A Brief Survey and Implementation on Refinement for Intent-Driven Networking

Abstract

Intent-driven networking (IDN) is emerging because the traditional network is complicated and error-prone when configured. As the key step of IDN, intent refinement is an alternative approach to convert intents from declarative language to a machine-readable policy, which possesses important scientific significance and application value. In this article, we first present a generic architecture to illustrate the process and functions of IDN. Then we give a definition of intent refinement to make the concept clear. In order to propose a standard classification method of intent refinement, we review the typical intent refinement schemes and distinguish them according to target users, input methods, and refinement approaches. Finally, in order to refine different kinds of intent, we design an intelligent intent refinement system based on natural language processing and deterministic finite automation. In summary, intent refinement plays a crucial role in IDN, which provides a convenient northbound interface for different users to express their communication requirements. Therefore, it is of vital importance to summarize its common methods and explore more straightforward and intelligent ways to enhance its utility.