Machine learning-based network modeling: An artificial neural network model vs a theoretical inspired model

Recent trends in networking are proposing the use of Machine Learning (ML) techniques for the control and operation of the network. The application of ML to networking brings several use-cases as well as challenges. In this paper we focus on a rather fundamental problem in networking: estimating the delays of a network. In particular, we aim to assess the performance of models inspired by the existing knowledge of network modeling in comparison to generic adaptive machine learning models. To do so, we present a M/M/1-inspired ML regressor and compare its performance to a neural network.