Privacy preserving spam email filtering based on somewhat homomorphic using functional encryption

Abstract

With the advent of cloud computing, there has been a recent trend of delegating the computation (of a specific function) from the client to the third party (cloud). The privacy condition in this scenario requires that the third party should be able to perform the computation of the required specific function over the private (encrypted) data, but learn nothing else about the data (apart from the specific computation performed on that specific encrypted data). We take into account this concern in scenario of proxy email server (semi honest) which is required to perform specific functions (filtering encrypted spam mails) with minimal or no knowledge about the input data. We also highlight the limitations and challenges over the set of functions that can be performed by proxy email server in context of functional encryption.