RECAP 5GC: Resilience and CAP aware 5G Core for Consistent and High Availability Service

Being scalable, distributed by nature, and stateless by design, the Service Based Architecture (SBA) of 5G Core (5GC) has become the popular approach today, to provide reliable services to 5G users. However, the well-known Consistency (C), Availability (A), and Partition (P) tolerance (CAP) theorem places a limit on 5GC, as it typically does for any distributed system to choose between consistency and availability, whenever a partition occurs in the system. Hence, choosing between ensuring consistency for users’ state information and the availability of the service while being resilient against unforeseen sudden failures becomes a critical need for Beyond 5G (B5G) and 6G networks. In this paper, we present RECAP 5GC: a REsilience and CAP (RECAP)-aware 5GC that keeps up the High Availability (HA) service by being resilient and consistent during sudden failures of stateless functioning. To prove the feasibility and effectiveness of the proposed solution, we implement it entirely on the 3GPP-compliant control plane of the 5GC testbed between the Access and mobility Management Function (AMF) and Unstructured Data Storage Function (UDSF). Experiments on the minimal two nodes active/active cluster configuration prove that the proposed RECAP 5GC is not only resilient by dropping 47% fewer user requests than the CP mode and 86% fewer user requests than the AP mode but also 1.6% additionally available than the CP mode while remaining consistent on the user’s state and information.