Trojan URL Detector: A Statistical Analysis Based Trojan Detection Mechanism

摘要

Trojans' threats are on the rise: malwares attempt to install and run automatically through the way so-called drive-by downloads. To enhance security of web applications, in present study, TrojanURLDetector is proposed for the detection and blocking of Trojans URLs. Traditionally, defenders detect malware simply based on signature, which is client-based, can't share the data with other machines of Internet. TrojanURLDetector detects malicious URL based on every URL's suspicious degree. It calculates suspicious degree per time period and ultimately marked the malicious URL. TrojanURLDetector is in server-side and therefore it has sufficient data from plenty of machines accessed to web to determine whether a URL is malicious or not. Both theoretic proof and simulation results manifest that TrojanURLDetector is high efficient in Trojan URL detection meanwhile it is low misdeclaration-rate.