The changes today in the data center with virtualization, blade chassis and storage devices brings changes in how we provide an Uninterrupted Power Supply (UPS) to them. The typical legacy data center was filled with underutilized 2U – 5U servers and networking gear. With Virtualization we have taken what used to be 2-3 racks and put it in 1 rack, by virtualizing your servers so you have higher utilization on your CPU or consolidating to a Blade Chassis and have more processing power in a smaller footprint. Whenever we change how a data center looks, we also have to change how that data center will receive UPS.
Since every critical IT device in a data center requires a power feed, a UPS is one of the most important units in a data center. The function of the UPS is to clean the power from the utility to your equipment and provide a desired runtime in the event your utility power is out. By cleaning the power from your utility the UPS ensures that your sensitive IT equipment is always getting the voltage that is desires. An online double conversion UPS does this best by passing the power thru a rectifier and inverter to condition the power. In the event of a power failure the UPS is designed to run on batteries for a given amount of time to either gracefully shut down your equipment or give your generator enough time to start up and provide power to your IT load.
Levels of redundancy are always up for debate in the datacenter. N, N+1, and 2N are some of the most common levels of redundancy for a datacenter UPS with 10-20 racks. 1N is just a single module UPS that is protecting the IT load. This provides you with UPS power in the event you lose utility power and in most cases a static switch in the event of a UPS failure. A static switch allows the UPS to pass straight utility to your load thru the UPS. If you lose power when you are in static bypass you will also lose power to your IT equipment. N+1 is a single UPS but has redundant components within the UPS. Depending on the manufacturer, the power module, battery module, and intelligent module are the most common redundant components. 2N configurations are when you have an A + B side. This give the customer two UPS’s with power feeds (A+B side). With 2N configurations you have the ability to have a complete UPS failure and still provide UPS power to your IT loads. Keep in mind, with higher levels of redundancy comes higher cost. Analyze the cost of the redundancy vs. the cost of downtime to make the best decision for your business.
A typical rack used to be 2-3kw so if a customer has 15 racks of equipment they have 30kw – 45kw worth of power. Today companies have the same computing utilization in 4-5 racks and still use 30kw – 45kw worth of power. With racks in the 7-11kw range we have to change how racks are being powered with Power Distribution Units (PDU’s) and UPS’s. So we don’t fill the rack with many single PDU we look to feed the rack with 3phase PDU’s. This allows us to bring 5.7kw – 20kw+ of power thru a single PDU, giving us more for cables and cable management in the back of the rack.