Privacy Preserving and Efficient Outsourcing Algorithm to Public Cloud: A Case of Statistical Analysis

The growth of the cloud computing services and its proliferation in business and academia has triggered enormous opportunities for computation in third-party data management settings. This computing model allows the client to outsource their large computations to cloud data centers, where the cloud server conducts the computation on their behalf. But data privacy and computational integrity are the biggest concern for the client. In this article, the authors attempt to present an algorithm for secure outsourcing of a covariance matrix, which is the basic building block for many automatic classification systems. The algorithm first performs some efficient transformation to protect the privacy and verify the computed result produced by the cloud server. Further, an analytical and experimental analysis shows that the algorithm is simultaneously meeting the design goals of privacy, verifiability and efficiency. Also, found that the proposed algorithm is about 7.8276 times more efficient than the direct implementation.