- Inventory and measure: List every device’s rack units (1U = 1.75"), depth (chassis + rear connectors), and weight. Include PDUs, shelves, cable managers, KVM, patch panels, and blanking panels. - Height (U): Sum all U, add 20–50% growth headroom, plus 1U per 2–3 switches for horizontal cable managers. Common choices: 42U (standard), 45–48U (dense), 24–27U (small/edge). - Depth: Check device max depth incl. power cords and cable bend radius. Add 6–8" rear clearance for cabling/PDUs. Typical cabinet depths: 600 mm (~24") for shallow gear; 800–1000 mm (~31.5–39.4") for deeper switches/servers. Prefer adjustable rails to match mounting depth. - Width: Internal rail standard is 19" per EIA-310. External cabinet width 600 mm fits tight spaces but limits side cable channels; 800 mm provides spacious vertical cable management and side-to-side airflow options. Telco 23" only if gear requires it. - Rack type: Two-post (relay) for light, shallow gear; four-post for heavier/deeper equipment. Open-frame for maximum airflow/access; enclosed cabinets for security, airflow control, and noise with perforated doors. - Airflow: Match device airflow. Many switches are side-to-side—use 800 mm cabinets with side channels or choose front-to-back models. Maintain front-to-back airflow and blanking panels to prevent recirculation. - Power and cooling: Plan A/B vertical PDUs; ensure depth allows PDU clearance. Confirm BTU load with room cooling; use hot/cold aisle, brush panels, and sealing. - Access and space: Ensure 36" front and 24" rear service clearance (local codes may vary). Check doorways, ceiling height, overhead obstructions, and floor loading. - Other specs: Weight rating, seismic/anchoring if needed, grounding, top/bottom cable entry, casters/levelers, and cage-nut (square) rails for flexibility.

- High-CFM fan trays: Top-mounted or rear-mounted trays pull hot air out; front-mounted trays assist cold-air intake. Choose variable-speed, ball-bearing fans with PWM control and hot-swap capability. Use fan doors or rear-door fans to enhance exhaust without adding noise up front. - In-rack fans for side-to-side gear: Side-to-rear airflow kits or cross-flow fans for switches prevent hot recirculation in switch-heavy racks. - Blanking (filler) panels: Seal every unused U of rack space to stop hot air from looping to the front. Use tool-less panels for speed; solid panels for isolation, perforated only where specified by vendor. - Brush panels (cable pass-throughs): Route cables while sealing openings to maintain pressure and block bypass airflow. Use dense bristles; pair with grommeted floor tiles for underfloor environments. - Airflow containment/direction: Rack chimneys or duct kits connect rack exhaust to ceiling return. Use aisle containment (hot or cold) where possible. Install baffles and side panels to prevent lateral leakage between racks. - Door and panel choices: High-perforation (≥70%) front doors for intake; rear doors with high free area or heat-exchanger doors for dense racks. Keep side panels installed to prevent cross-rack mixing. - Cable management: Vertical managers and short patch leads reduce face blockage. Avoid cable masses in the intake path. - Sealing and grommets: Floor grommets, foam kits, and baying seals close gaps around racks and tiles. - Filtration and maintenance: Removable filters on intake; clean fans/filters regularly, monitor ΔT and fan RPM. Add temp probes at top-front and top-rear to tune fan curves.

- Plan routes: define primary paths (ladder/basket tray overhead, underfloor ducts), then drop to racks/cabinets. Separate power from data (minimum 6–12 inches; cross at 90°). - Capacity: observe fill ratios (≤40–50% continuous), allow 25–30% growth. Use proper tray/duct sizing. - Horizontal management (within rack front/back): use slotted finger ducts for high-density patching; use horizontal D-rings for light/medium density. Pair with patch panels to keep patch cords straight and support bend radius. - Vertical management (rack height): use vertical finger ducts or large D-ring managers to carry bundles alongside equipment; provide tie-off points every 12–18 inches. - Rings vs ducts: D-rings are simple, visible, low-cost; finger ducts provide containment, better bend-radius control, and a cleaner look for dense environments. - Cable support: use J-hooks or cable tray for pathway support; spacing ~4–5 ft (J-hooks) and more frequently for heavier bundles. Avoid resting cables on ceiling tiles. - Fasteners: use Velcro (hook-and-loop) for all data cabling; avoid zip ties except for permanent infrastructure, and never over-tighten. Place straps loosely and at regular intervals. - Bend radius: maintain ≥4x cable diameter for UTP, ≥10x for fiber when static; avoid kinks and sharp bends. Provide strain relief at terminations. - Slack and service loops: provide modest service loops (0.5–1 m) at racks; avoid storing excess in managers—park in slack trays or rear bays. - Labeling/color: label both ends; use color conventions per function/VLAN; document routes and updates. - Airflow and access: keep fronts of switches clear; route patch cords to sides; avoid blocking fans. - Compliance: use plenum-rated cable where required; respect fire stops; follow local codes/NEC and BICSI best practices.

- Goal: Dual power paths (A/B) to every device, each able to carry the full load (2N) or at least N+1 upstream. - PDU choice: - Basic: Use only for non-critical or downstream of intelligent upstream monitoring. No visibility/remote control. - Metered (inlet-level): Minimum for production. Enables capacity planning and overload avoidance; place one per A feed and one per B feed. - Metered-per-outlet or Switched: Preferred for critical racks. Per-outlet metering for accurate load balancing; switched adds remote power-cycle and outlet-level control. - Add an Automatic Transfer Switch (rack ATS) for any single-corded device to achieve A/B redundancy. - Power capacity: - Provide 3‑phase where possible (e.g., 208V 3ϕ in NA or 400V 3ϕ in EMEA) to maximize outlet density and balance phases. - Size PDUs/feeds so each side (A or B) can support 100% of the rack’s steady-state load at 80% breaker derate. Aim for 40–60% typical utilization per side. - Common robust choices: 208V 3ϕ 30A or 60A (NA) or 230/400V 16–32A (EMEA) with mixed C13/C19 outlets. - UPS: - Online double-conversion for critical; line-interactive only for edge/non-critical. - Redundancy: N+1 minimum; 2N for high availability (each UPS sized to support full rack load). - Size at 1.2–1.5× total IT load with required runtime (e.g., 5–15 min to ride to generator). - Place A and B PDUs on independent UPS/panels/generators; avoid common upstream single points. - Monitoring and safety: - SNMP-enabled PDUs with environmental sensors. - Phase balancing, per-outlet metering alarms, locking IEC cords. - Proper grounding, surge protection/TVSS, EPO compliance.

- Follow applicable codes/standards: NEC Article 250 (US), UL 508A (industrial control panels), NFPA 79 (machinery), IEC 60364/61439 (international), and the AHJ/local utility requirements. - Establish a single protective earth (PE) bar in the cabinet, listed for grounding, rigidly bonded to the enclosure. Connect PE to the facility’s grounding electrode system via an equipment grounding conductor sized per code and the upstream overcurrent device. - Maintain a single-point bonding scheme: all metallic parts (enclosure body, doors, subpanels, DIN rails, cable trays, gland plates, heatsinks) bond back to the PE bar with short, low-impedance paths. Avoid daisy-chaining. - Use listed lugs/studs, serrated/star washers or grounding clamps; remove paint/oxidation at contact points; apply antioxidant if required; torque to manufacturer specs. Use green or green/yellow conductors for PE. - Bond doors with flexible grounding braids/jumpers near hinges. Ensure continuity across removable panels via dedicated bonding jumpers. - Keep PE conductors as short/straight as practical, separated from signal wiring where possible. Route surge protective devices and cable shields to PE at the entry point with wide, low-inductance connections. - Separate neutral from PE; do not bond N–PE in the cabinet unless it is the service disconnect point per code. Keep DC negative and signal/reference grounds isolated from PE except at a defined single bonding point if required by the design/EMC plan. - Provide equipotential bonding to external metallic systems (conduits, trays, piping) entering the cabinet using bonding bushings/glands per listing. - Label the PE bar and all grounding points; use consistent color coding; document the grounding diagram in as-builts. - Test and verify: continuity and bonding impedance per the governing standard; insulation resistance; fault-loop impedance/overcurrent clearing; SPD earthing; record torque values. Remedy paint/loose connections found. - Engage a licensed electrician and coordinate with the AHJ for inspection and approval.

- Verify rack specs: depth, U-space, rail type (square, round, or threaded), weight capacity, and ventilation pattern. Measure device footprint and weight. - Choose support: - Fixed vented shelf: simplest, best for light–moderate weight, good airflow. Use for devices rarely serviced. - Heavy-duty fixed shelf: for dense/heavy gear; derate for cantilever loads. - Sliding shelf: only when frequent access is needed; respect weight ratings; lock-out preferred. - Vendor rail kit: best fit, preserves airflow and serviceability if available. - Universal adjustable rails: for odd sizes; check minimum/maximum depth. - Hardware: - Cage nuts for square holes; choose correct thread: M6 (common), 10-32, or 12-24. Use matching screws and washers. - Install cage nuts with the proper tool; orient spring tabs horizontally; keep fingers clear. - Use four fasteners per U when possible; torque snug, not stripped. - Device securing: front and rear support when possible. Add rear L-brackets or rear rails to prevent sag. Use Velcro straps, rack ears, or tie-downs. Anti-slip pads for small devices. - Placement: heaviest gear low; balance front/rear weight. Avoid stacking loose devices; use shelves with lips. - Airflow: maintain front-to-back flow; use vented shelves; add blanking panels to prevent recirculation. - Cable management: rear cable bars, Velcro (not zip ties), service loops, strain relief; avoid obstructing slides. - Power/ground: use PDUs with proper cord lengths; bond grounding per rack standard. - Safety: two-person lift for heavy items; test slide clearance; avoid sliding shelves for very heavy gear; secure rack to floor if required (seismic).

- Follow manufacturer bend-radius specs; if unknown, use ≥10× cable OD (static) and ≥15–20× during installation. Use radius limiters, rings, and spools in panels and pathways. - Choose management hardware: horizontal/vertical managers with radius control, slotted ducts, troughs, and D‑rings. Avoid tight zip ties; use Velcro with light tension. - Route fibers on panel sides with 90° managed turns; avoid door pinch points and hinge paths. Maintain separation from power and sharp edges with grommets. - Right-size patch cord lengths to eliminate excessive slack; store necessary slack on dedicated spools, not in random loops. - Use cassettes and breakout modules (e.g., MPO/MTP to LC) for high-density panels; select front-access designs for maintenance in tight racks. - Provide strain relief at panel entries and for trunk cables; secure to lashing points before fan-out. - Standardize polarity and mapping (e.g., Method B for MPO). Keep consistent connector types and keying across the plant. - Label both ends of every link with unique IDs, destination, and port. Color-code jackets by type (SM/OMx) and purpose. Maintain up-to-date as-builts in a CMDB with photos. - Control pathway fill: keep ducts/trays ≤50–60% fill for serviceability and heat dissipation. Reserve spare capacity and defined growth paths. - Keep connectors clean and capped; implement “inspect-clean-connect” with appropriate tools. Dust management in panels. - Plan changes: pre-stage cords, define routes, and execute during windows. Document and back out if needed. - Test after changes with visual fault locator, power meter, and OTDR for critical links. Verify loss budgets and polarity.