AP Subject A 36 — Facility management: from power paths, UPS and N+1/2N to airflow, path diversity and physical access
AP Subject A supplement: utility-to-rack power paths, UPS runtime, N+1 vs 2N, cold/hot aisles, seismic isolation, surge and fire, MDF/IDF path diversity, physical access and 4 original questions.
AP Subject A · Installment 36 · An independent study resource for the 2026 exam and Syllabus Ver.7.2. It is not affiliated with IPA and does not guarantee a pass. Open the series hub
Based on the 2026 IPA Applied Information Technology Engineer Examination (AP), Subject A, Syllabus Ver.7.2; checked 2026-10-22. This is independent study material from Motitory: it is not affiliated with, reviewed or endorsed by IPA. It does not cover every item of Ver.7.2 or all of facility design and regulation, and does not guarantee a pass. All examples and questions here are Motitory originals, not IPA past questions. Real facility design, construction and inspection must follow applicable local law, manufacturer guidance and the judgement of qualified professionals.
Study index
- Terminology at a glance
- Scope and boundaries
- The power path from utility to rack
- Redundancy levels and Tier
- Cooling and airflow
- Building and disaster protection
- Communication facilities and path diversity
- Physical access management
- Operations and maintenance
- Quick sheet and traps
- Four original diagnostic questions
- Self-check and review schedule
- Official sources and revision history
This is AP36 (a supplement) of the fixed AP33–38 completion plan, not the last article of the series; every installment is listed in the series hub. No installment from AP1 to AP35 systematically covered facility management, cooling and cold/hot aisles, physical access to facilities, or the end-to-end power path, so this supplement fills that gap. The following were covered earlier and are only linked here:
- RAID, reliability and availability calculations, general reliability design: AP8
- Data backup and recovery (as distinct from facility backup): AP15
- User authentication, authorisation and logical access control: AP20
- SLA and incident, problem, change and release management: AP25
1. Terminology at a glance
| English | 日本語 (exam term) | One-line definition |
|---|---|---|
| facility management | ファシリティマネジメント | Lifecycle management that plans, builds, operates and improves buildings, equipment and their environment |
| facility and equipment management | 施設・設備管理 | Maintaining and inspecting equipment with cost, comfort and safety in mind |
| seismic isolation | 免震 | Devices between the building or floor and the ground reduce how much shaking is transmitted |
| earthquake resistance | 耐震 | The structure withstands shaking through its own strength |
| surge arrester, surge protection | 避雷器・サージ防護 | Diverts overvoltage from lightning and similar events to earth to protect equipment |
| fire suppression system | 消火設備 | Detects and extinguishes fire; the method is chosen by room use and law |
| uninterruptible power supply (UPS) | 無停電電源装置（UPS） | Supplies power from batteries without a break during outages and voltage dips |
| private / emergency generator | 自家発電設備（非常用発電機） | Generates power from fuel for long outages; needs time to start and transfer |
| HVAC | 空調設備 | Equipment that controls temperature, humidity and airflow |
| airflow management | 気流（エアフロー）管理 | Designing and keeping air paths so that cold supply and hot exhaust do not mix |
| cold aisle / hot aisle | コールドアイル／ホットアイル | The aisle facing rack fronts (intake) / the aisle facing rack rears (exhaust) |
| aisle containment | アイルコンテインメント（アイルキャッピング） | Closing a cold or hot aisle with a roof and doors to reduce mixing |
| MDF / IDF | MDF（主配線盤）／IDF（中間配線盤） | The frame where incoming carrier lines terminate / the intermediate frame serving a floor or zone |
| path diversity | 経路の冗長化（ルート分散） | Running lines or power over physically different routes |
| Tier | ティア（Tier） | A grading scheme for the redundancy and maintainability of data centre facilities |
| tailgating | 共連れ | Passing through without authenticating by following someone who did |
| anti-passback | アンチパスバック | Rejects authentication in an impossible sequence, such as exiting with no entry record |
| interlock (mantrap) | インターロック（サークルゲート等） | Links doors so that one opens only after the other has closed |
2. Scope and boundaries
The syllabus item ファシリティマネジメント asks for two things. One is the purpose and thinking behind the design, construction management and operation of computer and network facility infrastructure. The other is facility and equipment management for cost reduction, comfort and safety, with its concrete means (redundant power and lines, backup environments, access management for facilities, seismic isolation, surge protection, fire suppression, UPS, private generation, HVAC, MDF/IDF, Tier and so on).
- Facility management is not just repair. Fixing what breaks is one part; capacity planning, design, construction, inspection, renewal and disposal are managed across the whole lifecycle together with cost and risk.
- Facility availability and IT redundancy are different layers. Two clustered servers still stop together if both depend on the same UPS, the same cooling or the same incoming conduit. Server, RAID and cloud redundancy (AP8) does not replace facility redundancy, and a robust facility does not prevent software failures.
- Facility backup is not data backup either. A backup site or standby power provides a place and energy to keep processing; restoring data to a point in time is the backup and recovery of AP15.
3. The power path from utility to rack
A typical path looks like this (names and layout vary by site):
Utility feed → switchgear and transformers (受変電設備) → (automatic transfer switch) → UPS → distribution boards and PDUs → equipment power supplies (PSUs) in the rack
When the utility feed fails, the generator supplies the UPS input through the transfer switch.
| Device | Role | What it cannot do |
|---|---|---|
| UPS | Supplies power from batteries without a break during outages and dips; bridges the time until the generator starts and transfers, or until a clean shutdown | It is not long-duration backup by itself; it lasts only as long as its battery |
| Generator | Supplies power for as long as fuel lasts | Starting, stabilising and transferring take time, so a UPS is needed in between; without a fuel supply contract and regular start and load tests, operation in an emergency is not assured |
Single points of failure and common-cause failures: a device that exists only once on the path (a distribution board, a transfer switch, a cable tray) is a single point of failure. Even with two separate paths, sharing one room, one tray, one breaker or one work procedure means a single fire, flood or work mistake can stop both (a common-cause failure).
A dual path must be independent end to end. Power redundancy means something only when the A and B paths are actually separate from the feed all the way to the equipment PSUs. A dual-cord server (two PSUs) plugged into the same PDU, or into two PDUs fed from the same UPS, still has that PDU or UPS as a single point of failure.
Worked example 1 — UPS runtime
Assumptions (all set for this example; not official thresholds):
- Usable battery energy 12,000 Wh (= 12 kWh, decimal units)
- Inverter efficiency 0.95 (simplified: 95% of the battery’s DC energy becomes AC output)
- Load 20,000 W (= 20 kW, real power in W, constant)
Formula (linear simplification): runtime ≈ usable energy (Wh) × efficiency ÷ load (W)
- Deliverable energy: 12,000 × 0.95 = 11,400 Wh
- Runtime: 11,400 ÷ 20,000 = 0.57 h = 0.57 × 60 = 34.2 minutes
- Ageing margin: if only 80% of the usable energy is counted, 12,000 × 0.80 × 0.95 ÷ 20,000 = 9,120 ÷ 20,000 = 0.456 h = 27.36 minutes
- Load growth: if the load rises to 30,000 W, 11,400 ÷ 30,000 = 0.38 h = 22.8 minutes
Judgement: if this is comfortably longer than the time the generator needs to start and transfer (assumed to be a few minutes in this example), the UPS can bridge the gap. If the generator fails to start, however, the UPS time is only time for a clean shutdown. Real runtime depends on the manufacturer’s discharge curve (non-linear, with efficiency dropping at higher loads), battery age and temperature, conversion losses and design margin, so this calculation is only for checking the concept. When a question gives the load only in VA, check whether you must multiply by the power factor to get W.
Worked example 2 — finding a common-cause path
Setup (assumed): eight servers at 500 W each, 4,000 W in total, all with dual PSUs that share the load half and half in normal operation. PSU-A goes to PDU-A and PSU-B to PDU-B. Each PDU’s usable rating is assumed to be 5,000 W. An inspection finds that PDU-A and PDU-B are both fed from UPS-A. UPS-B serves other racks only.
- Single point of failure: UPS-A. If UPS-A stops, PDU-A and PDU-B go off together and all eight servers stop. Dual PSUs do not help at all.
- Corrective action: re-feed PDU-B from the UPS-B path (under a planned outage and work procedure), and confirm that cable routes and distribution boards are separate too.
- Capacity check: if one path stops, the remaining PDU carries the full 4,000 W alone. 4,000 W ≤ 5,000 W, so it holds under these assumptions. If the rack grows to 6,000 W, each path carries 3,000 W in normal operation and looks fine, but on one-side failure the survivor sees 6,000 W > 5,000 W and may trip on overload too. Keep each path’s normal load at a level the path can carry alone when the other fails.
4. Redundancy levels and Tier
Redundancy levels are stated against the required capacity N. Example: design load 300 kW, cooling units of 100 kW each, so N = 3 units.
| Notation | Units (example) | Meaning | Caution |
|---|---|---|---|
| N | 3 | Exactly the required capacity; no spare | One failure or service visit leaves a shortfall |
| N+1 | 4 | The required units plus one spare | Not two fully independent systems. A single shared part (piping, distribution board, controller) is a single point of failure |
| 2N | 6 (two paths of 3) | Two independent paths, each able to carry the full load | Remaining common causes (same room, same fuel tank, same operating mistake) can still stop both |
- N+1 holds at the stated design load. If the load grows to 350 kW, N becomes 4 units and a 4-unit system is no longer N+1.
- Maintenance bypass: switching the load to a bypass circuit that skips the UPS lets you service the UPS without stopping equipment. During bypass, though, the load is connected directly to the utility without UPS protection, and an outage at that moment stops it. Planned maintenance is a risky window in which redundancy is temporarily reduced.
- Tier: the data-facility Tier criteria listed by the IPA syllabus are examples of schemes that grade facility redundancy and maintainability. A Tier label alone does not establish a particular topology or uptime. Check the definitions and assessment scope of the standard being applied; actual availability also depends on operations and the IT design.
5. Cooling and airflow
Most rack-mounted equipment takes air in at the front and exhausts at the rear. Racks are therefore lined up with fronts facing each other to form a cold aisle (cold supply), and rears facing each other to form a hot aisle (exhaust return).
- The goal is to stop mixing. When exhaust loops back to the intake side (recirculation), a particular device’s intake temperature rises even if the room average looks fine. Cold/hot aisle layout, containment (closing an aisle with a roof and doors), blanking panels in empty U spaces and sealed floor cable cut-outs are all ways to reduce mixing.
- Cold/hot aisles organise airflow; they are not cooling redundancy. Protection against a failed cooling unit comes from capacity redundancy such as N+1 or 2N in section 4.
- Where you measure matters. What matters to equipment is the rack intake temperature. Watching only the room average or the cooling unit’s return temperature misses hot spots. Place sensors at top and bottom rack intakes and at likely hot spots.
- Humidity (too dry risks static, too humid risks condensation), leaks and dust are also monitored. Acceptable ranges come from equipment manufacturer specifications and the operator’s own criteria; this article does not give universal figures.
Worked example 3 — diagnosing hot spots
Assumptions: racks R1–R6 in one row. The intake warning value set by the operator for this example is T (not an official threshold). The room average is well below T. Measurements:
| Rack | Top intake | Bottom intake | Observation |
|---|---|---|---|
| R1–R3 | below T | below T | No issue |
| R4 | above T | below T | Six empty U at the top without blanking panels |
| R5 | above T | above T | One server mounted back to front, exhausting into the cold aisle |
| R6 | below T | below T | No issue |
Judgement:
- Even with a low room average, R4 and R5 are a problem because their intakes exceed the warning value. Monitoring only the average would miss this.
- For R4, suspect recirculation of rear exhaust through the empty U to the front. Installing blanking panels is the first action.
- For R5, the reversed server blows heat into the cold aisle. Correct its orientation as planned work. Lowering the set point spends more energy without fixing the cause.
- After the fix, measure again at the same positions to confirm the effect, and record the change.
6. Building and disaster protection
| Hazard | Typical measures | Exam point |
|---|---|---|
| Earthquake | Seismic isolation (免震): devices under the building or floor reduce transmitted shaking. Earthquake resistance (耐震): structural strength. Vibration control (制震): dampers absorb shaking | Even in an isolated building, anchoring racks and preventing tip-over is still needed. Conversely, anchoring racks does not reduce building shaking |
| Lightning and surges | Lightning protection (避雷針 etc.), surge arresters and surge protective devices (SPDs), earthing | Surges enter through communication lines as well as power lines |
| Water | Placement away from pipes and cooling drains, leak detection sensors, monitoring under the raised floor | Put sensors where water would collect; check pipe routes on the floor above too |
| Fire | Early smoke detection, fire suppression | See below |
| Crime | Access control (section 8), surveillance cameras | Access control for the plant rooms themselves |
Fire suppression is chosen for the situation. Server rooms often use gas-based systems (inert gas and others) to limit equipment damage, but there is no rule that water is always forbidden. Sprinklers (for example pre-action types) and water mist are also used; the choice is governed by room use, law and fire codes, manufacturer guidance and, above all, human safety. A gas system in a room people may occupy needs safety procedures under the applicable rules, such as pre-discharge alarms and evacuation, discharge delay, door interlocks and protection against accidental discharge during inspection.
7. Communication facilities and path diversity
- MDF (主配線盤): the frame where carrier lines first terminate on entering the building. IDF (中間配線盤): the intermediate frame that distributes from the MDF to floors and zones. The MDF is a concentration point for the whole building, so review flood and access risks at its location and whether it needs redundancy.
- The core of line redundancy is the physical path. Two lines on contracts with different carriers still fail together on a single excavation accident or exchange failure if they pass through the same incoming conduit, the same manhole, the same exchange (the carrier’s facility building) or the same MDF. Having two contracts alone does not assure physical diversity.
- What to check: building entry points (on different sides if possible), conduit and manhole routes, the serving exchanges, and in-building cable routes and frames. Ask the carriers for route information, and recheck periodically that construction work has not changed the routes.
8. Physical access management
Logical authentication and authorisation were covered in AP20. This section manages whether a person can physically touch the facility.
- Layered zones: site → building → floor → server room → rack (locked), stricter towards the inside. Electrical rooms, cooling plant rooms, MDF rooms and generator rooms matter as much as the server room.
- Least privilege and records: grant only the zones and periods each person needs for their work, keep entry and exit logs and review them regularly. Revoke immediately on leaving or transfer.
- Countermeasures against 共連れ (tailgating):
- Anti-passback: rejects authentication in an impossible sequence, such as exiting with no entry record or re-entering on a card already inside, exposing card lending and tailgating.
- Interlock (mantrap, circle gate): links the doors so the inner door opens only after the outer one closes. It makes people pass one at a time, but it does not by itself check who they are; identity checks come from combining it with cards, biometrics and so on.
- Visitors: prior request, ID check, visitor badge, escort by a staff member, logs.
- Moving media and equipment in and out: requests and approval, logs, and evidence of data erasure or shredding on disposal.
9. Operations and maintenance
| Activity | Content |
|---|---|
| Daily and periodic inspection | Check alarms, temperature and humidity, leaks, battery condition, generator fuel level |
| Capacity trend management | Track utilisation trends of power, cooling and rack space and plan expansion ahead; recheck whether N+1 still holds as load changes |
| Testing | Battery inspection and replacement plans, regular generator start and load tests, path switchover (failover) tests |
| Change records | Record wiring changes, added equipment and setting changes; keep drawings current |
| Alarms and emergency procedures | Alarm reception and contact chains, runbooks for outages, fire and flooding, regular drills |
| Planned maintenance | Risk assessment and approval before work, safe bypass and isolation procedures, verification after work that both paths and alarms are back to normal |
General change, incident and problem management follows AP25. Facility work is also subject to change management, and the work plan should state that the remaining path becomes a single point of failure during planned maintenance.
10. Quick sheet and traps
- Facility management = lifecycle management of facilities, not only fixing breakdowns.
- Server redundancy ≠ facility redundancy. Dependence on the same UPS, cooling or conduit means stopping together.
- UPS = no-break, short bridge and clean shutdown. Long duration = generator (plus fuel and testing). Generators take time to start and transfer.
- UPS runtime ≈ usable Wh × efficiency ÷ load W. Real values vary with discharge curves, ageing and temperature.
- Dual cords help only when both cords go to independent paths. The survivor must carry the full load when one side fails.
- N+1 = required N plus one spare (at the design load); not two independent systems. 2N = two full-load paths; common causes can remain.
- No UPS protection during bypass. Planned maintenance = temporarily higher risk.
- A Tier label ≠ a specific topology or an uptime promise.
- Cold aisle = rack fronts (intake), hot aisle = rears (exhaust). Airflow organisation, not cooling redundancy. Measure intake temperature per rack.
- Seismic isolation ≠ rack anchoring. Surges also enter through communication lines. Fire suppression is chosen by use, law and human safety.
- Two carriers sharing one conduit or exchange have no physical diversity.
- Anti-passback = rejects impossible authentication sequences. Interlock = links doors (not an identity check by itself).
11. Four original diagnostic questions
These questions were written by Motitory (unofficial); no IPA past question was used or adapted.
Q1 — UPS runtime
Assume a UPS battery with 6,000 Wh of usable energy, an inverter efficiency of 0.90 and a constant load of 9,000 W. Using the linear simplification (runtime ≈ usable energy × efficiency ÷ load), what is the runtime?
A. 40 minutes
B. 36 minutes
C. About 44.4 minutes
D. 60 minutes
- A: Incorrect. 6,000 ÷ 9,000 = 0.667 h = 40 minutes; the efficiency was not applied.
- B: Correct. 6,000 × 0.90 = 5,400 Wh, 5,400 ÷ 9,000 = 0.6 h = 0.6 × 60 = 36 minutes.
- C: Incorrect. 6,000 ÷ 0.90 ÷ 9,000 = 0.741 h ≈ 44.4 minutes; dividing by the efficiency turns the loss into a gain.
- D: Incorrect. 0.6 h was misread as 60 minutes; 0.6 h is 36 minutes.
Q2 — Cooling and airflow
A server room has a low average temperature, but the top intake temperature of some racks is high. Which action is most appropriate?
A. Because the room average is low, there is no need to check rack intake temperatures.
B. A cold/hot aisle layout is a cooling redundancy method that keeps cooling going when one cooling unit fails, so once the layout is changed no spare cooling unit is needed.
C. Align rack fronts to the cold aisle and rears to the hot aisle, and fit blanking panels in empty U spaces to reduce exhaust returning to the intake side.
D. Turn the fronts of racks with high intake temperatures towards the hot aisle so they reuse the exhaust as intake.
- A: Incorrect. What matters to equipment is the intake temperature; watching only the average misses hot spots.
- B: Incorrect. Cold/hot aisles organise airflow to reduce mixing of cold supply and exhaust. Protection against a cooling unit failure is separate capacity redundancy such as N+1.
- C: Correct. Aligning aisle orientation and fitting blanking panels are the basic measures against recirculation (worked example 3).
- D: Incorrect. The racks would draw in exhaust, raising intake temperatures further.
Q3 — Line path diversity
Company A has internet lines on contracts with two different carriers. An investigation shows both lines enter the building through the same conduit and terminate on the same MDF. Which statement is correct?
A. Because the carriers differ, the physical paths are automatically separate and no further check is needed.
B. Because they pass through the same conduit, a single excavation accident near it can cut both lines. Check whether the physical paths, including the entry routes and the carriers’ serving exchanges, are separate, and provide a different entry route if needed.
C. Duplicating the IDF on each floor prevents common-cause failure even with a single incoming conduit.
D. Doubling the bandwidth of each line gives the same effect as physical path diversity.
- A: Incorrect. Two contracts do not assure physical diversity.
- B: Correct. A shared conduit is a common-cause failure point; the exchanges and entry routes must be checked as well.
- C: Incorrect. IDF duplication concerns in-building distribution; with one incoming conduit, that conduit is still a single point of failure.
- D: Incorrect. Bandwidth is capacity; if the same physical path is cut, both lines go down regardless of bandwidth.
Q4 — Maintenance, access and safety
Which statement about operating data centre facilities is correct?
A. Anti-passback is a device that links two doors so that the inner door opens only after the outer door has closed.
B. While switched to maintenance bypass for UPS servicing, the load continues to be protected by the UPS batteries.
C. In a room with a gas-based fire suppression system, automatic discharge takes priority over evacuation procedures even during inspection.
D. While one path of a 2N system is stopped for planned maintenance, the remaining path is a single point of failure, so the procedure includes work approval and records, alarm monitoring, and verification after the work that both paths have returned to normal.
- A: Incorrect. Linking doors is an interlock (mantrap). Anti-passback rejects authentication whose entry and exit record sequence is impossible.
- B: Incorrect. During bypass the load is connected directly to the utility without passing through the UPS, so an outage at that moment is not protected.
- C: Incorrect. For gas systems, human safety procedures such as pre-discharge alarms, evacuation and protection against accidental discharge come first, under the applicable rules.
- D: Correct. Redundancy is temporarily reduced during planned maintenance, so the risk is managed and the return to normal is verified after the work.
Answers: B, C, B, D
12. Self-check and review schedule
- I can say why facility management is lifecycle management, not just repair.
- I can explain the difference between server redundancy and facility redundancy with an example of common-cause failure.
- I can draw the path from utility feed to rack and explain how the roles of the UPS and the generator differ.
- I can recompute the 34.2, 27.36 and 22.8 minutes of worked example 1 together with its assumptions.
- I can say why UPS-A is the single point of failure in worked example 2 and why 6,000 W causes an overload.
- I can explain N, N+1 and 2N with the 300 kW design load and 100 kW units, and state the risk during bypass.
- I can state the orientation and purpose of cold/hot aisles and why intake temperature is measured.
- I can distinguish seismic isolation from rack anchoring, anti-passback from interlocks, and two carriers from physical path diversity.
Suggested review rhythm (separate from the publishing schedule): same day, redo worked examples 1 and 2 without looking → after 1 day, Q1–Q4 → after 3 days, fill in a blank terminology table using only the quick sheet → after 7 days, change the battery energy, efficiency, load and PDU rating and recompute, then draw the utility-to-rack path on blank paper and mark the single points of failure.
13. Official sources and revision history
The exam scope rests on the IPA materials below. The numerical examples, configuration examples and judgement steps in this article are operational examples created by Motitory for study; they are not wording from the IPA syllabus or official thresholds.
- IPA exam outlines and syllabus overview — the IPA page where each exam category’s syllabus is published (last updated 2026-07-10); confirmed that Syllabus Ver.7.2 is listed as the AP learning guide. Checked 2026-10-22.
- IPA Applied Information Technology Engineer Examination Syllabus Ver.7.2 (PDF) — the syllabus text that defines the exam scope. This article maps to major category 6 / middle category 15 Service management / 5. ファシリティマネジメント, item (1) ファシリティマネジメント (printed p.92). Checked 2026-10-22.
- IPA syllabus revision information — the page for the syllabus revision history and changes. Checked 2026-10-22.
- IPA 2026 exam information — the page for administering the 2026 (令和8年度) exams; used for exam administration details, not for the exam scope. Checked 2026-10-22.
Revision history:
- 2026-10-22: First published. The official materials above were checked on the same day.
- 2026-10-28: Added the link to the next installment, AP37, and stated that AP38 is the final installment.
This article is the fourth installment of the AP33–38 completion plan and not the last one. The next installment is the AP37 technology mixed diagnostic, and the series ends with the final installment, AP38.