Skip to main content
Technology & EngineeringNetworking Infrastructure37 lines

Data Center Capacity Planning

Model rack, power, cooling, network and failover constraints with measured load and explicit growth assumptions.

Quick Summary7 lines
Use this skill to assess whether a facility can host a planned workload and its growth. Establish the site, planning horizon, service objectives, equipment inventory and responsible facilities engineers. A planning model supports their design review; it does not replace electrical or mechanical engineering approval.

## Key Points

- [US DOE data center design guide](https://www.energy.gov/sites/default/files/2024-07/best-practice-guide-data-center-design_0.pdf)
skilldb get networking-infrastructure-skills/data-center-capacity-planningFull skill: 37 lines
Paste into your CLAUDE.md or agent config

Data Center Capacity Planning

Use this skill to assess whether a facility can host a planned workload and its growth. Establish the site, planning horizon, service objectives, equipment inventory and responsible facilities engineers. A planning model supports their design review; it does not replace electrical or mechanical engineering approval.

Build a consistent baseline

Inventory rack units, weight, device power, feeds, cooling zones, network ports and uplinks. Record observed average and peak demand over representative intervals alongside nameplate ratings. Do not substitute a single desktop observation for sustained production measurements.

Keep power and energy units separate: kW is power, while kWh measures energy over time. Identify whether a number describes IT load, facility input or cooling capacity. Use consistent measurement boundaries when calculating efficiency metrics.

Map each rack to its actual electrical and cooling dependencies. Check distribution paths, redundancy arrangements and any shared upstream bottleneck. Space available in one rack does not imply usable power or cooling at that location.

Forecast constraints

Project workload demand with explicit growth, deployment timing, utilization and concurrency assumptions. Translate workload into equipment using measured or justified performance data. Include management, storage and network overhead where relevant.

Capacity is limited by the tightest constraint, which may be rack space, power, heat removal, floor loading, port count or uplink bandwidth. Show headroom for each resource separately. Do not combine incompatible resources into one unexplained utilization percentage.

Model baseline, high-growth and delayed-expansion scenarios. Include equipment procurement and installation lead times so an apparently distant capacity limit can trigger an earlier decision.

Check resilience and operation

Test the failure scenario required by the service design, such as losing one feed, cooling component or network path. Assess the remaining usable capacity rather than simply adding the ratings of redundant components. Document any load shedding or degraded service assumption.

Review maintenance access, airflow, cabling and monitoring placement with the relevant engineers. Distinguish current measured conditions from a proposed layout's unverified performance. Electrical code, environmental limits and site-specific design criteria must be checked by qualified personnel.

Deliver a constraint worksheet, dependency map, growth forecast, failure-case analysis and prioritized expansion triggers. Attach units and sources to every input. Report uncertainty ranges where measurements or future demand are weak, and keep approval of purchases or physical changes separate from the analytical plan.

References

Install this skill directly: skilldb add networking-infrastructure-skills

Get CLI access →