Data Center Power Redundancy and Its Importance

October 11, 2023

Data centers rely on power for nearly everything. Losing utility power, malfunctioning hardware or an end-of-life replacement can result in thousands or even millions of dollars lost per hour. These risks are why backup power is crucial in overall data center design.

What are a data center’s redundant power supplies? These infrastructure units include uninterruptible power supply (UPS) systems, standby generators and power distribution components like switchboards and lines. Because each part is vital in delivering essential power to the data center, having one or multiple equipment duplicates can maximize availability.

The Importance of Data Center Power Redundancy

Data center power redundancy is important because it mitigates risks, helps companies protect sensitive data and reduces significant downtime costs. Approximately 60% of system failures cost companies over $100,000 to repair damage. Challenges such as power outages, natural disasters, equipment maintenance and networking complications are highly likely occurrences, so companies must take them seriously to protect their data and operations.

How to Define N for Your Data Center

To understand the approaches to data center power redundancy, we need to define N. This variable can designate a data center’s total power needs (measured in kW) or the number of non-redundant components in the power supply and distribution chain. For example, if your data center requires six UPS units for full operation, N is six. The same rule applies to other components, such as power generators, switchboards and HVAC units.

N is the baseline capacity and means a single point of failure exists in the system. A facility cannot withstand a disruption such as a power outage, hardware malfunction or scheduled maintenance if it operates with an N architecture at its full capacity.

Types of Data Center Power Solutions

You can implement several different redundancy models in your data center depending on your facility’s needs. For example, you can apply an N+1 redundancy model for your cooling system while applying a 2N model for your UPS systems. Consider the following most common data center redundancy models and how they work:

N+1 Redundancy

The N+1 configuration supplies a minimal resiliency level with a single component. For example, you may add one UPS system or one generator to your N architecture. This single component can take over one primary system’s load during an outage, malfunction, replacement or scheduled maintenance. Design standards typically recommend installing one spare unit for every four units a data center needs to support full operational capacity. The N+1 model is the most affordable and most energy-efficient option, but it offers the least backup.

N+2 Redundancy

An N+2 redundancy model is an extension of the N+1 configuration. To implement this model, you would add two backup systems to your architecture instead of one. Since the N+1 redundancy model can only support one unit and presents a risk if multiple units fail, the N+2 model allows data centers to minimize this risk.

2N Redundancy

2N design keeps a separate system capable of handling a data center’s entire power needs. For example, a data center may install five spare UPS units if its system requires five primary units to function at full capacity, creating a system architecture with ten units total. A 2N model often features two mirrored, identical systems but may also comprise different equipment makes and models. This configuration results in a 100% required capacity and 100% stranded capacity.

The 2N redundancy model significantly reduces the risk of downtime because it implements independent distribution systems. An operator can perform maintenance on a system’s entire set of units without halting operations when a data center has a 2N model in place.

2N+1 Redundancy

The 2N+1 model takes the 2N model a step further and is the highest obtainable redundancy level. It provides a full set of backup units plus an additional component for extra protection. This configuration can maintain operation in the event of multiple unit failures and withstand an entire primary system shutdown, sustaining an N+1 redundancy if a center’s main units lose power or undergo maintenance.

3N/2 Redundancy

3N/2 redundancy allows data centers to partially reach 2N redundancy but with operating and capital costs closer to the N1 model. This distributed redundancy model implements additional capacity based on a system’s load. It can help companies reduce costs, but it can also present load management challenges. To use this configuration effectively, the system must carefully manage the load from the power supplies in use.

Note that these same formulas above can also serve as data cooling system configurations.

How to Choose a Data Center Power Redundancy Configuration

You may assume that having more backup systems is better by default. Instead, the redundancy level you need depends on multiple factors such as:

  • Budget: Consider how much your company can afford to spend on a backup system and the cost of downtime if your units malfunction or lose power.
  • Business goals: A growing business requires dependable system operation. It’s important to implement a high-level redundancy model if you plan to expand your operations significantly.
  • Risk tolerance: An IT department’s risk tolerance is also an essential consideration. Your company can withstand more risk and implement an N+1 model if you handle minimal sensitive data or perform non-critical operations, but a 2N or 2N+1 model is more appropriate if you operate in a highly regulated industry, provide critical services, rely on the cloud or are vulnerable to cyberattacks.
  • Security: Consider your company’s security requirements. Having adequate backup systems can help increase security.

The larger your business and data center operations, the more likely you are to invest in a power infrastructure with more resiliency. For example, large companies that cannot allow even minor operational disruptions depend on a 2N+1 model for optimal reliability.

In critical data center applications like health care facilities where backup power is necessary, redundancy can help mitigate wider-scale disasters from affecting operations. Certain data center certification levels, like the upper classifications of ANSI/TIA-942, require some form of redundancy. According to Uptime Institute, a vast majority of data center managers are seeking higher levels of redundancy than their baseline levels.

Minimizing Interruptions

While power redundancy helps prevent interruptions, you can further minimize interruptions with an automatic transfer switch (ATS). An ATS enables equipment to switch from a failed power source to its backup power source instantly. Installing an ATS eliminates the need for a technician to switch power manually, minimizing equipment downtime.

Contact DataSpan for Power Redundancy Management

Data center power redundancy is one of many areas of expertise we’ve gained throughout our over 45-year history by serving facilities of all sizes. DataSpan can help analyze your data center’s power needs and determine how to configure redundant power supplies for maximum uptime or cost efficiency. Contact us online or locate a representative to start mitigating risk and increasing uptime.