EXIN CDCE cheat sheet

Tier I–IV (Uptime Institute) describe topology and concurrent maintainability, not measured availability; Tier IV requires fault tolerance plus continuous cooling and electrical paths.

Redundancy notation: N = capacity needed; N+1 = one extra unit; 2N = two full capacity systems; 2(N+1) = two full systems each with one spare unit.

Key design load metrics: IT load (kW/rack), power usage effectiveness PUE = total facility power / IT equipment power, and design PUE is a target, not a guarantee.

Cooling capacity rule of thumb: air-side economizers need temperature and humidity within ASHRAE allowable envelope; ASHRAE TC 9.9 recommends supply air 18–27 °C for Class A1–A4 equipment.

Raised-floor design: tile airflow is limited; perforated tile open area typically 25–60 %; hot-aisle/cold-aisle containment reduces bypass and recirculation.

Electrical single-line diagram must show utility feeds, transformers, UPS, STS, PDU, RPP and generator; point of common coupling and transfer time drive ride-through requirements.

Fire suppression: clean-agent (e.g. FM-200, Novec 1230) or inert gas for occupied spaces; water mist or pre-action sprinkler where code permits; VESDA early detection is common.

Structural floor loading: typical design 8–12 kPa (approx. 170–250 lb/ft²) for white space; verify against rack and UPS battery weights.

Cable pathway separation: power and data in separate trays; fibre and copper bend radius limits; redundant paths must be physically diverse to avoid common-mode failure.

MOP (Method of Procedure) is step-by-step for a specific task; SOP (Standard Operating Procedure) is routine recurring work; EOP (Emergency Operating Procedure) covers failure response.

Change management: every change needs risk assessment, approval, rollback plan and post-implementation review; emergency changes are retro-approved within a defined window.

Preventive maintenance intervals: UPS battery impedance/capacity test annually; generator load bank test at least annually; IR thermography on electrical joints periodically.

Incident vs problem: incident restores service; problem removes root cause. Major incident triggers post-incident review with timeline and corrective actions.

Capacity management tracks power, cooling, space, network ports and weight; thresholds (e.g. 80 % utilisation) trigger action before exhaustion.

Access control: two-person rule for critical switching, visitor escorts, badge audit trails, and CCTV retention per policy.

Mean time metrics: MTBF = total operating time / number of failures; MTTR = total repair time / number of repairs; availability = MTBF / (MTBF + MTTR).

Battery testing: IEEE 450/1188 cover vented and VRLA batteries; capacity test verifies actual runtime against design autonomy.

Runbooks and escalation matrices define on-call tiers, response times and vendor contacts; drills (failover, generator start) validate them.

ISO/IEC 22237 series covers data centre facilities and infrastructures (general concepts, building construction, power, cooling, ICT, security, management).

ISO/IEC 27001 is the ISMS standard; Annex A controls cover physical and environmental security; certification is by accredited body.

EN 50600 mirrors ISO/IEC 22237 and defines availability classes 1–4 and protection classes; it is the European counterpart.

TIA-942 specifies telecommunications infrastructure for data centres with Rated-1 to Rated-4; ANSI/BICSI 002 covers design and operational best practice.

Uptime Institute Tier certification is a commercial rating; it is not a national or international standard and does not equal ISO certification.

PCI DSS v4.0 requires physical access controls, CCTV, visitor logs and annual review; applies to cardholder data environments.

SOC 2 Type II reports cover a period and test operating effectiveness; Type I is a point-in-time design assessment.

GDPR and local data protection laws affect data residency and physical security of processing; breach notification timelines apply.

NFPA 75/76 and local fire codes govern fire protection for IT equipment and information technology spaces; compliance is mandatory, not optional.

PUE = total facility energy / IT energy; DCiE = 1 / PUE. PUE of 1.0 is theoretical; values below ~1.1 require substantial economizer use.

WUE (water usage effectiveness) = annual water use (litres) / IT energy (kWh); CUE = carbon usage effectiveness = total CO2 / IT energy.

Total cost of ownership includes capex (build), opex (power, cooling, staff, maintenance) and cost of downtime; availability targets drive cost non-linearly.

Service level agreements should define availability (e.g. 99.982 % = ~1.6 h/year downtime), response times, credits and exclusions.

Availability math: annual downtime minutes = (1 − availability) × 525,600; 99.9 % ≈ 8.76 h/year; 99.99 % ≈ 52.6 min/year.

Business continuity: RTO is maximum tolerable downtime; RPO is maximum tolerable data loss; both drive redundancy and replication design.

Lifecycle planning: refresh cycles for UPS batteries (~5 years), servers (~3–5 years), cooling (~10–15 years); depreciation affects refresh timing.

Sustainability strategy: renewable PPAs, heat reuse, refrigerant selection (lower GWP), and reporting under frameworks such as EU Energy Efficiency Directive for data centres.

Outsourcing models: colocation (space/power), wholesale (shell), managed services; contract terms cover power density, cross-connects, SLA credits and exit.