A Multi-agent System for Smartphone Intrusion Detection Framework

There has been significant growth in the number of malware using the Android platform, especially malware that target Short Message Services (SMS). The mobile botnet has been using SMS as a channel to distribute spam, send unauthorized SMS messages without user knowledge, use command and control (C&C) channel, and attach malicious URLs. With the limitation of Android smartphone resources, a multi-agent technology can make our framework to be more robust and efficient. In this paper, we propose a multi-agent system that is currently being developed using JADE platform for observing Android Smartphone features and monitoring SMS services, as well as creating Android profiles. Our framework applies hybrid detection approaches in order to counteract botnet attacks, by investigating damaging SMS botnet activities through the examination of Smartphone behaviour. These approaches utilize multi-agent technology to recognize malicious SMS and prevent users from opening these messages, by applying behavioural analysis to find the correlation between suspicious SMS messages and the profiles reported by the agents.