Conceptual design of proactive SONs based on the Big Data framework for 5G cellular networks: A novel Machine Learning perspective facilitating a shift in the SON paradigm
Bharath Keshavamurthy, Mohammad Ashraf · 2016
Self-Organizing Networks (SONs) are being researched extensively in the existing 3G and 4G landscape primarily to facilitate a convenient yet cost-effective approach in the configuration, optimization and troubleshooting of networks. However, the existing SONs will be no match for the operational complexity of the envisioned 5G networks. The promise of 5G revolves around the premise of infinite capacity and zero latency. 5G networks are set for commercial availability by 2020 and these networks will be part of a flexible and dynamic telecom ecosystem supporting cross-domain integration and multi-RAT environments. Network Densification, Network Function Virtualization, Flexible spectrum allocation, E2E security and Massive MTC are a few of the features promised by 5G networks. With this assortment of numerous technologies, it is obvious that SONs have to evolve beyond the existing reactive paradigm without which the maintenance and management of 5G networks would prove to be a herculean task. Hence, we propose a novel Proactive SON methodology based on the Big Data framework to enable the shift in the SON paradigm. In this article we present a comprehensive Big-Data based SON framework involving innovative Machine Learning techniques which would cater to scalability and programmability of 5G networks with respect to availability, reliability, speed, capacity, security and latency.