How an LLM query works
A single LLM query travels through far more infrastructure than most users realise.
It begins on an office computer, where the request is encrypted and sent through the local network.
From there, it passes through the router, ISP, and high-capacity internet backbone.
The request then reaches the AI provider’s edge network and is routed into a data centre.
Inside the facility, load balancers and network switches direct it to the correct inference servers.
…..