Unlock Efficiency and Performance in Data Centers

Note: This article first appeared on Area Development’s website.

Aging data centers face mounting pressures to evolve as 55 percent of facilities are over a decade old. Strategic upgrades can unlock new levels of reliability, scalability, and energy efficiency.

The digital world is evolving at breakneck speed, yet many data centers are struggling to keep up due to aging infrastructure. According to the Uptime Institute, 55 percent of data centers are over a decade old, and nearly 25 percent are older than 15 years. With such a significant portion of facilities nearing or exceeding their intended lifespans, modernization is no longer a luxury—it’s a necessity. From optimizing power distribution to upgrading cooling systems, there are numerous ways to breathe new life into these critical facilities and meet the demands of high-performance computing.

A fresh look at power distribution

Modernizing a data center often begins with its power distribution system. Today’s IT equipment is designed to handle voltages between 85 to 264 volts AC, making voltage conversion an enticing opportunity for boosting efficiency. Distributing power at 415 volts AC for conversion from 240 volts AC, for example, can yield significant benefits—but compatibility is key. Servers, cooling systems, UPS (uninterruptible power supply) units, and transformers must be assessed to ensure they can handle the shift—or upgraded if they can’t. Equipment originally designed for 480 volts AC, for instance, might encounter performance issues when operated at lower voltages.

The payoff of voltage conversion is substantial. Systems running at higher voltages, such as 415 volts AC, require less current to deliver the same power, minimizing losses due to heat and resistance in cables and transformers. When executed with high-quality equipment, voltage conversion introduces only minimal losses while improving overall system efficiency. Additionally, IT devices like servers tend to perform best when operating closer to the upper end of their voltage range, reducing internal stress and heat generation. This, in turn, eases the load on cooling systems, creating a ripple effect of efficiency gains.

Upgrading the backbone: The UPS

The uninterruptible power supply (UPS) is the lifeline of any data center, and upgrading it offers a prime opportunity to improve reliability and efficiency. Modern UPS systems range from traditional battery-based units to rotary and flywheel systems. Each has its strengths. Battery-based systems are familiar and versatile, while flywheel and rotary options excel in efficiency and longevity, requiring less maintenance.

When selecting a new UPS, it’s crucial to consider both current and future power needs. Will it handle the growing load as your data center expands? Can it provide redundancy to ensure uninterrupted operations if a single unit fails? In the world of high-performance GPU servers, additional considerations must be made to ensure the UPS can support the sudden load increases associated with AI workloads. Be sure to assess the power requirements these dynamic computing environments require and design the UPS system accordingly.

Elevating power with overhead busway

Overhead busway systems are gaining traction in data center design for their ability to optimize space and streamline power distribution. By routing power overhead instead of underfoot, these systems free up valuable floor space, making room for additional racks and improving airflow.

The modular nature of the overhead busway is another advantage. Expanding or reconfiguring your data center becomes much simpler, allowing you to adapt to changing requirements with minimal disruption. Better airflow is also a direct benefit, as removing cables from the floor eliminates obstacles to efficient cooling. Maintenance becomes easier and safer, with power cables clearly visible and accessible. Over time, the improved efficiency and reduced infrastructure complexity can offset the higher upfront costs of busway installation.

Cooling for the future

Cooling is a perennial challenge for data centers, especially as power densities rise. Traditional air cooling can only go so far, and hybrid liquid cooling methods are becoming the gold standard. Techniques like direct-to-chip cooling, where liquid directly cools the most heat-intensive components, offer unmatched efficiency.

If a complete overhaul isn’t feasible, retrofitting existing racks with rear-door heat exchangers or immersion cooling systems can still make a significant impact. While these options may require infrastructure modifications, such as adding liquid distribution piping or upgrading chillers, they often pay for themselves by reducing overall energy consumption. Hybrid liquid cooling systems also help lower the facility’s Power Usage Effectiveness (PUE), a critical metric for data center efficiency.