Real time face recognition applications use face recognition algorithms, which extract information such as shape, size, and position of facial features (eyes, nose, ear, lips) and then compare this extracted feature to stored featured database to locate matching images. Face recognition application used in airport security, in passport office, in R.T.O office, in government organization, in development and research department. In this paper we implement a application which recognize a face and minimize computation time using our mobile-cloudlet-cloud architecture named MOCHA. And also develop an algorithm for percent variance. Our experimental result demonstrate that we can communicate anywhere from the world by using jelastic cloud. We will run application on mobile devices using jelastic cloud as the backend server.