DCEP flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

PUE formula

Power Usage Effectiveness equals total facility energy divided by IT equipment energy over the same period. A value of 1.0 means all energy reaches IT; higher values indicate overhead from cooling, power distribution, and other loads.

DCiE definition

Data Center infrastructure Efficiency is the reciprocal of PUE expressed as a percentage (1/PUE × 100). A PUE of 1.5 corresponds to DCiE of about 66.7 percent.

PUE measurement boundary

The defined set of meters whose energy is counted. Total facility energy is measured at the utility service; IT energy at UPS output or rack PDUs. Both must cover identical time windows.

Meter accuracy for PUE

Utility-side meters should be revenue-grade (roughly 0.5 percent class); IT-side branch or rack meters should be 1 percent or better. Poor meter accuracy dominates PUE uncertainty at low overhead ratios.

IPMVP M&V options

Measurement and Verification protocol options: A (retrofit isolation, key parameters), B (retrofit isolation, all parameters), C (whole facility utility meter), D (calibrated simulation). Savings claims should cite the option used.

Baseline normalization

Adjusting a baseline for changes in IT load, weather (cooling degree days), and occupancy before comparing periods. Without normalization, seasonal PUE changes are confounded with efficiency changes.

Server power proportionality

Server power approximates idle power plus utilization times the difference between max and idle power. Because idle power is often 30 to 50 percent of max, low utilization wastes substantial energy.

Consolidation headroom rule

Consolidate workloads onto fewer servers while keeping peak utilization below roughly 70 to 80 percent. This preserves latency headroom and avoids queuing while eliminating idle-server overhead.

Nameplate vs measured power

Nameplate ratings are conservative maxima that real servers rarely reach. Planning should use measured draw or vendor power curves such as SPECpower data instead of nameplate values.

Power capping

Enforcing a hard power ceiling on a server via BMC or RAPL controls, trading performance for reduced peak draw. Useful for matching rack power budgets and limiting demand charges.

BIOS power management

CPU P-states and C-states reduce voltage, frequency, and idle power. Disabling these features in BIOS for perceived latency gains forfeits large idle-energy savings across a fleet.

Idle server fraction

The share of servers running below a low utilization threshold, such as 20 percent. A high idle fraction signals the largest consolidation and decommissioning opportunity in a fleet.

Hot-aisle containment

Enclosing the hot aisle to prevent hot exhaust from mixing with cold supply air. Containment reduces recirculation and bypass, allowing higher supply temperatures and more economizer hours.

ASHRAE recommended envelope

The TC 9.9 recommended inlet range for Class A1/A2 equipment, roughly 18 to 27 degrees Celsius. Operating within it permits wider chilled water reset and economizer use.

Chiller kW per ton

Chiller efficiency expressed as electrical input power divided by tons of cooling delivered. Lower is better; centrifugal chillers commonly run around 0.5 to 0.7 kW per ton at design.

Cooling tower approach

The difference between leaving condenser water temperature and ambient wet-bulb temperature. A smaller approach enables more water-side economizer hours and lower condenser water temperatures.

Airflow per kW rule

Typical raised-floor designs supply roughly 100 to 160 cubic feet per minute per kW of IT load. Actual airflow should be verified by measurement or CFD rather than assumed.

Return-supply temperature delta

The difference between return and supply air temperatures. A delta well below design, such as under 8 to 10 degrees Celsius, indicates bypass air or low load rather than excess cooling.

Chilled water supply reset

Raising chilled water supply temperature, for example from 7 to 12 degrees Celsius, improves chiller and economizer efficiency but must respect coil capacity and dew-point limits.

Fan affinity laws

Fan power scales approximately with the cube of rotational speed. Variable frequency drives running fans at part speed therefore deliver large reductions in fan energy.

Distribution chain efficiency

Overall electrical efficiency is the product of stage efficiencies from transformer through UPS, PDU, and rack PDU. Each stage's losses compound and reduce energy reaching IT.

UPS efficiency curve

Double-conversion UPS efficiency peaks near full load, commonly 94 to 96 percent, and falls sharply below about 30 to 40 percent load. Eco-mode bypass can reach 98 to 99 percent.

Transformer loss components

No-load core losses occur continuously regardless of load, while copper losses scale with load squared. Lightly loaded transformers waste core loss around the clock.

Three-phase power formula

Real power in kW equals the square root of three times line-to-line voltage times current times power factor, divided by 1000. Use line-to-neutral voltage with the three-times form instead.

Power factor correction

Raising power factor reduces reactive current and resistive line losses, frees distribution capacity, and avoids utility low-power-factor penalties. Capacitor banks or active filters are common remedies.

Modular UPS sizing

Scalable UPS frames let modules run near their efficient load band as IT load grows. Oversized monolithic UPS units idle at low load and operate below peak efficiency.

Redundancy and efficiency tradeoff

A 2N topology splits load across two systems, so each runs at roughly half load and lower efficiency than an N+1 arrangement at the same IT load. Redundancy level directly affects PUE.

Higher distribution voltage

Delivering power at 415 or 480 volts instead of 208 volts reduces current for the same power, cutting I-squared-R losses in conductors and busway.

Partial PUE

pPUE applies the total-to-IT energy ratio to a defined subsystem such as one hall or module. It is useful without full-facility metering but is not comparable to facility PUE.

CUE and WUE metrics

Carbon Usage Effectiveness is total CO2 emissions divided by IT energy; Water Usage Effectiveness is liters of water divided by IT kWh. Both are reported alongside PUE for sustainability reporting.

Energy Reuse Effectiveness

ERE equals total energy minus reused energy, divided by IT energy. Values below 1.0 indicate that heat reuse is offsetting facility energy consumption.

Time-weighted PUE reporting

PUE for a period must be computed as total energy divided by IT energy over that period, not as an average of instantaneous PUE samples, which biases results when loads vary.