Control and management mechanisms in 5G networks: operator networks

• Autor(a) principal: Ferreira, Diogo Daniel Soares
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/29643
• topic: Time-series prediction; Forecasting; 5G; Network Monitoring; Machine Learning; Prediction Architecture; Network Policies; Network Slicing

In 5G networks, time-series data will be omnipresent for the monitoring of network metrics. With the increase in the number of Internet of Things (IoT) devices in the next years, it is expected that the number of real-time time-series data streams increases at a fast pace. To be able to monitor those streams, test and correlate different algorithms and metrics simultaneously and in a seamless way, time-series forecasting is becoming essential for the pro-active successful management of the network. The objective of this dissertation is to design, implement and test a prediction system in a communication network, that allows integrating various networks, such as a vehicular network and a 4G operator network, to improve the network reliability and Quality-of-Service (QoS). To do that, the dissertation has three main goals: (1) the analysis of different network datasets and implementation of different approaches to forecast network metrics, to test different techniques; (2) the design and implementation of a real-time distributed time-series forecasting architecture, to enable the network operator to make predictions about the network metrics; and lastly, (3) to use the forecasting models made previously and apply them to improve the network performance using resource management policies. The tests done with two different datasets, addressing the use cases of congestion management and resource splitting in a network with a limited number of resources, show that the network performance can be improved with proactive management made by a real-time system able to predict the network metrics and act on the network accordingly. It is also done a study about what network metrics can cause reduced accessibility in 4G networks, for the network operator to act more efficiently and pro-actively to avoid such events