University Of California, San Diego (Calit2)
Global data has been growing much faster than Moores Law, creating an \"Energy wall\" in designing systems to exploit that data. This problem is exacerbated by the increasing demand for security to ensure both secrecy and integrity of all the data in the cloud. Those security mechanisms can consume significant energy overhead if designed in an ad-hoc manner. In this paper, the authors estimate the energy consumption per byte for several major security mechanisms, and compare them with the energy consumed in useful computation. They estimate that security mechanisms can consume almost 30% of computation energy, becoming a giant target for energy efficiency and optimization.