Intent-based networking: policy to solutions recommendations

Low, Jun Sheng (2020) Intent-based networking: policy to solutions recommendations. Final Year Project, UTAR.

Preview

PDF Download (2916Kb) | Preview

Abstract

Network design and solution architecting becomes challenging when multiple constraints are involved to comply with individual network policy. The semantic diversity of policies written by people with different IT literacy to achieve certain network security or performance goals created ambiguity to otherwise straightforward networking solution implementations. In this project, an intent aware solution recommender is designed to decode semantic cues in network policies written by various demographics for robust solution recommendations. A novel policy analyzer is designed to extract the inherent intents using a custom ML model to recognize network constraints and goals to provide context-specific recommendations. There are two components: (1) a custom intent recognizer A.I. trained with network logs first normalize spectrums of policies ranging from layman to domain-specific to detect entities of interests; such as data quota, access-controls, sharing permission, etc. (2) a recommendation system based on crowd-sourced ground truth to suggest optimal solutions to achieve the goals outlined in these policies. The experimental results showed that the proposed expert system is effective in general purpose recommendations with an average score of 69% precision for different use cases and workload types.

Actions (login required)