In this paper, we propose PRAMOD – a novel framework for enabling efficient and secure database-as-a-service. We consider a setting in which data is stored encrypted on the untrusted cloud and data dependent computations are performed inside a trusted environment. The proposed framework protects against leakage caused by observable data movement between different components (due to limited private memory) by using a special component called scrambler running in O(n) time. It supports popular algorithms underlying many data management applications, including sort, compaction, join and group aggregation. The algorithms implemented in PRAMOD are not only privacy-preserving but also asymptotically optimal. They can be used as building blocks to construct efficient and secure query processing algorithms. The experimental study shows reasonable overheads over a baseline system assuring a weaker level of security. More remarkably, PRAMOD shows superior performance in comparison with state-of-the-art solutions offering similar privacy protection: up to 4.4× speedup over the alternative data-oblivious algorithms.
Category / Keywords: implementation, applications Date: received 22 Sep 2015, last revised 1 Oct 2015 Contact author: hungdang at comp nus edu sg Available format(s): PDF | BibTeX Citation Note: This is an updated version of the report 2015/919 Version: 20151002:044653 (All versions of this report) Short URL: ia.cr/2015/919 Discussion forum: Show discussion | Start new discussion