When you run a modern day data center, there is one metric that towers above almost everything else, power usage effectiveness, or PUE. It's simply the ratio of total facility energy use over the IT equipment's energy use. A PUE of 1.0 is perfect efficiency, i.e. all that power goes to servers, and nothing is wasted. Most facilities don't get anywhere near that, so when you see that the PUE is 1.4 or 1.5, you're doing well. Studies indicate an industry average of around 1.8 in the case of many data centers.
What does this mean in terms of practice? If your total energy use is 100 kW, and your IT equipment total energy use is 70 kW, your PUE is approximately 1.43. Every non-IT watt is an added overhead.
I remember going to a facility years ago and the PUE there was over 2.0, meaning as much energy went into cooling the computer facility down and infrastructure as compute energy. That was a wake-up call.
If you have a data center in Albuquerque, the stakes are specific and real.
First of all, power usage has a direct effect on your bottom line. Higher PUE implies higher overhead. If you're paying $0.10 per kWh in your facility and you're wasting 30% more power than you need, that's a cost that reoccurs.
Second, location matters. Albuquerque's climate is semi-arid and the temperature fluctuates widely. That leads to the fact that cooling loads might be different from humid zones. You'll have to find strategies that fit accordingly.
Third, being sustainable is no longer choice. Your facility's carbon footprint and data center energy strategy are important to your clients, regulators and your brand.
A Congressional Research Service report points out that data centers remain among the fastest-growing electricity consumers in the country, which makes PUE management vital for both cost and resilience.
So when I recommend that you "reduce energy consumption", "optimize energy usage" or "improve PUE" I am referring to it in this local, operational context. Making sure that your data center energy, power usage and infrastructure is optimized for the region is your real advantage.
Several factors influence the PUE for a modern data center. Let's break them down.
Every transformer, every UPS, every piece of switchgear is a count! If your uninterruptible power supplies and power distribution units aren't efficient, then you're throwing watts away.
When I had audited a facility, we found one transformer here that was operating at 95% efficiency. Upgrading it improved power conversion losses leading to the lowering of PUE.
You should monitor power supplies, measure losses and plan upgrades if the ROI pops up.
Cooling, this is often the largest non-IT loading in a data center. You ignore it, and your PUE suffers.
Free-air cooling and outside-air economization programs are popular, but in Albuquerque you need to consider dust, temperature waves and ASHRAE guidance.
A study reported that most modern data centers are around PUE values of 1.4-1.6 and much of this is due to inefficiencies in cooling.
If your racks are under-utilized then cool above minimum thermal conditions and blow more air than required, you'll increase your total energy use.
EPA’s ENERGY STAR program reports that inefficient cooling remains one of the largest contributors to wasted energy in data centers, often accounting for more than 30 percent of total facility use.
Let me share: I went to one location where the cold aisle containment was not properly closed. Fans ran more intensely, cooling overshot. Fixing the seals shaved off around 0.1 in PUE.
Your IT equipment has to function effectively. Idle servers continue to consume power. When your load is low, the non-IT overhead is low and PUE is high.
If you use up utilization, you retire old gear, you virtualise workloads, then your data center is more energy efficient and your PUE goes down.
It's easy to overlook this. You may only care about cooling or power distribution, but IT load is important.
Older buildings, old designs, poorly managed airflow or too much unused rack space and your PUE gets dragged up. Big data center operators indicate lower PUE in new generation big sites.
Make sure your layout allows for containment, wasted air flow is minimized, hot & cold aisles are segregated and airflow paths are monitored and any recirculation is minimized.
Add continuous measurement of data center operations, data center infrastructure metrics and energy usage to remain ahead.
You want practical, right? Here are strategies to do in Albuquerque that actually move the needle.
The aridity of Albuquerque's air affords you options. Consider outside-air economization, where the conditions permit. In regions where humidity permits, use adiabatic or evaporation cooling.
I once worked with a facility that switched into "free-cooling" during night-time hours and saved nearly 15% in total energy use during summer months.
When you optimize energy usage this way your data center energy footprint is improved.
Switch to high efficiency UPS systems. Consider direct-to-rack or high voltage distribution if the load you have supports it.
Check with your power conversion losses. Sometimes an upgrade of a component will save more than a total cooling retrofit.
You'll reduce power usage, overhead and efficiency of power distribution, thus reduce your PUE.
Consolidate servers, retire old machines, virtualise if possible. Increase the utilization of resources without impairing uptime.
If your servers are at 15% load but drawing 50% of peak power, you are wasting money and creating more energy usage for no purpose.
Monitoring your IT loads and matching loads with business demand and turning off hardware that is not needed helps.
Albuquerque enjoys good sun. On-site solar plus solar energy renewable energy contracts to reduce your energy cost and help your carbon footprint.
Pair this with demand response programs to mitigate utility bills in peak hours.
Using renewable energy ties into the story of reducing your total energy demand, optimizing energy and improving sustainability.
If you don't measure you can't improve. Use continuous metrics of PUE, power consumption, energy use, total energy in the facility and data center energy efficiency.
Set targets, for example, aim to improve PUE by 0.1 in 12 months. Review on a monthly basis, adjust on a quarterly basis.
When I questioned data center operators at a conference one common complaint was "we didn't know where the numbers were". Don't let that be you.
If you're ready to take action you'll probably get in the experts. It's important to find the right electrical contractor in Albuquerque.
Look for experience in data center power systems, knowledge of high efficiency UPS, power distribution units and energy monitoring systems.
Ask how they'll measure up to pre- and post-intervention, how they define success, how they plan to manage the transition without disrupting the data center operations.
Your facility may require upgrades for distribution, metering, monitoring and controlling systems. A competent contractor helps you in reducing power usage, reducing energy consumption and improving PUE in a measurable way.
Yes you can aim high but you will be up against blocks.
Due to age, there may be limitations to upgrade options.
Load variability can be a skewing issue, if IT load is falling, your PUE can add up if power consumption is flat.
Metrics matter, if you mis-calculate PUE or exclude parts of your power chain you'll mislead yourself. Some experts are warning that PUE is only part of the story in data center energy efficiency.
Don't ignore water usage and other resource impacts, data centers are now coming under the gun for much more than power.
Rack densities are increasing with the growth of AI and big data workloads. Some facilities are looking at 100kW racks or more. That drives new demands at the cooling and distribution level.
Liquid cooling immersion cooling modular designs are proving themselves.
Renewables and grid integration will be more important. As you develop your facility in Albuquerque you may already begin thinking about battery storage, onsite generation and active energy management.
The metric of PUE may evolve but for now it is a key indicator of data center energy use and data center energy efficiency.
A target around 1.3-1.4 is solid. Top performers may reach 1.1 or below. For instance, one large operator reported a fleet-wide PUE of close to 1.10.
Continuously if possible. At a minimum on a monthly basis to identify trends and anomalies.
Not always. If your IT load is low or your cooling system is over-sized, your power usage may be high. You need to match load, infrastructure and cooling.
No. You'll also need to look at things like water usage, carbon footprint, total energy use and other resource metrics.
You'll be helped with dry air while cooler nights help with outside air cooling. But summer peaks still require solid systems. Your energy use pattern may be different as compared with humid coastal zones.
Baseline: Measure current PUE, data center power, energy consumption, total energy and cooling load.
Target: Set realistic goal e.g. decrease PUE by 0.2 in 12 months.
Plan: Audit on power distribution, cooling system, server utilization and monitoring.
Execute: Upgrade key components (UPS, transformers), sealing of racks, optimising airflow.
Monitor: Monitor monthly, adjust strategies, report to stakeholders.
Scale: Implement strategies on scale of facility, renewables & more aggressive cooling as next phase.
Getting a lower PUE isn't just a numbers game. It means your data center will use less waste, cost less to run and send out a clear message about sustainability and performance.
For your facility in Albuquerque, it is a real opportunity. As you optimize energy usage, reduce energy consumption and improve data center energy efficiency you become a smarter operator, a better partner and a more competitive business.
Inspired by what you read?
Get more stories like this—plus exclusive guides and resident recommendations—delivered to your inbox. Subscribe to our exclusive newsletter
Resident may include affiliate links or sponsored content in our features. These partnerships support our publication and allow us to continue sharing stories and recommendations with our readers.