CDCP cheat sheet

Availability tiers: Tier I and II are single-path with no concurrently maintainable capacity; Tier III adds concurrently maintainable paths (one distribution path active, one alternative); Tier IV adds fault tolerance with multiple independent, simultaneously active paths.

Tier objectives are measured as annual downtime: Tier I ~28.8 h, Tier II ~22.0 h, Tier III ~1.6 h, Tier IV ~0.4 h (99.671%, 99.749%, 99.982%, 99.995%).

Redundancy notation: N = capacity needed for load; N+1 = one spare module; 2N = two full capacity sets; 2(N+1) = two full sets each with a spare.

Concurrently maintainable means any component can be removed/replaced for planned work without affecting the IT load; fault tolerant means a single failure of any component does not affect the load.

Common standards to cite: ANSI/TIA-942 (telecommunications infrastructure), ISO/IEC 22237 (facility and infrastructure), EN 50600 (data centre facilities and infrastructures), Uptime Institute Tier Standard.

Structural floor loading is typically designed around 8-12 kPa (roughly 1,200-1,500 kg/m2) for high-density white space; verify against project specification rather than assuming.

Separation of pathways: power and data cabling in distinct pathways; fibre and copper segregated where required; hot and cold aisles arranged to prevent recirculation.

Design documents typically include the basis of design (loads, densities, redundancy, growth), single-line diagrams, cooling schematics, and a commissioning plan aligned to a tier objective.

Power formula: P (W) = V x I x power factor for AC single phase; for three phase P = √3 x V(line-line) x I x PF.

UPS output sizing uses real power (kW) and apparent power (kVA); kVA = kW / power factor, so a 0.9 PF load of 900 kW needs 1,000 kVA.

Battery autonomy: runtime depends on load fraction and battery end-voltage; doubling load roughly halves runtime, and runtime falls faster than linearly at high discharge rates.

Typical distribution chain: utility MV feed -> transformer -> LV switchgear -> UPS -> PDU/static transfer switch -> rack PDU; A and B feeds should be independent to the source.

Generator sets are rated for standby/prime duty; continuous rated output is lower than standby rating, and load steps (block loading) must be checked against the alternator capability.

Power usage effectiveness: PUE = total facility energy / IT equipment energy; ideal is 1.0, and a PUE of 1.5 means 0.5 units of overhead per unit of IT energy.

Neutral and earthing: TN-S arrangements keep protective earth separate from neutral in the distribution; UPS output neutral bonding must follow the manufacturer's and local code requirements.

Redundancy decisions: N+1 UPS modules share a parallel bus; 2N uses two independent UPS systems feeding A/B rack PDUs so a single UPS failure cannot drop the load.

Sensible heat load: Q (kW) ≈ airflow (m3/s) x 1.2 (kg/m3) x 1.005 kJ/(kg·K) x ΔT (K); equivalently CFM x ΔT(°F) x 1.08 gives Btu/h.

Airflow rule of thumb: about 1.7-2.0 m3/min per kW of IT load at a 10-12 K supply-to-return temperature difference; verify with the actual design ΔT.

ASHRAE recommended envelope for IT equipment intake is roughly 18-27 °C with humidity bounded to avoid ESD and condensation; allowable envelope is wider but reduces reliability margin.

Containment: hot-aisle or cold-aisle containment prevents bypass and recirculation; uncapped racks and open gaps defeat the containment benefit.

Economisers: air-side or water-side free cooling is usable when outdoor conditions allow; the number of free-cooling hours depends on the local climate and the supply temperature setpoint.

Chiller plant efficiency is expressed as kW/ton or COP; 1 refrigeration ton = 3.517 kW of cooling, and COP = cooling kW / input kW.

CRAH/CRAC units control sensible cooling; humidity control is often centralised to avoid simultaneous humidification and dehumidification by competing units.

Liquid cooling: direct-to-chip and rear-door heat exchangers support high-density racks (tens of kW per rack) where air cooling becomes airflow-limited.

Change management: every change is logged, risk-assessed, approved, scheduled, implemented, verified and closed; emergency changes are retrospectively reviewed.

MOP (method of procedure) documents step-by-step execution with prerequisites, rollback and verification; SOP covers routine operations; EOP covers emergencies.

Preventive maintenance intervals follow manufacturer guidance; UPS batteries are typically capacity-tested on a defined schedule, and results are trended, not just pass/fail.

Incident versus problem: an incident is an unplanned interruption or degradation; a problem is the underlying cause requiring root-cause analysis.

Capacity management tracks power, cooling, space and network against thresholds (e.g. rack power, PDU capacity, cooling kW) with alerting before limits are reached.

Commissioning levels: factory testing, site acceptance, functional performance testing (Level 4), and integrated systems testing (Level 5) including full-load and failure scenarios.

Access control and CCTV logs are retained per policy; visitor escort rules and mantrap procedures protect the white space.

Metrics to report: PUE, availability (uptime / total time), MTTR, MTBF, and number of change-related incidents.