Teaching My Lab to See the Future: Bandwidth Prediction with Two Routers, a Time-Series Database, and a 337-Number Brain
How a CiscoU post sent me down a rabbit hole, and what came out the other side: a tiny, fully reproducible lab where real packets train a real neural network to predict congestion before it happens. This is the whole chapter every step, every line of code, and what actually happens under the hood.
1. The post that started it
A few weeks ago I was scrolling through a CiscoU article about predictive networking the idea that a network shouldn’t just tell you it’s congested, it should tell you it’s about to be. The article was good, but it stayed at the concept level, and I kept thinking the same thing I always think when I read these pieces:
Fine. But could I build one?