A Vision For Hierarchical Federated Learning in Dynamic Service Chaining
Abdullah Bittar, Changcheng Huang · 2022
We have seen the tremendous expansion of machine learning (ML) technology in Artificial Intelligence (AI) applications, including computer vision, voice recognition, and many others. The availability of a vast amount of data has spurred the rise of ML technologies, especially Deep Learning (DL). Traditional ML systems consolidate all data into a central location, usually a data center, which may breach privacy and confidentiality rules. The Federated Learning (FL) concept has recently emerged as a promising solution for mitigating data privacy, legality, scalability, and unwanted bandwidth loss problems. This paper outlines a vision for leveraging FL for better traffic steering predictions. Specifically, we propose a hierarchical FL framework that will dynamically update service function chains in a network by predicting future user demand and network state using the FL method.