Leveraging Artificial Intelligence to Improve Quality of Service in Next-Generation Broadband Networks

Abstract

In recent years, Operators are confronted with new challenges regarding the Quality of Service (QoS) they provide to their users. Changes on the traffic load after the emergence of smartphones and other mobile devices with capacity for using real-time applications have produced an overloading excess of many mobile networks. New paradigms such as the Internet of Things (IoT) and Smart Cities are expected to generate a massive traffic growth that current fixed and mobile networks may be unable to support. In addition to this, the quality of service perceived by the users is becoming the most important selection criterion in the choice of the service provider. Providers need an exhaustive management of the Internet Protocol (IP)-based services that they offer, and specifically, of the related QoS parameters. A joint approach is proposed to leverage Artificial Intelligence (AI) techniques to achieve an accurate monitoring of QoS parameters that impact on user experience. Data imported from the provider, application and content server domains enables to train Random Forest, Support Vector Machine and Decision Tree models, which are used to predict target QoS parameters values. A congestion classifies conditions of the multimedia content targeting. An extensible framework for dynamic managing of the QoS of the services offered is provided. Monitoring tools. QoS Monitoring Framework is devised for “on-the-fly” monitoring QoS controlling of usage by classifiers and models. Classifications of the IP-based services offered are needed and relevance of the parameters is estimated to identify targets for applying those policies. A clear approach should be defined on the policies available in the context of controlling the service provisioning (“active”), specification of the QoS to achieve (“passive”). A framework for evolving the adequate policy actions according to the usage of the services is required. Ensuring that customers experience a certain level of QoS has become an important issue in the design considerations of networks. To achieve this goal, it is important to identify those parameters that need to be controlled and those that impact on the customer perceived service QoE.

Topics & Concepts