Design and Implementation of a Trust-based Framework for Network Mobile Traffic Detection

Nowadays, with the widespread growth of mobile devices and their usages in most people’s daily lives, a main concerns is handling and responding their related security incidents in the area of mobile network traffic analysis. The main objectives of security researchers include mobile application identification, mobile malware detection and user’s personal information leakage. Identification and traffic analysis of mobile devices is more difficult than other network devices like traditional ones due to their dynamic nature. Although a lot of research has been done in this area, most of them are not efficient for identifying in real traffic due to challenges such as encrypted traffic and data manipulation. In this paper, we propose a trust-based mobile traffic detection system which is able to handle various real traffic. The proposed framework consists of four detection subsystems. First, the environment identification subsystem examines validity of the information for each network layer. If the information of each layer is valid, then traffic is sent to the detection subsystem. After detection of each subsystems, the final detection is performed according to the obtained factors in each subsystem. For the evaluation purposes, a suitable dataset is used which is collected from wireless network traffic of Ferdowsi University of Mashhad campus in Mashhad, Iran. The experimental results show that aggregation of subsystem detection based on trust and confidence is efficient in analyzing real mobile traffic with maximum accuracy, 0.97, and minimum error rate, 0.09, under special circumstances.