Energy Redundancy in Data Centers: What It Is and How to Calculate Energy Needs

Data centers are the backbone of the digital world. Any failure in their power supply can lead to significant economic losses, critical disruptions, and damage to reputations. Therefore, power redundancy is an essential pillar to ensure operational continuity. Below, we explore this concept in depth, including practical examples of systems and equipment that implement it.

What is power redundancy in data centers?

Power redundancy involves having backup systems that automatically come into operation if the main power supply fails. These systems are designed to prevent disruptions, protect data, and comply with stringent industry standards, such as those defined by the Uptime Institute in its Tier classifications.

Key components:

1. UPS (Uninterruptible Power Supply):
   Systems like the Eaton 9PX or APC Smart-UPS are examples of equipment that provide immediate power through batteries to prevent outages until generators are activated.
2. Backup generators:
   Diesel generators from Caterpillar or Cummins are popular choices in data centers, offering long-term power during prolonged outages.
3. Redundant power lines:
   Data centers such as those of Equinix often have multiple power lines from independent substations, ensuring an alternative power source.
4. Battery banks:
   Tesla Powerpack and Schneider Electric are leaders in scalable battery solutions to provide power during critical events.

Why is energy redundancy important?

1. Minimizes downtime

Companies like Amazon Web Services (AWS) or Microsoft Azure operate Tier IV data centers, which guarantee less than 26 minutes of downtime per year thanks to complete redundancy. Every second of downtime on these platforms can translate into million-dollar losses.

2. Protects data integrity

Data centers like those of Google Cloud implement UPS systems with automatic transfer to prevent sudden outages that could corrupt databases or damage hardware.

3. Meets industry standards

Most large data centers operate under Tier III or IV standards, maintaining high levels of redundancy to meet the demands of enterprise customers.

4. Builds customer trust

Companies like IBM have built their reputation on offering highly available data centers thanks to the implementation of N+1 and 2N+1 configurations.

Levels of redundancy and practical examples

1. N Redundancy (no redundancy):
   • Description: A single power source with no backup.
   • Example: A small data centerA data center or data processing center (DPC) … for startups with a limited budget.
   • Use: Ideal for non-critical applications.
2. N+1 Redundancy:
   • Description: An additional backup component is added.
   • Example: Tier II data centers that use an extra generator alongside the main ones.
   • Use: Enterprise applications that require high reliability.
3. 2N Redundancy:
   • Description: Two completely independent systems.
   • Example: Data centers like those of Facebook, which maintain two completely separate power sources to supply the entire facility.
   • Use: Critical infrastructures that cannot afford failures.
4. 2N+1 Redundancy:
   • Description: Two complete systems plus an additional component for extra security.
   • Example: Equinix data centers in London feature two main sources and an extra system for complete redundancy.
   • Use: Companies with mission-critical services.

How to calculate energy needs: a step-by-step example

Step 1: List the equipment

Suppose a small data center has the following components:

• 10 servers with an average consumption of 400 W each.
• 5 network switches consuming 100 W each.
• 2 storage systems of 500 W each.

Step 2: Calculate energy consumption

Total IT consumption:
(10 × 400 W) + (5 × 100 W) + (2 × 500 W) = 5,500 W

Step 3: Add the cooling system consumption

If cooling accounts for 40% of IT consumption:
5,500 W × 0.4 = 2,200 W

Step 4: Sum total consumption

Total consumption = 5,500 W + 2,200 W = 7,700 W

Step 5: Add a safety margin

For redundancy and future growth, add 30%:
7,700 W × 1.3 = 10,010 W (or 10 kW)

Step 6: Convert to kVA

Using a power factor of 0.9:
10 kW ÷ 0.9 = 11.1 kVA

Selecting backup systems

UPS:

For a consumption of 11.1 kVA, a APC Smart-UPS VT 15 kVA would be recommended, offering sufficient capacity and a 15-minute autonomy to start the generators.

Generator:

A 20 kVA Cummins diesel generator would cover the needs, providing an additional safety margin.

Real-world implementation cases

1. Google Cloud: Their data centers implement 2N redundancy systems to ensure uninterrupted availability.
2. Amazon Web Services: AWS has geographically distributed data centers with 2N+1 configurations to maximize resilience.
3. Equinix: Their data center in Ashburn, Virginia, operates with Tier IV redundancy, ensuring that disruptions are almost nonexistent.

Conclusion

Power redundancy is essential to ensure the continuous operation of data centers. From small startups to giants like AWS and Google, proper planning of energy needs and implementing reliable backup systems is crucial for protecting data, meeting standards, and maintaining customer trust. Designing a robust infrastructure with real-world examples as reference ensures reliable and scalable performance over time.