Scalable detection of web malware on smartphones

Recently, the smartphone industry has seen tremendous growth due to the widespread adoption of devices based on Google's Android and Apple's IOS platforms. The worldwide market penetration of Android based smartphones has attracted the attention of malware developers. This paper presents a scalable classifier to detect web malware on Android smartphones. Limited computational and energy resources on smartphones make it infeasible to deploy rich featured security mechanisms. We extracted network and URL inspection features from over two million mobile URLs. A scalable classier is applied to the measured features and implemented on MapReduce/Hadoop based cloud computing platform. Performance evaluation of the system has shown that 99.8% accuracy is achieved with a response time of 120 ms.