BICSI DCDC cheat sheet
BICSI DCDC · Keep the key ideas within reach.
Data Center Design and Project Requirements
Availability tiers: Tier I single path no redundancy; Tier II redundant components; Tier III concurrently maintainable; Tier IV fault tolerant with compartmentalization.
Concurrently maintainable means any planned maintenance activity can occur without shutting down IT load; fault tolerant adds tolerance of an unplanned failure event.
Design documents progress through program/concept, schematic design, design development, construction documents, then commissioning and handover.
Redundancy notation: N is required capacity; N+1 adds one unit; 2N duplicates the full system; 2(N+1) duplicates each system plus one spare.
Design loads must include IT critical load plus mechanical, lighting, and losses; add 10-25 percent growth margin per owner program.
Concurrent maintainability requires both power and cooling paths, not just electrical redundancy.
Mean time between failures (MTBF) and mean time to repair (MTTR) drive availability; availability = MTBF / (MTBF + MTTR).
A basis of design records assumptions, codes, load densities, redundancy, and growth so later changes can be traced to requirements.
Data Center Site and Space Planning
Floor loading for high-density halls commonly targets 150-300 lb/ft2 (7-15 kPa); verify against actual rack and UPS weights.
Slab-to-slab height must accommodate raised floor depth, rack height, cable trays, and clearance for sprinklers and ducts.
Separate hot and cold aisles; orient racks face-to-face on cold aisles and back-to-back on hot aisles.
Electrical rooms, UPS/battery rooms, and generator yards are placed to shorten feeders and isolate fire risk from the IT hall.
Site selection weighs power availability, fiber routes, flood and seismic risk, fuel delivery, and latency to users.
Growth planning reserves white space and power capacity for at least one expansion phase; track kW per rack trends.
Raised floor heights of 24-36 in (600-900 mm) support underfloor cabling and air delivery; deeper plenums serve higher densities.
Physical security zoning places the perimeter, building entry, data hall, and cage levels in progressively restricted layers.
Data Center Power and Electrical Systems
Critical power path: utility service, medium-voltage switchgear, transformers, UPS, power distribution units, remote power panels, rack PDUs.
UPS types include static double-conversion, delta conversion, and rotary; double-conversion isolates load from all utility disturbances.
Battery autonomy is sized in minutes at full load; valve-regulated lead-acid (VRLA) and lithium-ion are common chemistries.
Generators are sized for full critical load plus mechanical loads; fuel storage typically supports 24-72 hours of runtime.
Power usage effectiveness (PUE) = total facility power / IT equipment power; lower is better and 1.0 is theoretical minimum.
Rack power density planning: a 42U rack at 5 kW needs about 119 W per U; verify PDU and branch circuit ratings before deployment.
Single-corded IT equipment requires rack-level ATS or dual-path PDU design to preserve redundancy.
Grounding and bonding follow IEEE 3003 and local code; a single-point ground reference limits noise and fault currents.
Data Center Cooling and Mechanical Systems
ASHRAE recommended inlet envelope: 18-27 C (64-81 F) and dew point 5.5-15 C; allowable range extends to 15-32 C (59-90 F).
Heat load: sensible heat from IT plus lighting and people; 1 kW of IT load equals about 3,412 Btu/h.
Airflow rule of thumb: about 100-160 cfm per kW of IT load; verify with rack-level airflow and containment design.
Cooling capacity: 1 ton of refrigeration = 12,000 Btu/h = 3.517 kW; size CRAH units to peak load plus redundancy.
Containment of hot or cold aisles prevents supply/return mixing and raises return air temperature for economizer hours.
Free cooling economizers use outside air or water-side economizers when outdoor wet-bulb is below the design threshold.
Chilled water supply/return commonly 12-18 C (55-65 F) supply; higher supply temperatures improve chiller efficiency.
Humidity control prevents ESD below about 30 percent RH and condensation above the dew point; use dew point, not RH alone.
Data Center Cabling and Telecommunications Infrastructure
Structured cabling hierarchy: entrance facility, equipment room, backbone cabling, horizontal cabling, and distribution area.
TIA-942 defines data center cabling topology, redundancy, and area types; ANSI/TIA-568 covers generic cabling performance.
Fiber types: OM3/OM4 multimode for short reach, OS2 single-mode for long haul; 40/100G uses MPO parallel optics.
Copper categories: Cat 6 to 55 m for 10GBASE-T, Cat 6A to 100 m; Cat 8 supports 25/40GBASE-T to 30 m.
Cross-connect architecture uses main distribution area (MDA), horizontal distribution area (HDA), and zone distribution area (ZDA).
Cable pathways must maintain separation from power: typically 12 in (300 mm) for unshielded, less for shielded or fiber.
Labeling follows a documented scheme tied to the cabling administration system; every port and cable end is identified.
Cabinet airflow and cabling: overhead or underfloor trays should not block cold aisle supply; use vertical cable managers.
Data Center Operations, Security, and Commissioning
Commissioning levels: Level 1 factory, Level 2 site, Level 3 start-up, Level 4 functional performance, Level 5 integrated systems test.
Integrated systems test (IST) verifies full failure scenarios: utility loss, UPS transfer, generator start, and cooling recovery.
Change management requires documented request, risk review, approval, scheduled window, and rollback plan.
Security layers: perimeter, building access control, data hall access, cabinet locks, and video surveillance with retention policy.
Environmental monitoring covers temperature, humidity, water detection, power quality, and access events with alert escalation.
Preventive maintenance schedules include battery testing, generator load banks, filter changes, and infrared electrical scans.
Incident response includes escalation matrix, communication plan, and post-incident review with corrective actions.
Capacity management tracks power, cooling, space, and network utilization against thresholds to trigger expansion.