Abstract
The ever-increasing number of mobile devices is heavily modifying the traffic observed in cellular networks. From smartphones and tablets to Machine-to-Machine (M2M) devices, the traffic volumes and patterns generated by end-user and M2M applications introduce novel challenges to cellular network operators. One of these relates to the detection and diagnosis of network traffic anomalies caused by specific devices and applications. We introduce a simple yet effective approach to detect and diagnose such anomalies, applying entropy-based analysis on top of device/application-related descriptors. As case study, we present the analysis of a large scale traffic anomaly observed in a real cellular network, linked to smartphones. Our diagnosis approach promptly revealed a failure of a specific OTT (Over The Top) service not linked to the operator, showing its paramount advantage from an operational point of view.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
