الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 134 |  |  |
| | | from | 134 | | |
Abstract
The integration between fog computing and software defined networks introduce new concept called network function virtualization. 5G networks will depend on this NFV which simplify Big data analysis through better spectrum usage and network planning. In this thesis, we present a proposed platform that integrates the Software-defined networks (SDN) technology with fog computing. This platform employs docker as a virtualization technique to implement a container on any available IoE device with high computational power. Also, we discuss the main requirements of this platform such as data analytics engine, management module, the main building blocks of SDN controller and IoT devices. We think that the proposed platform will provide a suitable environment for IoE applications demands. In addition to that, we introduce a new offloading distributed decision making mechanism based on the game theory, which helping user nodes to organize their transmission through learning from the environment. We formulate the communication between a single fog node and multiple IoE devices as a minority game. This suggested approach organizes the user nodes transmission around an equilibrium level defined by the fog node controller. The suggested approach helps in reducing network usage and the number of dropping tasks with lower complexity approach. Moreover, we present a shared algorithm that encourages the fog controllers to share task processing between fog nodes based on a multilevel placement strategy. We apply several algorithms such as greedy, dynamic programming, genetic and mixed-integer programming algorithms. The multilevel placement algorithm based on mixed integer programming achieves good services placement with low space and time complexity, and without firing the application deadline. We also investigate the impact of increasing the number of fog nodes with reducing the transmitted tasks to the cloud. The results show that with increasing the numbers of fog nodes, the transmitted tasks to the cloud decrease. Finally, we utilize data mining and machine learning tools to analyze the generated data from IoT devices and discover the nodes that can be used as fog nodes.