To enable global digital connectivity, data centers have proliferated in every corner of the world. Today’s distributed data centers have also become more interconnected, providing the foundation for our modern digital economy and society. At the same time, new threats to these data centers have also arisen.
These threats include extreme weather, power shortfalls, malicious attacks and political instability. Couple these threats with the fact that our world is now more reliant on digital infrastructure than ever before, and the challenge becomes clear: Data center operators must do everything we can to make our facilities more reliable.
Threats to data center reliability fall into two main categories: operational threats and security threats. In this blog post, the first of a two-part series, we’ll explore the operational threats.
Designing data centers for reliability despite extreme weather
It may help to stop and think about what colocation data center providers actually do for customers: in the most basic terms, they provide power and cooling for IT equipment and a secure space in which to host that equipment. To ensure reliable delivery of these services, data center operators must prepare for extreme weather conditions that have the potential to interrupt the power or cooling supply and/or damage the facility itself.
A colocation provider can’t stop these unexpected events from happening, so the goal should be to predict them as far in advance as possible. This allows the provider to inform customers earlier and help them reduce the impact on their operations. Like many other data-driven sciences, meteorology is advancing through the application of AI models. By finding patterns in training datasets composed of decades of observational weather records, AI models can help human meteorologists make better, faster forecasts.[1] Data center operators can act on these forecasts to give customers the warning they need.
At Equinix, we’re working to automate notifications to give our customers ample opportunity to divert traffic away from the impacted facility before a predicted incident occurs. Thanks to our globally interconnected data center platform—and the fact that most customers are deployed in more than one Equinix facility—we can remove the risk of our customers over-relying on any single facility. (Also, it doesn’t hurt that we have an industry-leading average uptime record of 99.999%+ across our global data center platform.)
Ensuring energy stability for data centers
Extreme weather patterns also impact energy consumption, thus creating another potential threat to data center reliability. We design our data centers with 50-year weather extremes in mind, but it’s undeniable that what was once extreme weather is now becoming more common. We’re working to address this by expanding the operating temperature ranges of our data centers to align with ASHRAE A1 Allowable (A1A) standards.
Specifically, in the facilities where we’ve piloted the transition, we’re moving operating temperatures closer to 80°F (27°C). In contrast, the industry average is about 72°F (22°C). Operating our data centers in a higher temperature range enables us to use less energy for cooling, adapt to changing weather conditions and reduce the risk of outages.
Data center operators are also uniquely incentivized to help increase the supply of renewable energy to the grid. There’s a direct correlation between the stability of energy grids and the reliability of the data centers on those grids. Although Equinix has no intention of becoming a grid operator ourselves, we do target projects that provide additionality—that is, projects that add new renewable energy capacity to local grids. In particular, we’ve invested in power purchase agreements (PPAs) to support new renewable energy projects in regions throughout the world.
In addition to powering IT workloads with renewable energy whenever possible, data center operators can apply AI to help run workloads more efficiently. For instance, AI can help identify how to schedule workloads during off-peak hours to reduce energy consumption while still providing acceptable results. Closer cooperation between a colocation provider and its customers is required to get this functional and optimized.
Managing data centers for resource optimization
We believe that effective management of resources—including but not limited to energy—is essential for ensuring the reliability of data centers. All data center operators and their supply chain partners should participate in the circular economy. This means reusing or repurposing resources whenever possible to reduce waste, and therefore help ensure those limited resources will be available when they’re needed in the future.
At Equinix, we do this through our Responsible Electronics Disposal (RED) program, which aims to refurbish and remarket the servers leaving our business to give them a second useful life. From an energy perspective, we do it by scaling our heat recovery and heat export efforts. This involves recapturing heat created by IT equipment in our facilities and exporting it to heat homes in local communities. This process gives a second useful life to energy we pull from the grid, while also replacing more expensive, carbon-intensive energy sources that would otherwise heat those homes.
Our 2023 sustainability report shows that we put 4,000 MWh of recovered heat back into communities. This amount by itself is not game-changing, but we’re working to continue scaling up our program. More importantly, it speaks to the fact that we all must approach the challenge of building a reliable, sustainable energy grid from a systemic perspective.
The data center industry has traditionally been a leader in working toward energy stability, and Equinix has been a leader within the industry. That said, the scope of the challenge is so great that no company can address it working alone. We believe that companies in all industries should invest in renewable energy production in the same way we have. Doing so will give us all greater control over power supply on the grid. In turn, this will help ensure we’re able to meet the growing energy demands of our global digital economy—now and into the future.
Learn more about the Equinix approach to sustainable, reliable data center operations
At Equinix, we believe in pursuing sustainability in every sense of the word—both in terms of protecting the planet and the communities that populate it, and in terms of ensuring our own ability to continue delivering the global digital infrastructure services our customers rely on.
To learn more about the sustainability progress we made last year—including some of the initiatives mentioned above—access our interactive sustainability report today.
[1] Lauren Leffer, AI Weather Forecasting Can’t Replace Humans—Yet, Scientific American, January 9, 2024.
Subscribe to the Equinix Blog