Bare Metal Sizing Calculator

Work out bare metal sizing instantly with clear inputs, formula shown and shareable results.

About Bare Metal Sizing Calculator

Bare-metal sizing takes the larger of the CPU and memory requirements, each derated for target utilisation — 75 percent on cores and 85 percent on memory is typical — then adds one server for N+1 redundancy so a single failure does not reduce capacity below peak demand. Which resource drives the count tells you what to specify differently next refresh.

Formula

Server count

CPU count = ceil(vCPU demand / (cores x 0.75)); memory count = ceil(memory demand / (RAM x 0.85)); servers = max + 1 for N+1

Worked example

With the default values below, Bare Metal Sizing Calculator returns Servers to buy including N+1 of 8. Change any input above and every figure here is recalculated in your browser.

Inputs used

Peak vCPU demandminimum 1: 320
Peak memory demandminimum 0.1 GB: 1,400 GB
Cores per serverminimum 1: 64
Memory per serverminimum 1 GB: 512 GB

Results

Servers to buy including N+1: 8
Servers needed for capacity alone: 7
Count driven by CPU demand: 7
Count driven by memory demand: 4

Step by step

1Values used
2Server count
3Servers to buy including N+1= 8
4Servers needed for capacity alone= 7
5Count driven by CPU demand= 7
6Count driven by memory demand= 4

Assumptions and limits

The result is derived only from the 4 inputs listed above. Anything not among them is not part of the calculation.
In particular, this calculator does not take inflation, tax, fees or charges or compounding frequency into account.
Figures are rounded for display to whole numbers, so totals built from several rounded parts can differ slightly from the unrounded arithmetic.

How these calculators are built and tested is described on the Methodology page. If a figure looks wrong, please report it.

Frequently Asked Questions

Why derate CPU more heavily than memory?

Because CPU demand is spikier. Memory is allocated and held, so it can be planned closer to the limit provided there is no overcommit.

Is N+1 enough?

For a single rack or fault domain, usually. Across availability zones you need enough spare that losing a whole zone still leaves capacity, which is N+1 per zone.