Monitoring Traffic Classification in an ISP

• Autor(a) principal: Henriques, João Pedro de Figueiredo
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/159640
• topic: Traffic Classification; Monitoring; Automation; Robot Framework; Appium; ADB; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática

The use of mobile applications has been increasing exponentially in recent years, which in turn leads to the need to manage a large flow of Internet traffic. Service providers offer their customers plans that include exemption from taxation of traffic consumption for some applications and certain quality guarantees for them, therefore monitoring the traffic classification is important to maintain the quality of the service offered. In this dissertation, a prototype was developed in order to monitor the accuracy of the traffic classification by the 4G network core for some of the most popular mobile applications. The solution presented for monitoring the core classification was to measure the amount of data consumed by the application directly on the test devices (only Android mobile phones were used in this work) and compare it with the result given by the core classification. The project was developed in a 4G network of a service provider. This project was also an automation work, since the entire system works automatically and there is no need for human interaction. Real devices were used to test the classification. Navigation in applications was automated using tools such as Robot Framework, Appium, and ADB. The Robot Framework proved to be very strong in automating tasks, both at the mobile level and in validating the core classification. A dashboard with graphs was built to analyze the performance of the classifiers over time, using the Grafana tool. To manage the launch of tests periodically and automatically, the Jenkins tool was used. A network data collection mechanism for mobile applications was developed, using automation techniques, which can be used to implement other types of classifiers, such as ML or DL.