Resource-Aware Service Prioritization in a Slice-Supportive 5G Core Control Plane for Improved Resilience and Sustenance

Abstract

Providing resilient and sustained service is quite challenging in the Service Based Architecture of distributed 5G Core (5GC) as multiple Network Functions (NFs) are involved to help serve the various User Service Requests (USRs) arriving in the control plane. In this regard, the continuous monitoring of individual NFs in a Closed Loop Automation (CLA) is a need of hour to keep up the robust and resilient functioning of the 5GC overall. Any unforeseen situations like the sudden failure, overload, or congestion of the NFs of the 5GC can drop the critical USRs unnecessarily. This paper proposes the proactive monitoring of the NFs of the 5GC in the control plane and utilizes it to intelligently schedule and serve the frequently arriving USRs and prioritize the critical slice service requests. Specifically, the Ford-Fulkerson algorithm popularly known as the Max-Flow problem solver is leveraged to proactively assess the NFs' performance and availability and use it effectively to serve critical service requests arriving during unexpected situations of failure and overloads. Our experiments based on the 3GPP-compliant 5G testbed show that, with the proposed solution, the native 5GC can serve 20% more predominant USRs, and the slice-supportive 5GC can serve 33% more massive Machine Type Communications (mMTC) slice USRs, and 47% more ultra Reliable Low Latency Communications (uRLLC) slice USRs while handling their respective peak traffic.