Analysis Of Iot Scenarios And IoT Trffic Modeling

Promoted by sensor, big data, mobile computing and network virtualization technologies, Internet of Things (IoT) services are increasing rapidly. At the same time, due to the deployment of large numbers of intelligent terminal equipments, a considerable amount of IoT traffic is generated and transmitted to the core network. Such substantial amount of the IoT traffic data have a significant impact on the traditional network and the network performance. However, currently not much work related to the analysis of the characteristics of IoT applications has been done, not many results related to the IoT traffic modeling have been obtained.There is a paucity of research that analyzes the characteristics of IoT application. Consequently, limited results are available regarding IoT traffic modeling in the extant literature. First of all, by reviewing the literature on the IoT application, we analyze and summarize the network and traffic characteristics of the applications and classify the application according to the different business areas. As the scenario of Smart City involves several representative IoT applications, this thesis selects Smart City to perform the traffic modeling.For the purpose of traffic modeling, this thesis investigates the traditional mathematical modeling methods, evaluate the advantages and disadvantages of various mathematical models and their applicable scenarios. Then, the Gamma model with wavelet modulation is proposed,which can fit the data of IoT applications. The traffic self-similarity and network scale related to the proposed model have also been analyzed.Moreover, we set up traffic generation rules of intelligent campus,intelligent transportation, intelligent shopping malls, intelligent hospitals and other sub-scenes on NS-3 simulation platform, and collect traffic in the various sink nodes. The collected data of Smart City are fitted using Matlab tools. The model parameters and flow characteristics are compared and analyzed, and several conclusions are drawn. The accuracy and general applicability of this model are verified by several tests.