A secure multi-dimensional partition based index in DAS

Database-as-a-Service is an emerging data management paradigm wherein data owners outsource their data to external untrusted server. To keep server from unauthorized access, sensitive data are encrypted before outsource. Although partition based index is an effective way of querying encrypted data, several such indexes would lead to extra information leakage because of un-uniform multi-dimensional data distribution. In this paper we first introduce two security constraints based on minimal confidential interval and minimal occupation number, then propose a multi-dimensional partition which satisfies previous security constraints. Since optimal multi-dimensional partition is a NP-hard problem, we propose a heuristic based greedy algorithm which is simple and efficient. Experiments show that our index could achieve a stable trade-off between security and efficiency compared to multiple single-dimensional indexes.