KlonetAI: Automating (Com)²Nets Management With Human Language Intents

The communication and computation integrated network architecture ((Com)2Net) simplifies the deployment of applications such as distributed machine learning. However, the management and operation of the (Com)2Net pose significant burdens and consume a substantial amount of time due to its large-scale and complexity nature. Intent-based networking technology and Large Language Models (LLMs) are beneficial for the automation configuration and management of (Com)2Nets. In this article, we introduce a network automation configuration and management architecture. It implements network automation configuration based on user intents in natural language, reducing the network management complexity. The network management agent, based on a LLM, is proposed to translate user intents in natural language into machine instructions. Finally, we present two use cases and real-world evaluations to validate the performance of this architecture. The demo and code are open available on: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/Lizonghang/KlonetAI</uri>.