Dog ORAM: A Distributed and Shared Oblivious RAM Model with Server Side Computation

Outsourcing to the Cloud is becoming an attractive option for many organisations dealing with large amounts of data. However, there is still a reluctance amongst companies dealing with highly regulated data because traditional Cloud storage does not support the level of privacy required to prevent access pattern leakage. Oblivious Random Access Machines (ORAM) have been a hot topic of research over the past number of years, proposing various cryptographic techniques to obtain the privacy levels required. We propose a new model, Dog ORAM - a distributed and shared oblivious RAM model with server side computation, that merges several models existing in the literature and includes a new method of access right management for multi-party data access. To achieve this, we use an additive homomorphic encryption scheme and a chameleon signature.