Data Center Design and Project Requirements
Start With Requirements, Not Equipment
Every defensible design begins with a requirements document that captures availability target, IT load forecast, growth horizon, redundancy philosophy, and budget. Candidates must know the four classic availability tiers (often mapped to Tier I–IV) and what each adds: non-redundant capacity, redundant components, concurrently maintainable paths, and fault-tolerant paths. Understand the difference between concurrently maintainable (any component can be serviced without shutting down IT) and fault tolerant (a single failure of any path does not disrupt IT).
Concepts to Master
- Capacity vs. redundancy: N, N+1, 2N, and 2(N+1) topologies and their cost/space implications.
- Design load versus actual load; right-sizing avoids stranded capacity.
- Project phases: programming, schematic design, design development, construction documents, commissioning.
- Standards landscape: recognize that TIA-942, Uptime Institute tiers, and BICSI guidance serve different purposes; questions often test which document governs what.
- Sustainability targets (PUE goals) must be set as requirements, not retrofitted later.
Common Traps
Questions love to conflate reliability (fewer failures) with availability (uptime despite failures). Another trap: assuming a higher tier always means higher efficiency - fault tolerance usually costs efficiency. Watch for distractors that treat a growth forecast as a fixed number rather than a range with assumptions.
Worked Example / Rule of Thumb
If IT load is 800 kW and the owner requires N+1 on UPS modules sized at 400 kW, you need three modules: two to carry the load plus one redundant. Rule of thumb: size modular infrastructure so the loss of the largest single module still leaves full IT load supported, and document the assumption in the basis of design.