A Brief Survey and Implementation on AI for Intent-Driven Network
Jiaorui Huang, Chungang Yang, Shiwen Kou, Yanbo Song
Abstract
Intent-driven network (IDN, or intent-based network, IBN) is a novel networking paradigm, which can enable user intents to drive network management autonomously and improve the network’s operational efficiency. Although artificial intelligence (AI) has been found for several applications to the IDN, there lacks a systematic discussion and research on this topic. In this work, we present a survey of the application of AI at each layer of IDN. Then, a general IDN management architecture, State-Action-Intent (SAI), is proposed. The presented SAI is a new IDN implement framework to automate the operational intents in a closed loop to overcome the challenges of complex network services. To verify the availability and effectiveness of SAI, a proof-of-concept demonstration is provided, and the obtained performance is discussed.