Many companies choose the cloud as their data and IT infrastructure platform. The remote access of the data brings the issue of trust, and the potential risk of compromising sensitive information should not be underestimated. Despite the use of strong encryption schemes, adversaries can still learn valuable information regarding encrypted data by observing the data access patterns. To that end, one can hide the access patterns, which may leak sensitive information, using Oblivious RAMs (ORAMs). Numerous works have proposed different ORAM constructions. Nevertheless, many such ORAM constructions are of only theoretical interest, hence, are notuseful in practice. Several more practical ORAM constructions do exist, but they have never been thoroughly compared against and tested on large databases. There are no open source implementation of these schemes, making such a study challenging to carry out (since most ORAMs are quite contrived in terms of both theoretical analysis and practical implementations).These limitations make it difficult for researchers and practitioners to choose and adopt a suitable ORAM for their applications. To address this issue, we provide a thorough study over several practical ORAM constructions, and implement them under the same library. We perform extensive experiments to provide insights into their performance characteristics with respect to efficiency, scalability, and communication cost. Lastly, we plan to release our ORAM implementations through GitHub so that the communities at large may benefit from and contribute to an open source ORAM library under one unified framework.
The wide presence of large graph data and the increasing popularity of storing data in the cloud drive the needs for graph query processing on a remote cloud. But a fundamental challenge is to process user queries without compromising sensitive information.This work focuses on privacy preserving subgraph matching in a cloud server. The goal is to minimize the overhead on both cloud and client sides for subgraph matching, without compromising users%u2019 sensitive information. To that end, we transform an originalgraph G into a privacy preserving graph Gk, which meets the requirement of an existing privacy model known as k-automorphism. By making use of the symmetry in a k-automorphic graph, a subgraph matching query can be efficiently answered using a graphGo, a small subset of Gk. This approach saves both space and query cost in the cloud server. In addition, we anonymize the original query graphs to protect their label information using label generalization technique. To reduce the search space for a subgraph matching query, we propose a cost model to select the more effectivelabel combinations. The effectiveness and efficiency of our method are demonstrated through extensive experimental results on real datasets.