Data centers are a critical part of the global economy and are central to the increased innovation and digitization of human life. Global data creation and consumption is estimated to have grown from 9 zettabytes (or billion terabytes) to 120 between 2013 and 2023. This data usage includes data transmission, cloud computing, and big data analysis, underpinning digital technologies such as remote working, IoT devices, streaming services, and digital communications.
Up to 83% of enterprise workloads are held in the cloud, making data centers vital for businesses worldwide. This significant usage has changed what data centers look like, with enterprises needing to scale to meet data storage needs through enterprise data centers. Now the vast majority of the estimated 8,000 global data centers are either colocation or hyperscale, the latter of which are generally larger than 10,000 square feet.
How Much Power Do Data Centers Need?
Data center energy usage has been a major talking point, with data centers estimated to use as much as 3% of the world’s produced energy, which could grow to 4% by 2030. Some hyperscale data centers use energy on a very large scale, with Microsoft’s 700,000 sq. ft data center in Northlake, Illinois having the capacity to consume 198 MW of power or the equivalent of 150,000 homes.
In terms of energy needs, there are four main factors in a data center’s requirements: the facility and its operations team, industrial equipment including its mechanical, electrical and plumbing (MEP), its hardware, and its software (that is, the operations that the storage stacks are performing).
The CPUs that perform the operations in a data center’s servers are made up of billions of transistors prone to overheating, especially as they are used 24/7. For this reason, around 40% of a data center’s energy requirements are dedicated to cooling the space and the servers. It is essential, according to the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), to keep the ambient temperature in data centers between 64 and 81 degrees Fahrenheit (18 to 27 degrees Celsius) for optimum performance and minimizing equipment failure.
The majority of racks in the US use 16-20kW for cooling and operations. Since large data centers run thousands of racks, the energy demands are very high per sq. ft. This has led many large data centers to mass buy energy production, such as Amazon buying the output of an entire 50MW wind farm in Scotland.
Importance of Data Center Backup Power
Data centers are billion-dollar operations, with the equipment they contain and the purposes they serve being highly cost-sensitive to even relatively small downtime. Reliable and continuous power supply is vital for effectively running data centers and preventing downtime. Power failure has a major impact on data center operations and costs attributed to:
Equipment Failure: The heat created by server cabinets in a data center is drawn away by fans before being cooled with water or refrigerant. In the case of power failure, both the fans and pumps will stop working, leading to sudden rises in the temperature of servers. This means the servers must either be shut down completely or result in melting, fires, arc flash, and other destructive incidents.
Downtime: If data center servers are damaged or turned off to prevent damage, they can’t perform their data operations. This downtime causes significant losses, with enterprises unable to use applications, access data, or serve their customers, such as with streaming or games services. For data centers, these costs are measured per minute and can be very large. Surveys of data center operators show that over 60% of failures cost over $100,000, with 15% costing over $1 million.
Lost Data: Cloud operation transfers data elsewhere rather than stored permanently on specific servers. As such, destructive incidents due to power failure can potentially lead to data being completely lost. This would result in compensation payments and lost customer confidence for the operator.
Data center backup power, mostly generators, forms the backbone of the uninterruptible power supply (UPS) technology deployed at data centers. This backup power is designed to kick in to prevent equipment downtime and avoid destructive incidents. A data center’s backup power is crucial for powering critical loads and avoiding the major costs and reputational damage associated with equipment failure, service downtime and lost data.
How Data Center Backup Power Is Deployed
Data centers must ensure virtually 100% uptime and protect sensitive and expensive server racks and other equipment that perform their operations. This is achieved through multiple gensets, generator controls, and switches that automatically switch power from non-critical to critical loads.
The UPS system will be designed specifically for the site and according to the data center’s needs and business goals. The most effective will have multiple points of failure built into the system and include high levels of redundancy, such as 3N/2 or 2(N+1). In the event of power failure, the emergency generator system will start, and the switchgear transfers the load to the UPS system until normal or utility power supply is resumed. In some UPS systems, the data center backup power will include large-scale batteries to ensure no momentary delay for loads requiring an uninterrupted power supply.
A UPS setup for data center backup power involves planning for multiple and simultaneous scenarios. The gensets, starters, switchgear, and other equipment need to be regularly load-tested and certified for rated use to ensure effective operation when needed.
At Enercon, we have helped design, build, and install backup power setups in thousands of projects around the world. This expertise allows us to ensure high-quality deployments that meet a customer’s specific needs and are built to perform. To find out more about how Enercon can help you with your data center backup power or any backup power project, connect with our team today.