| Popis: |
Homomorphic encryption (HE) allows the execution of queries over encrypted data without requiring decryption. However, we found a lack of studies for investigating the use of HE in data warehouse (DW). In this article, we propose a framework that defines how a HE scheme can be used to encrypt numeric measures of a DW, and how sum aggregation functions of analytic queries are processed over encrypted DWs created according to our framework. Also, we propose a system architecture for safely processing encrypted DWs and describe performance tests that investigated the impact of HE schemes on the performance of analytic queries executed over encrypted DWs. Results indicated that HE generated performance gains of up to 47.46% w.r.t. an encryption scheme that was unable of computing aggregates over encrypted data. Also, results showed that the use of HE caused overheads of up to 16.55% on the computation of aggregates w.r.t. an unencrypted DW. |
