypted Databases

Abstract

Encrypted database draws much attention as it provides privacy-protection services for sensitive data outsourced to a third party. Recent studies show that the security guarantee of encrypted databases are challenged by several leakage-abuse attacks on its search module, and corresponding countermeasures are also proposed. Most of these studies focus on static databases, yet the case for dynamic has not been well investigated. To fill this gap, in this paper, we focus on exploring privacy risks in dynamic encrypted databases and devising effective mitigation techniques. To begin with, we systematically study the exploitable information disclosed during the database querying process, and consider two types of attacks that can recover encrypted queries. The first active attack works by injecting encoded files and correlating file volume information. The second passive attack works by identifying queries’ unique relational characteristics across updates, assuming certain background knowledge of plaintext databases. To mitigate these attacks, we propose a two-layer encrypted database hardening approach, which obfuscates both search indexes and files in a continuous way. As a result, the unique characteristics emerging after data updates can be eliminated constantly. We conduct a series of experiments to confirm the severity of our attacks and the effectiveness of our countermeasures.