AI-Native for 6G Core Network Configuration

3GPP envisions exploiting AI-Native for the day-to-day operations of 6G core networks (CNs). As opposed to 5G CNs whose uses for artificial intelligence (AI) are limited and not yet standardized, 6G CNs seek a revolutionary redesign where AI will no longer simply be an overlaid service but rather the foundation upon which all network functions run, i.e., AI-Native. By leveraging knowledge of the CN control plane, we utilize an AI framework to optimize our three proposed CN configurations and minimize the CN execution time which is imperative for end-to-end routing in the user plane.

I. Introduction

Although 5G networks are still yet to be fully standardized, it is well-accepted by 3GPP that the 6G core network (CN) will be a complete overhaul of its predecessor. It is projected to exploit a system level artificial intelligence (AI) framework known as AI-Native which will oversee all aspects of management, orchestration, and operation. All CN network functions (NFs), tasks, and services will be executed autonomously. Currently as it stands, AI is designed as an external NF which is an after-the-fact service. However, in complete contrast, AI will be the system in 6G CNs. AI-Native will naturally have access to all NF data, performance metrics, statistics, and alerts to inform its decision making which will undoubtedly impact NF execution, resource allocation, network prediction, security, service restoration and redundancy. The specifications of AI-Native are still evolving [1]; the standardization bodies have yet to reach that milestone. However, there have been preliminary investigations into integrating AI at a level expected to be conducive to AI-Native (also known as Native-AI in some literature) within the radio access network (RAN) [2], [3] and CN [4].