Towards Cloud Data Warehouses of Multivalued Encrypted Values

Abstract

The motivation behind generating multivalued encrypted values by multivalued encrypted techniques (METs) is the minimization of encrypted data redundancy. In the literature, there are METs that enable the processing of sum aggregations, selection constraints and sorting operations over multivalued encrypted values. However, we found a lack of METs for processing data groupings over multivalued encrypted values. In this article, we propose two novel METs that allow the execution of data groupings over multivalued encrypted values and we specify two novel data schemas for encrypted DWs (MV-HOM and MVSE-HOM) that allow analytical queries with data groupings be performed over multivalued encrypted values. Also, we conducted a performance evaluation of encrypted DWs stored in a cloud and results showed that MV-HOM's overhead was decreased up to 10.41% (with 16 nodes) when compared to a non-encrypted dimensional data schema, while the overhead imposed by MVSE-HOM was high (89.91% with 16 nodes).