The Influence Of Packet Traffic On Mobile Communication System:a Study Based On Data Collection Platform

According with the rapid growth of the mobile Internet, packet switched data traffic has being the major traffic type on mobile communication networks, promoting the human society development.The original purpose of mobile communication system is carrying circuit switched voice traffic, so the architecture, protocol stack and signaling system has been well-designed for voice traffic. Before the third generation, such design effort has a good performance. However, with the prosperity of the mobile Internet, packet switched data traffic has replaced circuit switched traffic. Thus, the traditional optimization is no longer appropriate for current mobile communication system. The characteristic of packet switched data traffic is differ from circuit switched traffic, as its duration is shorter, its origination is more frequent and its traffic volume is small. Under its high occurrences, the different character of packet switched data traffic bring mobile communication a big challenge. Entering the third generation era, many mobile operators have encountered the so called signaling storm problem. The signaling storm is aroused by the high occurrences of packet data traffic as mentioned before. Because of the short duration and frequent origin, packet data traffic lead the connection between user equipment and network to re-establish and release frequently. This malfunction is a strong evidence to prove that current mobile communication system is improper for the high occurrences of packet switched data traffic.This article aims to analyze the influence of packet switched data traffic on mobile communication. To solve this problem, we first study the signaling system and protocol stack of common mobile communication system. User equipment should establish the connection to network before send or receive any data. Packet data traffic is short in duration and its occurrences is frequent, such characters may lead the connection between users and network to be built largely. Large number of connection establishment cause large signaling numbers of network, even block the control plane. The signaling procedure of service request, connection management and conversation management is examined carefully. Based on the research of existing protocol, we study the high occurrences of packet data traffic how influence the mobile communication system theoretically. Using probability theory and convex optimization, we describe the relationship among packet data traffic occurrences, network signaling load and radio resource utilization. To better understanding this problem, we build a data collection platform in Android system. Current studies on how packet data traffic influence mobile networks often model the behaviors of data traffic. Without real data, this method is inaccurate and lack of flexibility. Our data collection platform provides an easy way to accurately gather traffic data trace. Deploying this platform large-scale, researchers can not only study data traffic, but also study other user behaviors.