查看文章

Currently, network operators take much time to learn network configuration commands for their equipment. To support the learning, Intent-based networking (IBN) is proposed, which aims to translate the network operator's intent described by natural language to real network configuration. Ref. [1] is the latest IBN scheme, which can translate the operator's intent into the network commands. However, as shown in the left of Fig.1, it must pass the intermediate language translation step, which needs lots of translation rules. To save this cost, we propose a new IBN scheme without the intermediate language translation. Our IBN scheme directly finds corresponding network commands for an operator's input by using deep learning. In evaluation, we confirmed our IBN scheme can extract network operator's intents and generate configurations with a high probability.