An Evolutionary Edge Computing Architecture for the Beyond 5G Era
Elli Kartsakli, J. Pérez-Romero, Nikolaos Bartzoudis, O. Sallent, Oluwatayo Y. Kolawole, Xin Tao, Swarup Kumar Mohalik, Tomasz Mach, Sige Liu, Yansha Deng, Gianluca Mandò, Angelos Antonopoulos, Valerio Frascolla, Semiha Koşu, Gökhan Kalem, Fred Buining, Eduardo Quiñones
Abstract
Beyond 5G (B5G) communication networks face the challenge of meeting the demanding requirements of various service types, including uRLLC, mIoT, eMBB, and emerging technologies like Extended Reality (XR). Edge computing can address these demands effectively because of the ability to bring computational power and resources closer to the source of data. Nevertheless, the realization of this potential necessitates an open, flexible, and automated architectural framework capable of supporting disaggregated applications and network designs. In this context, this paper introduces a novel architecture designed to advance the evolution of edge computing in B5G, developed within the EU-funded project VERGE. The proposed architecture is modular and scalable, guided by artificial intelligence (AI), and founded on three essential pillars: “edge for AI,“ “AI for edge,“ and “security, privacy, and trustworthiness for AI.“ After presenting this architecture, the paper showcases its applicability through the examination of two vertical use cases within the industrial and transportation domains.