Exploring Behavioral Patterns for Security in Cloud Computing: a Case Study

The Internet has now become a powerful enabler to access data and services offered on cloud platforms. Ensuring the security of cloud services for customers leveraging it is a challenge. The use of passwords, OTPs, certificates, biometrics, Multi-factor, OpenID, Kerberos, challenge-response, and behavioral-based authentication techniques is in use in the industry. Habit, which is a behavior that becomes automatic and is repeated often, with little to no thought, while interacting with the environment is recordable due to smart electronic gadgets. These can record the user interactions in logs using different software applications that can be analyzed and then used for providing safety in a cloud computing environment. The patterns of interest include time of login, likes, dislikes, comments, keystroke speed, errors in typing, reading, and comprehension, use of different types of devices, machines used to access the internet, trends, search phrases, reading types of news feeds can be explored for possible enhancement of security. This paper is focused on exploring the use of reading, comprehension, and typing behavioral patterns of the users to provide proactive threat-driven security while accessing cloud applications. It is the first step towards analyzing the behavioral patterns that can be used to re-authenticate the users for enhanced security of devices, applications, and services like cloud computing. The simulations tested the system’s ability to differentiate between fake and pivot users as group size is increased. While accuracy was 100% for smaller groups upto 30, it slightly decreased to 97.5% and 94% for larger groups with 2 to 4 pivot users and total of more than 40 users in all. The study is useful to minimize the malicious users’ access to services meant for the valid intended users of cloud computing.