Capacity planning, sizing and total cost of ownership, modelled as code. A self-study book and toolkit built around one question — how big, how much, and how wrong could I be? — and one rule for answering it: every number says where it came from.
Topics Covered¶
What You’re Sizing: point estimates, what a workload is, where the numbers come from, peak, mean and growth
Ceilings: Little’s Law, queueing and the knee, when adding servers stops helping, regime changes
Sizing: capacity, bandwidth and the binding constraint, headroom and failure domains
Uncertainty: Monte Carlo from first principles, correlation and convergence
Cost: capex, opex and lifecycle, power first, unit economics, the five-year model
Sensitivity: which input is really the answer, and the missing node
Presenting It: a TCO for finance, and comparing two TCOs
Appendices: DSL reference, the Monte Carlo module, distributions, units, two worked models, glossary
What Makes It Different¶
Models are code — a model is a graph of named quantities in a YAML file, evaluated and version controlled like anything else
Units fail the build — every node declares a unit, so a model that multiplies the wrong two things is rejected rather than published
Provenance is mandatory — every input declares whether it is a measured fact, a vendor’s claim, or an assumption
Gaps stay visible — a constant nobody has measured leaves the nodes below it empty instead of quietly filled in
Monte Carlo from scratch — inverse-transform sampling, rank correlation and convergence written to be read, not called
Access the Full Book¶
The complete book is available as an interactive resource, with worked models you can adjust and re-run in the browser. Source and toolkit on GitHub.