Applying Big Data, Machine Learning, and SDN/NFV to 5G Traffic Management

• Main Authors: Le Luong Vy, 黎梁偉
• Other Authors: Lin, Bao-Shuh Paul
• Format: Others
• Language: en_US
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/94283u

博士 === 國立交通大學 === 電機資訊國際學程 === 107 === Traffic management plays a crucial role in improving network efficiency, network quality, load balancing (LB), and energy saving of mobile networks. Especially, in 5G networks, a dense heterogeneous architecture of various types of cells (macro cells and small cells) makes traffic management more complicated. Therefore, investigating and understanding traffic patterns of a huge number of cells are challenging issues, but valuable for network operators. Moreover, due to the rapid growth of mobile broadband and IoT (Internet of Thing) applications, the early-stage mobile traffic classification becomes more important for traffic engineering to guarantee Quality of Service (QoS), implement resource management, and network security. Hence, identifying traffic flows based on a few packets during the early-state has attracted attention in both academic and industrial fields. However, a powerful and flexible platform to handle millions of traffic flows is still challenging. On the other hand, big data, machine learning (ML), software-defined network (SDN), and network functions virtualization (NFV) have recently been proposed as emerging technologies and the necessary tools for empowering the SON of 5G to address the intensive computation and optimization issues. In this dissertation, the authors applied those technologies to build a practical and robust framework for clustering, forecasting, and managing traffic behaviors for a huge number of base stations with different statistical traffic characteristics of different types of cells (GSM, 3G, 4G). Besides, several applications based on traffic forecasting are also introduced such as energy saving and abnormal detection. Moreover, based on this framework, we successfully implemented an early state traffic classification, network slicing, and QoS control to configure priorities per-flow traffic to enable bandwidth guarantees for each mobile broadband traffic application. Finally, the performance of the proposed models is evaluated by applying them to a real dataset that collected traffic KPIs (key performance indicators) a real network.