DCSW common mistakes
DCSW · Understand the traps and how to avoid them.
Treating Tier I–IV as a simple ladder and answering that a Tier III facility must contain every Tier IV feature, instead of evaluating the specific topology (concurrently maintainable vs. fault tolerant) against the scenario described.
Candidates memorise the four tier names from marketing material and assume higher tier means a superset of features, so they pick the 'highest tier' option regardless of what the question actually asks about redundancy or concurrent maintainability.
What to do instead
Learn each tier by its defining performance outcome: Tier I basic capacity, Tier II redundant capacity components, Tier III concurrently maintainable with dual-corded paths, Tier IV fault tolerant with continuous cooling. For every practice scenario, state aloud which single property is being tested (maintainability, fault tolerance, or component redundancy) before selecting an answer.
Assuming a dual-corded server is fully protected without checking that its two power paths are truly independent - separate UPS, separate distribution, separate upstream source - and answering as if any two cords satisfy redundancy.
Diagrams in study material show two cords and candidates stop reading at that point; the subtle failure (both cords landing on the same PDU or the same UPS module) is exactly what the exam tests.
What to do instead
For any redundancy question, trace each path end to end: utility feed, transformer, UPS, PDU, rack A/B. If the two paths share any single point of failure, the system is not redundant. Practise drawing two-column A/B path diagrams from memory until shared-component traps are obvious.
Confusing N, N+1, 2N and 2(N+1) when sizing UPS or generator capacity, and answering with the configuration that sounds most resilient rather than the one that meets the stated load and redundancy requirement.
The notation looks similar and candidates mix up 'redundant modules' with 'redundant systems'; they also forget that 2N means two complete independent systems, not two modules in one system.
What to do instead
Build a one-page reference: N = load only; N+1 = one extra component; 2N = two full independent capacities; 2(N+1) = two full capacities each with one spare. Work numerical examples where you calculate installed capacity for a given IT load under each configuration, and check the arithmetic against the stated redundancy goal.
Mixing up sensible-heat ratio, wet-bulb, dry-bulb and dew point when evaluating cooling coil performance or economiser eligibility, and choosing an answer based on dry-bulb alone.
Candidates treat temperature as a single number and skip psychrometrics because it feels like physics rather than data centre content; the exam deliberately uses humidity-dependent scenarios.
What to do instead
Learn to read a psychrometric chart: locate a state point from dry-bulb plus relative humidity, then read wet-bulb, dew point and enthalpy. Practise deciding whether an air-side economiser is usable given a stated outdoor condition, and remember that coil latent capacity depends on dew point, not dry-bulb.
Answering cooling questions with a single metric (e.g., only supply air temperature) instead of checking airflow, bypass, containment and return path, and therefore missing the actual root cause in a hot-spot scenario.
Hot-aisle/cold-aisle and containment concepts are learned as vocabulary, so candidates pick the term they recognise rather than diagnosing the described symptom (recirculation vs. bypass vs. insufficient capacity).
What to do instead
Use a symptom-to-cause table: hot spot at top of rack = recirculation or missing blanking panels; cold supply but hot exhaust return = bypass or short-circuiting; uniform overheating = capacity shortfall. For each practice scenario, name the physical air path before choosing the remedy.
Treating structured cabling distances and media limits as approximate, and selecting a copper or fibre option that exceeds the standard's channel length for the required data rate.
Candidates remember '100 metres for Ethernet' without distinguishing permanent link from channel, or which categories and fibre types support 10G, 25G or 40G at what distance.
What to do instead
Memorise a compact table: twisted-pair channel limits by category and speed, multimode fibre reach by OM class and data rate, and single-mode as the long-reach default. When a question gives a distance and a speed, check the media against that table before answering.
Applying clean-agent concentration and hold-time rules without checking the enclosure integrity, room volume basis, or whether the design accounts for occupied egress time and pressure relief.
Candidates learn a single design concentration figure and ignore the surrounding requirements (altitude correction, enclosure leakage, discharge within ten seconds for some agents, signage and egress), so they accept an option that is only partly correct.
What to do instead
Study each suppression system as a package: agent, design concentration, discharge time, hold time, enclosure integrity test, pressure relief venting, and egress/notification. For scenario questions, verify every one of those elements is satisfied before selecting the answer.
Confusing detection technology selection - spot vs. aspirating, photoelectric vs. ionisation, linear heat detection - and choosing the most sensitive-sounding option rather than the one matched to the hazard and environment.
Very-early-warning detection is marketed heavily, so candidates assume aspirating detection is always best, forgetting airflow, dust, and cost factors that make it unsuitable in some spaces.
What to do instead
Map hazard to detector: smouldering cable in a return plenum favours aspirating or beam; high-airflow spaces need detectors rated for the velocity; dusty or outdoor areas need alternatives to spot photoelectric. Justify each selection by the stated environment, not by sensitivity alone.
Answering access-control questions by choosing the strongest-sounding credential (biometrics, multi-factor) without checking the required egress path, fail-safe vs. fail-secure behaviour, and life-safety override.
Security and life safety objectives pull in opposite directions, and candidates prioritise the security framing of the question while forgetting that egress must never depend on a credential or a powered lock.
What to do instead
For every access-control scenario, first establish the egress requirement (free egress, request-to-exit, break-glass, fail-safe on power loss for egress doors), then layer the credential. Learn which doors must be fail-safe and which may be fail-secure, and check that the chosen option preserves egress before it improves security.
Skipping the commissioning sequence and answering maintenance or testing questions with the final integrated test rather than the correct stage (factory, site, functional performance, integrated systems).
Candidates treat commissioning as one event and do not know the level structure, so they misplace load-bank testing, point-to-point verification, or integrated systems testing in the sequence.
What to do instead
Memorise the commissioning levels in order and what each proves: component verification, system functional performance, and integrated systems test under load. For each scenario, ask which level is being described and what evidence it produces before selecting an answer.