Abstract
Advances in micro-electro-mechanical systems (MEMS), computing hardware, and software algorithms for wireless sensor networks (WSNs) have boosted the adoption of WSNs, also known as machine-to-machine (M2M) communications systems, in many fields, including vehicle tracking, supply chain management, security, and healthcare. Due to the large scale of the deployments of many commercial applications, and the diversity of the hardware and software, the management of these M2M communications systems is becoming more and more cumbersome. Motivated by the challenges posed by a real-life commercial M2M communications system featuring millions of heterogeneous devices connected to hundreds of applications using a GSM cellular network, we developed a cloud-based solution for cellular traffic analysis aimed at M2M communications systems. The proposed system captures and stores all traffic generated by the M2M communications system 24/7, and can process and analyze one day worth of traffic in 2.5 hours for $2–3 using cloud computing. We also report on case studies, where the proposed solution was employed to detect misbehaving devices and test different configuration for select devices.