Quality of service management in telecommunication network using machine learning technique

Abstract

Designing and implementing a fail-safe, real-time telecommunications network is complex. In modern networks, traditional quality of service (QoS) methods for monitoring and analyzing data have some problems, such as accuracy and efficient processing of big data in real time. To solve this problem, should use an appropriate intelligent crash classification system to detect and diagnose run-time errors. The article proposes to use a comprehensive fault detection system that includes QoS and machine learning technologies using information about the state of a point-to-point protocol over ethernet (PPPoE) session on PPPoE active discovery termination (PADT) virtual local area network (VLAN) routes. This intelligent system is built using the machine learning method and is independent of the main real-time system. Demonstrated the operation of seven machine learning algorithms and presented the results of training and fault detection. Based on the received information about the state of the PPPoE session, the PADT type allows you to control the behavior of the real-time system.