Application-Aware Fine-Grained QoS Framework for 5G and Beyond

In the face of diverse traffic types, existing 5G standards have adopted a flow-based Quality of Service (QoS) framework and introduced dual mapping to enhance resource allocation flexibility. However, the cognitive ability of these standards, particularly in terms of application traffic recognition (based on network features, eg. IP five-tuples), remains relatively rudimentary. This limitation results in coarse-grained QoS flow classification. Consequently, it’s unable to cater to the diverse requirements of priority requirements or service quality needs of applications. To overcome this limitation, we extend our previously proposed cognition-driven core network architecture by developing a cognition-based fine-grained QoS control framework. This framework employs service tags to accurately disclose the QoS requirements of application traffic, enabling 5G networks and beyond to provide more precise services. We provide an in-depth discussion of the framework’s internal processes, service tag definitions, and implementation. Through experiments conducted on both the core network and Radio Access Network (RAN) sides, our framework effectively enhances the user’s quality of experience (QoE).