Traffic Classification of QoS Types Based on Machine Learning Combined with IP Query and Deep Packet Inspection

Due to many network services using encryption protocols, it poses a massive challenge to accurately identify traffic and provide quality of service (QoS). Although deep packet inspection (DPI) technology can identify most traffic, there are still many disadvantages to identifying encrypted traffic. In this paper, we apply the IP (Internet Protocol) autonomous system information query with machine learning (ML) and deep packet detection technology for traffic identification. By combining these methods, it can quickly provide good QoS for different types of encrypted traffic. The experimental results show that different encrypted traffic can be effectively identified. The services with different QoS requirements can be effectively identified even if they use the same protocol.