Keystroke dynamic authentication in mobile cloud computing

Abstract

Contemporary mobile sets are not used just for making calls and sending messages. They are increasingly being used in Mobile Cloud Computing (MCC) to store sensitive and critical information as well as to access sensitive data using the Internet via Cloud Service Provider (CSP). MCC is combination of Cloud Computing (CC) and mobile communication. Using MCC helps to decrease running cost and expansion of mobile applications. One of the important challenges in MCC is security and privacy. Furthermore, authentication plays an important role in preserving security and privacy of shared information in MCC. In fact, the majority of mobile handsets use inherently weak authentication mechanisms, based upon passwords and Personal Identification Number(PINs). But it is not secure way for authenticating users because of its limitation, as well as it is difficult to confirm that the demand is from the rightful owner. This study focus on a type of behavioral biometric authentication that is called Keystroke Dynamic Authentication (KDA) to identify mobile users and improve the authentication mechanism in cloud server. There are different parameters of measuring keystrokes, we defined keystrokes duration as an attribute to identify user. For implementing this method, we use Android SDK development. It includes mobile device emulator, Android Virtual Device (AVD) that helps to develop and test Android applications without using a physical device. Test of the the application is based on three different approaches (client side, local server, cloud server).In the first approach, experimental results is obtained from mobile device (client side) shows that this method works 94% correctly. In the second approach, application is connected to the php server and obtained results show that the application can works 96:15%correctly. In the last approach, obtained results from cloud server (google drive) show that the application can work 95:34% correctly. The important point in these approaches is that, the username and password were known for people cooperate in testing the proposed application. In addition, KDA is based on unique characteristics of users (here keystrokes duration) and it is hard to pretend as legible user. Therefore, applying KDA in mobile set helps to improve security and privacy of authentication.