On SPIM detection in LTE networks

The telecommunication enterprise networks are continuously facing increasing amount of messages and multi-media traffic due to growing deployment of messaging applications. The detection of SPAM over Instant Messaging (SPIM) is essential for the mobile service providers to achieve user satisfaction and efficient network load balancing. However, accurate SPIM detection is challenging at the server side due to large volume of data packets and near-real-time communication of messages. In this setting, we propose a message classification approach to collect valuable insight on SPIM inside LTE cellular telecommunication framework. The contribution of this paper is threefold: First, it identifies the location and scope of the SPIM detection component inside LTE architecture. Second, we compare the detection accuracy, training and testing time of different classification algorithms. Finally, we evaluate the additional delay in messaging for adding such a module inside LTE architecture using an OPNET simulation environment.