This article presents an open-source network simulator and digital twin focused on Metropolitan Area Networks (MAN). Our tool is capable of estimating average latency and delay percentiles between every pair of nodes with minimal computational effort, with the ultimate goal to see if the MAN is capable of supporting emerging services and applications that have strict latency requirements. We show that probability bounds like Vysochanskij–Petunin’s inequality can be used to quickly estimate latency bounds with minimal computational effort. Theoretical results are validated with extensive simulations leveraging the simmer package, an open-source Discrete Event Simulator (DES) written in R, along with igraph for visualization of networks and shortest path calculations. The code is publicly available as open-source, and published in Github for the research community to use and build upon.

This article overviews OpenAI’s ChatGPT and its capability at solving classical communications and networking problems, ranging from physical layer, medium access control, queueing theory and TCP/IP. We asked ChatGPT v3.5 to solve ten classical problems in this field of science and engineering and we observed that, although ChatGPT knows the theoretical background and sounds convincing, it fails to give the final result due to intermediate mistakes in eight cases. This may be dangerous for students using ChatGPT v3.5 as an assistant in their studies of Computer Science and Engineering and should be therefore used carefully.