Maintenance of electrical infrastructure is critical to reducing damage from unplanned outages in data centers. Maintenance of UPS power supplies and generators and disaster planning should be part of routine data center operations.
Sudden power outages can paralyze data center operations. Power outages are detrimental to IT systems as they result in data loss, file corruption and equipment damage. On-premises data centers require a backup power system integrated directly into the power infrastructure to ensure critical systems remain operational.
The data center power infrastructure should support two modes of operation: regular operation and emergency operation. In normal operating mode, the data center is connected to the power lines of one or more utility companies, while some businesses may use diesel generators to meet their power needs.
However, diesel generators often have another important role in the power infrastructure: providing backup power in an emergency. In an emergency, businesses use one or more diesel generators to power the data center and provide energy in the form of alternating current (AC).
The importance of UPS power
Data centers typically deploy one or more automatic transfer switches to switch between utility and backup power. In the event of a mains interruption or failure, the transfer switch sends a start command to the connected backup diesel generator, which then provides power as it begins to generate. Once the utility power is restored, the transfer switch switches to the utility power supply and sends a shutdown command to the diesel generator.
The power infrastructure of a data center includes one or more UPS power supplies. UPS power supplies serve two key purposes: to protect IT equipment from voltage spikes and to provide short-term power in the event of a power outage.
Another common power infrastructure component is a power distribution unit (PDU), which is an advanced electrical outlet that draws power from a UPS power source and distributes it to IT systems. Power distribution units (PDUs) do not generate power spikes, which is why power distribution units (PDUs) are often used with UPS power.
Its power conditioning and routing is done through switchgear or main switchboards and transformers. This ensures that the power supply voltage and current are output properly. Its equipment configuration depends on the size of the data center, the number and type of IT systems.
ensure continuous operation
The data center power infrastructure must provide power to other equipment in the data center facility, such as lighting, alarms, sensors, fire monitors, cooling units, and dehumidifiers. Its power infrastructure must also provide power to IT systems, such as servers, storage devices and network components, so that they can operate without interruption even in the event of a power outage.
During normal operation, the power targeted to the IT system will run continuously through the UPS power supply. UPS power ensures that connected batteries are fully charged and ready to support emergency operations. Depending on the number and type of batteries, batteries can provide enough power to run IT systems for short periods of time.
Some UPS power supplies use flywheels or supercapacitors instead of batteries. A flywheel uses an accelerating rotor to store kinetic energy that can be converted into electricity. Supercapacitors use electrostatic charge to store electrical energy on demand.
UPS power supply is divided into three types: offline type, online interactive type, and online type. If the utility power is interrupted, the UPS power will switch to the powered state, convert it to AC power and send it to the PDU. A line-interactive UPS works much the same way, except that it regulates the mains as it powers the device to prevent voltage spikes.
An online UPS converts alternating current to direct current and charges the battery. Any remaining power is conditioned and converted back to AC for PDU output. Large data centers running critical workloads often deploy online UPS power supplies because they provide the highest level of protection, albeit at increased operating costs.
Regardless of which UPS is used, or how an organization configures backup power, the goal is the same: provide enough time to shut down IT systems or allow backup generators to start generating electricity. Ideally, less than a minute after the failure is detected, the backup diesel generator will start, restoring the UPS power to normal operation.
Most data center backup generators have enough fuel to run for 24 to 48 hours. The number of backup generators and the total voltage required depends on the specific power requirements of the data center.
Organizations must ensure that their generators operate safely, comply with applicable environmental regulations, and have administrators monitor exhaust gases such as carbon monoxide or nitric oxide.
Proactively plan outages
Every data center should have a disaster recovery plan that outlines the steps to take in the event of a power outage and the role each piece of equipment should play; it is critical to follow this plan during an outage.
This means shutting down IT systems in a prescribed sequence, ensuring backup generators are running and properly ventilated, monitoring temperatures to prevent overheating, and verifying that emergency systems, such as pumping equipment, are functioning properly.
There must also be a recovery plan once power is restored. IT systems should be recalibrated in the prescribed order and tested to confirm that everything is working as expected. Emergency equipment should be turned off to prepare for the next crisis. Most importantly, participants should maintain proper communications during and after a power outage, so that workers affected by power outages are kept up-to-date on the state of the infrastructure.
Data center power infrastructure also requires regular equipment maintenance and testing. In the event of a power outage, an improperly maintained UPS power supply or diesel generator can affect the operation of a data center, which can affect its troubleshooting time.
