Service-oriented intelligent and extensible RAN

Network slicing is considered to be the enabler for a coexistence of a multitude of services on a multi-tenant 5G infrastructure. It is supported through software-defined radio access networking (SD-RAN), bringing programmability to the network in order to enhance performance according to the needs of slice owners. However, SD-RAN so far remained limited to a mere reconfiguration of the base station. In this work, we demonstrate a prototype of a service-oriented RAN on top of the OpenAirInterface and Mosaic5G platforms that brings programmability and extensibility to the RAN with a range of network applications for the purpose of intelligent slicing. We implemented a slice control and management framework, and plug a traffic analysis application that significantly improves the performance of slice users. We observe an improvement of 30% in application round-trip time with negligible variability for the considered traffic. Further, we demonstrate how to extend control plane functionality from a network store to improve slice performance.