Net-in-AI: A Computing-Power Networking Framework with Adaptability, Flexibility, and Profitability for Ubiquitous AI

Along with the unprecedented development of artificial intelligence (AI), a considerable number intelligent applications are universally recognized to significantly facilitate evolution anthropogenic activities. The abundant AI computing power is one main pillars fuel booming ubiquitous applications. As proliferates multitude network edges, even end devices, networking function bridges gap, on hand, among ends-edges-clouds, other between multiple and heterogeneous requirements. emerging new opportunities have spawned deep integration networking. However, complete integrated system under-addressed, including adaptability, flexibility, profitability. In this article, we propose computing-power framework for by establishing Networking in pool, denoted as Net-in-AI. We design enable adaptability users, flexibility networking, profitability providers. then formulate computing-networking resource allocation problem, joint perspective these three aspects. Experimental results prove superior performance proposed comparison current popular schemes.