Internal Attacker Detection by Analyzing User Keystroke Credential
Dwijen Rudrapal, Smita Das, Nikhil Debbarma, Swapan Debbarma · Lecture Notes on Software Engineering · 2013
Abstract—Considering the recent Internet-based attack, many organizations are so focused on firewall and web server security. As a result, many organizations fail to realize the looming threat of attack from the inside. Based on recent few studies it could be said that insider attacks represent the most prevalent of all threat. More complicated issue is that insider attackers are not detectable with only the classic cryptography based techniques. In an organization employees used to login by ID and password. But password is not enough strong to maintain its privacy in this era of technology. In this paper the proposed algorithm considers multiple attributes of user keystroke dynamics which in addition of traditional authentication has applied in an organization for distinguishing one user than another. In earlier studies of keystrokes, we observed degree of disorder for keystroke duration and latency and their uniqueness. Along with standard deviation of keystroke duration feature, we proposed this algorithm which helps to find an internal attacker easily. More over it is done without letting inform any user and without use of any extra hardware. The obtained experiment result disclose that even though internal attacks can't be eliminated but based on keystroke dynamics values internal attackers can be identified.