Volume of signaling traffic reaching cellular networks from mobile phones

摘要

Cellular networks are facing a big challenge emerging from excessive increase in signaling traffic due to background activities on smartphones. The excessive signaling if not properly managed contributes to signaling storms causing outages in mobile networks and poor quality of service, especially during peak times. For smartphone users, device-specific excessive signaling traffic caused by mobile apps can lead to poor device performance. Additionally, the cost of supporting this traffic is increasing and becoming unsustainable compared to the revenues generated. This paper aims to provide an understanding of the magnitude, scale, and the sources of signaling traffic, which is important for researchers and network operators if they are to devise effective solutions to manage the excessive signaling. We provide an analysis of characteristics of traffic resulting from background activity on smartphones and cellular networks based on datasets collected using passive (traces from a leading mobile operator) and active experiments (over 29 million logs from selected mobile apps). From the findings we deduce that the volume of signaling traffic is determined by the number of mobile devices, the kind of technologies supported by the devices, and roaming of mobile devices between different networks. Apps on the mobile devices generate background traffic without the active intervention of the device user(s). The traffic is generated in small quantities as the devices connect and tear down connections to the network periodically. Though, the average signaling traffic is tolerable, high peaks of this traffic can cause signaling floods, when some elements in the network become unbearable; this can render the network unavailable. Our findings provide network operators with a quantified understanding of the magnitude of the signaling challenge. For smartphone owners, our results provide an awareness of and insights into the impact of signaling traffic caused by individual mobile apps.