EPI Data Centre Associate common mistakes
EPI Data Centre Associate · Understand the traps and how to avoid them.
When a question describes a raised-floor data hall, the candidate treats the void under the floor as a general cable pathway and answers that fibre and copper should be routed there, ignoring that the design intent is for cooling air distribution and that cable obstructions reduce airflow and complicate maintenance.
The candidate has seen underfloor cabling in real sites and generalises from that experience, without distinguishing between legacy practice and current design guidance that segregates power, data, and air paths and keeps the plenum clear.
What to do instead
Before answering any raised-floor pathway question, state the primary function of the void in the scenario (air delivery, containment, or service routing) and then check whether the proposed cable route would obstruct it. Learn the standard segregation rules: power and data in separate trays, fibre protected from crush and bend-radius violation, and cable entry through sealed grommets to preserve containment.
The candidate calculates cooling capacity or power density using nameplate ratings for IT equipment rather than design or measured load, and then selects a cooling or UPS topology that is oversized or mismatched to the actual redundancy requirement.
Nameplate values are the most visible number on a device label, and candidates confuse rated maximum input with expected operating load. The exam scenario usually gives both, and the distractor answers are built from the nameplate figure.
What to do instead
When a scenario provides nameplate and design load, always compute with the design load and note the diversity or utilisation factor explicitly. Practise converting between kW, kVA, and tons of cooling, and check that the chosen topology (N, N+1, 2N) matches the stated availability target rather than the largest number in the options.
The candidate answers standards questions from memory of a single clause and applies a requirement across the whole facility, for example assuming a tier rating applies to every system in the building rather than to the specific topology being described.
Tier and availability classifications are frequently summarised in training material as a single label for a site, so candidates lose the nuance that ratings apply to defined topologies and that concurrent maintainability and fault tolerance are distinct properties.
What to do instead
For every standards question, identify which property is being tested: fault tolerance, concurrent maintainability, or basic capacity. Then map the described topology to that property before choosing an answer. Build a one-page comparison of the four topology levels covering redundant components, distribution paths, and maintainability, and revise it until you can state the differences without notes.
The candidate treats a standard as a legal requirement and selects an answer saying non-compliance with a voluntary framework is a regulatory breach, or conversely dismisses a mandated code as optional guidance.
Standards bodies, national codes, and customer specifications are often presented together in study material, and candidates do not separate voluntary industry frameworks from enforceable regulation and contractual obligation.
What to do instead
Classify each document you study into three buckets: law or regulation, voluntary standard, and contractual or customer requirement. When a question asks about obligation, check which bucket the cited document sits in. Practise writing one sentence per document stating who enforces it and what happens if it is not met.
On maintenance questions, the candidate chooses the activity that restores the failed component fastest rather than the one that follows the approved change and permit process, skipping isolation, lockout, or verification steps in the answer.
Exam scenarios reward operational urgency, and candidates project real-world pressure onto the question, forgetting that the correct answer in a governed environment is the controlled sequence, not the quickest fix.
What to do instead
For any maintenance scenario, run a fixed checklist: risk assessment, change approval, isolation and lockout, permit to work, execution, verification and functional test, and record update. Eliminate any option that skips a step even if it sounds efficient. Learn which tasks require a formal change request versus routine planned maintenance.
The candidate confuses preventive maintenance intervals with condition-based triggers and answers that a component should be replaced on a fixed calendar schedule when the scenario describes monitoring data indicating degradation.
Maintenance terminology overlaps, and candidates default to the familiar calendar-based model without noticing that the question supplies sensor readings, trend data, or performance drift that point to a condition-based response.
What to do instead
Read the scenario for the trigger word: schedule, elapsed time, or calendar points to preventive; trend, threshold, vibration, or temperature rise points to condition-based; failure or breakdown points to corrective. Match the response to the trigger, and remember that condition-based action still requires the change and permit process.
The candidate treats risk assessment as a one-off document produced at project stage and selects answers where the register is filed and not revisited, rather than recognising continuous review, ownership, and treatment tracking.
Governance material often presents risk registers as deliverables, so candidates remember the artefact and not the cycle. Questions that describe a new threat or a changed environment are testing whether the candidate knows the register must be updated.
What to do instead
Memorise the risk management cycle as identify, assess, treat, monitor, and review, and attach an owner and review frequency to each risk. In scenario questions, look for signals of change such as new equipment, new tenants, or new threats, and prefer answers that update the register and reassign treatment over answers that archive it.
The candidate answers access-control questions by choosing the strongest technical control, such as biometric entry everywhere, instead of the layered control set that matches the zone and the threat, and ignores procedural controls like escorting, logging, and visitor management.
Technical controls feel more definitive than procedural ones, and candidates equate security with hardware. Exam scenarios often describe a lower-risk zone where procedural controls are the proportionate and specified answer.
What to do instead
For each zone in the facility, define the required control layers: physical barrier, electronic access, surveillance, and procedural supervision. Then check the scenario's zone classification before choosing. Remember that tailgating, escort, and visitor logging are controls in their own right, and that proportionate response is usually the marked answer.
The candidate answers availability and redundancy questions by selecting the design with the most components, assuming more redundancy always means better availability, without accounting for single points of failure in the distribution path or the added complexity of more components.
Redundancy is taught as a count, so candidates add components rather than trace the full path from utility to load. They miss that a redundant UPS feeding a non-redundant switchboard still has a single point of failure.
What to do instead
Trace the power or cooling path end to end in every topology question and mark each point where a single failure interrupts the load. Choose the answer that removes single points of failure along the whole path, not the one with the largest component count. Practise drawing N, N+1, and 2N paths and labelling the failure points.
The candidate answers governance questions about roles by assigning accountability to the person or team performing the task, rather than to the accountable owner who approves and is answerable for the outcome.
In practice the engineer doing the work is visible and the accountable owner is not, so candidates map responsibility onto activity. Governance frameworks separate who does the work from who owns the outcome, and the exam tests that separation.
What to do instead
For each governance scenario, write two labels: who performs and who is accountable. Choose answers that place accountability with the named owner and execution with the operational team. Revise a simple responsibility matrix covering operations, security, change management, and risk ownership until the split is automatic.