The writing here comes out of running the thing: what it takes to hold connections open across every region we cover, why clock synchronization turns out to be a correctness problem rather than a housekeeping one, and what a latency figure has to include before it means anything.
We would rather publish one piece with a measurement in it than four with a thesis. If you want the vocabulary rather than the argument, the glossary defines the terms these pieces lean on, and the edge network page explains the architecture they keep referring back to.
The recurring subjects are the ones that decide whether a dataset can be trusted: how event time is established and why arrival time is not a substitute, what a delivery guarantee obliges a consumer to handle, why point-in-time history is the difference between a backtest and a story, and what changes for a research desk when provenance travels with every record rather than living in a report.