Toward hardware-accelerated QoS-aware 5G network slicing based on data plane programmability

The diverging requirements from various vertical industries have driven the paradigm shift in the next-generation (5G) mobile networks, where network slicing has emerged as a major paradigm for this purpose by sharing and isolating resources over the same 5G physical infrastructure. To truly fulfill the different quality-of-service (QoS) requirements imposed by different network slices for different vertical applications, it is essential to introduce a programmable data plane that is aware of QoS and is configurable to enforce the QoS commitments. In this paper, we focus on designing, prototyping, and evaluating a novel QoS-aware data-plane network slicing framework for the edge and core network segments of a 5G network. The proposed framework is capable of dealing with differentiated services through hardware-based traffic classification, priority configuration, and traffic scheduling. By leveraging the latest open-source field-programmable gate array platform, we prototype the proposed framework and empirically evaluate the performance of the prototyped system. Experiment results demonstrate the capabilities of the proposed framework in terms of achieving QoS-aware network slicing at the data plane.