Open-Source Edge AI for 6G Wireless Networks

Abstract

Multi-access Edge Computing (MEC) has been recognized as a key enabler for next-generation networks in supporting a large variety of compelling applications with challenging requirements. With its widely proved strength and successes, AI has to become an integral part of MEC. In this paper, we present a novel open-source edge AI (OpenEAI) framework that introduces a native AI plane into the recently proposed open-source MEC framework. The AI plane is designed based on two principles: decoupling the edge AI services into independent AI functions; and recomposing the independent edge AI functions into customized OpenEAI instances based on users’ specific requirements. Typical use cases of OpenEAI are characterized with the aid of a small-scale test network. Finally, we discuss the opportunities and challenges facing OpenEAI.