PUE Impact On Data Center Costs

Clarissa GarciaArticles, Blog

PUE impact to data center cost

Importance of PUE on Data Center Costs

Efficiency measurements such as Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE) assist data center owners and operators. By assessing the entire operations it is easier to identify the areas to improve. This metric is not only beneficial for identifying a data center’s environmental impact and carbon footprint, but also from a business operation perspective, it can help identify cost-saving measures.

PUE is a scale of one to three, with one being the highest possible score. If the PUE was 1, it meant that the data center’s entire power supply was used to power only the IT equipment. PUE has fallen from 1.8 in 2011 to 1.53 in 2018, according to a recent survey done by the Uptime Institute of 900 data centers and service providers. While this is encouraging, the study also revealed that outages and the severity of outages have grown in the past year. Consequently, the data center sector still has work to do in terms of improving PUE and outage rates.

efficiency formula power

How to calculate PUE?

Datacenter operation budgets can be destroyed by high power costs. In a colocation facility, it is also critical to know how a data center bills clients for their power consumption. Most data center providers bill on a per-customer basis. Customer equipment power usage and shared overhead costs for the entire data center are included in the charges. Operating numerous utility feeds, multiple generators, and multiple UPS systems are examples of shared overhead.

The below diagram illustrates the percentage of data centers operating at specific PUE. As per the uptime institute, the majority of data centers (47%) operate a PUE of 1.2-1.5. 38% are in the 1.5-2.0 range. This 38% of data centers could likely do some simple improvements to their data center, such as aisle containment and turning up the thermostat. This would result in lower PUE and cost savings. You can use our free online PUE calculator to see how much you can save by improving your energy efficiency.

pue ratings pue calculations

Photo Credit: uptimeinstitute.com

The lower the PUE, the more efficient a data center is. And this also means reduced operating costs for both the operator and the client. A good data center makes the best use of its power to maintain and manage its facility. Which will, as a result, also benefit the customers.
Inefficient facilities, data centers that run with more power, are likely to have higher running expenses. As a result, these providers must charge their consumers a fortune.

How Can Data Centers Lower Their PUE?

Datacenter operators can use a variety of techniques to reduce their PUE rating. By simply increasing the temperature within the facility. Some data centers keep temperatures substantially lower than those suggested by ASHRAE. But raising the temperature a little higher is still safe. This will not harm the equipment and may even assist to improve PUE.

Additionally, hot or cold aisle containment is advantageous. This strategy prevents the cold air from mixing with the hot air exhausted from IT equipment. This makes cooling more efficient, thus saving energy spent on cooling, which improves your PUE. Another advantage is that you can increase server density for the same volume of cooling, which can further improve PUE numbers by increasing IT load for the same or less cooling load.

factors of pue in data center

Photo Credit: gstatic.com

Data centers can also be resourceful with free cooling. Reducing reliance on energy-hungry cooling systems. The climate of the location must be considered. Of course in colder regions such as Iceland, Norway, Finland, and Alaska, where there is good infrastructure accompanying cold weather, and good air quality, free cooling is a viable option.
It is also suggested that the facility’s electricity distribution be examined. Power loss is reduced by using high-efficiency UPS systems and removing superfluous voltage converter stages. This has an added benefit because this power loss converts to heat therefore it decreases the cooling load, resulting in even more savings.

Drawbacks in Measuring The PUE?

PUE is an excellent tool for the data center’s facilities department. It enables facility engineers to assess the impact of infrastructure modifications. This includes:

  • Boosting the temperature
  • Upgrading to a more efficient UPS
Datacenter administrators are being forced to lower their PUE. The pressure to save expenses and match the stated PUE from other firms may have unintended consequences. The data center may do it without any strategic planning.
Let’s look at an example of how this may happen. Assume we have a data center with 100 kW input power, of which 50 kW is used to power IT equipment. This would result in an initial PUE value of 2.0, as shown earlier.
The data center administrator may decide to virtualize a few servers. And we thus cut the power used by IT equipment by 25 kW while also reducing the power used by our data center by the same amount. What will the PUE fate be?
Isn’t it true that having a higher PUE number is what we want to avoid? Certainly not. It’s critical to comprehend what might cause the PUE to rise or fall. It may be paradoxical, any decrease in IT demand without a corresponding decrease in cooling burden will result in a greater PUE.

When we break down PUE into its constituents, it becomes clearer:

When the IT load is lowered, the PUE will always rise, resulting in a higher PUE. Increasing the IT load, on the other hand, will always lower the PUE.

So, if our PUE has gone up, does this mean the data center is now less energy efficient? No, the data center is now more energy efficient. We are accomplishing the same work while using less energy at less cost.

To illustrate this, let’s calculate the annual energy usage and cost both before and after virtualization.
Before Virtualization:
Annual Energy Use = 100 kW * 8,760 hrs/yr = 876,000 kWh
Annual Electric Cost = 876,000 kWh * $0.10/kwh = $ 87,600
After Virtualization:
Annual Energy Use = 75 kW * 8,760 hrs/yr = 657,000 kWh
Annual Electric Cost = 657,000 kWh * $0.10/kwh = $ 65,700
The virtualized data center is clearly more energy efficient. In fact, the data center can even be more energy-efficient if the cooling load is now decreased by increasing the set point. 

If we don’t know how to utilize PUE to assess the effects of changes in the data center, it’s a useless benchmark.

Should we avoid virtualization because we know it would certainly increase the PUE? Absolutely not! When looking at our PUE across time, it’s crucial to notice when the virtualization occurred. We must track any changes in the infrastructure or IT load in addition to PUE so that we can link the changes to the PUE value.

It’s important to remember that PUE might be influenced by different variables. Tradeoffs between availability and energy efficiency will always exist. When data center equipment is loaded, it runs more efficiently, from cooling to UPS to server power supply. Maximizing the support infrastructure utilization to match the server loads is essential. It’s a constant balancing act to maintain the temperature to match the IT loads. Avoiding overcooling, but at the same time avoiding hotspots.
While redundancy improves availability, it also reduces the strain on many systems. As the load is lowered, the energy efficiency is reduced as well. Virtualization and consolidation will raise the PUE while lowering total energy use. Raising the server intake temperature may lower the PUE. But if the extra power required for server fans is higher than the cooling savings, the overall energy use may actually increase.

What Else Should I Be Measuring?

PUE is best used for tracking the impact of changes made to the data center infrastructure. It is less useful for tracking the improvements resulting from reducing the energy consumption of IT equipment.
While it is important to reduce losses in the power system and the power used for the support infrastructure, we need to realize that the bulk of the power consumption in the data center goes to the IT load itself. If we can reduce the IT load, we will reduce the overall power required for the data center.
Lowering the IT load has a multiplier impact. It lowers power system losses and the amount of power needed for support infrastructure. This is referred to as the “cascade effect. This explains that “reductions in energy consumption at the IT equipment level have the greatest impact on overall consumption because they cascade across all supporting systems.”
Let’s have a look at how the cascade works. If one watt can be saved at the IT load, it will cut losses in the server power supply, power distribution, UPS, cooling requirements, and building transformer and switchgear. As a cascade effect, saving one watt at the IT load may result in total energy savings of two or more watts.

AKCP Online PUE Calculator

Not only do you save operational overhead, but you lower the carbon footprint of the data center, which can open the possibility of government rebates and grants from green energy initiatives.

But how do you go about making these savings and implementing the changes in your data center that are required? AKCP has a wide range of sensors for your data center including cabinet thermal maps and differential air pressure sensors. These sensors allow you to make changes in your data center and with AKCPro Server see in real-time the effect it has on your PUE. You can ensure that your data center is running at optimal condition, without violating ASHRAE recommended rack inlet temperatures and ∆T as well as ∆P values.

Online PUE Calculator calculating energy efficiency

Online Power Usage Effectiveness Calculator

Power Monitoring Sensor

The AKCP Power Monitor Sensor gives vital information and allows you to remotely monitor power eliminating the need for manual power audits as well as providing immediate alerts to potential problems. The AKCP Power Monitor Sensor is specifically designed to be used with AKCP’s sensorProbe+ and securityProbe base units. It has been integrated into the sensorProbe+ and securityProbe’s web interface with its own “Power Management” menu, allowing multiple three-phase and single-phase Power Monitor Sensors to be set up on a single sensorProbe+ or securityProbe depending on which readings are required. Please check the sensorProbe+ Modbus manual or the PMS manuals on our website for more detailed information on this. Power meter readings can also be used with the sensorProbe+ and AKCPro Server lives PUE calculations that analyze the efficiency of power usage in your data center. Data collected over time using the Power Monitor sensor can also be viewed using the built-in graphing tool. Combining this durable Power Monitor Sensor with the sensorProbe+ and securityProbe creates an IP-enabled power monitoring capable of monitoring:

  • Phase Line Voltages
  • Current
  • Power Factor
  • Active Energy
  • Active Power

By employing the complete AKCP ecosystem of products, cabinet thermal maps, AKCPro Server, and power monitoring sensor work together to give a complete analysis and assistance in cutting your power costs and improving your PUE.

Reference Links:




Clarissa GarciaPUE Impact On Data Center Costs