Why Software Is the New Layer of Critical Infrastructure

By Javier Zurera

In the data center industry, we often say that what cannot be measured cannot be improved. But in today’s context of energy challenges and increasing climate-related requirements, that principle is no longer enough. Today’s mantra should be:

What cannot be visualized in real time is being wasted.

For years, cooling management in data centers has been a mix of conservative engineering and, let’s be honest, a certain degree of “fear of heat.” When in doubt, operators have tended to overcool. The result is an industry that, while essential to the digital economy, continues to carry a silent but costly burden of thermal inefficiency.

This is where software-driven optimization and AI solutions, with EkkoSense at the forefront, are changing the game.

The Colocation Challenge: Managing Uncertainty

If there is one environment where thermal management becomes a highly complex puzzle, it is the colocation data center.

Unlike a corporate data center with relatively homogeneous hardware, colocation facilities host a diverse ecosystem of customers, rack densities, and airflow configurations that are constantly evolving.

In this environment, operators face a paradox: they must meet strict temperature and humidity SLAs (Service Level Agreements) for demanding customers while simultaneously reducing their PUE (Power Usage Effectiveness) to remain competitive and sustainable.

Historically, the answer has been overcooling. If the SLA states that cold aisle temperatures must remain below 26°C (78.8°F), operators routinely deliver air at 20°C (68°F) “just to be safe.”

That “just to be safe” approach is responsible for millions of dollars in unnecessary electricity costs and a carbon footprint that is no longer acceptable, particularly given reporting requirements under the European Union’s Energy Efficiency Directive (EU) 2023/1791.

The Digital Twin and Real-Time Analytics Revolution

The value of solutions like EkkoSense goes far beyond monitoring. Their real strength lies in creating an interactive Thermal Digital Twin.

By deploying a network of strategically placed wireless sensors, a previously “blind” data hall is transformed into an environment where every airflow pattern and every degree of temperature variation becomes visible.

We recently analyzed a project that perfectly illustrates this transformation: a colocation data hall with an IT load of 230 kW.

By today’s standards, this is a mid-sized facility, but one with significant thermal complexity due to the diversity of customer deployments.

Sensor installation and software configuration were completed in just three days.

No major construction work. No service interruptions.

Within 72 hours, the facility moved from theoretical, model-based management to real-time operational decision-making driven by actual data.

The 20 Days That Changed the Strategy

Following deployment, the project entered a 20-day analysis phase.

During this period, machine-learning algorithms began correlating IT load behavior with the performance of the cooling infrastructure, including the CRAC and CRAH units.

What emerged was a familiar story:

• Hotspots caused by airflow recirculation or insufficient cooling.
• Areas receiving far more cooling than the installed hardware actually required.

Data hall conditions before optimization, showing hotspots and areas operating below 20°C (68°F).

The Goal: Safely Raise the Operating Temperature

For many data center operators, increasing room temperature feels like walking a tightrope.

A small adjustment can appear risky when there is limited visibility into how thermal conditions will evolve throughout the room.

With complete thermal visibility, however, supply-air temperatures can be raised safely and confidently, ensuring that even the most remote server supporting the most critical customer remains within ASHRAE recommendations and contractual SLA requirements.

As room temperatures are increased in a controlled manner, cooling-system efficiency improves dramatically.

Every degree increase in the cooling setpoint delivers meaningful energy savings that are both measurable and fully verifiable.

Anomaly Detection: Getting Ahead of Failure

The software’s value is not limited to energy savings. It also provides peace of mind.

One of EkkoSense’s most powerful capabilities is its ability to detect anomalies in cooling equipment performance.

In many colocation environments, degradation in an air-conditioning unit is often discovered only after room temperatures have already begun to rise dangerously.

The software identifies these issues much earlier.

By continuously comparing actual thermal performance against expected behavior, the platform can pinpoint units that are working harder than necessary or operating inefficiently.

During the 20-day analysis period, for example, the system identified CRAH units delivering significantly colder supply air than their configured setpoints because internal temperature sensors had drifted by approximately 2°C.

Cooling unit configured with a 20.5°C (68.9°F) supply-air setpoint but actually delivering air at 18.7°C (65.7°F).

This predictive management approach enables intervention before failure occurs.

Rather than waiting for equipment to break down, operators can take action as soon as software detects behavior that deviates from optimal performance.

In the data center industry, resilience is ultimately the product we sell, and software has become one of its strongest guarantees.

The Business Case: An ROI That Challenges Skepticism

Let’s talk about the numbers.

For this 230 kW IT deployment, the objectives were ambitious but entirely achievable thanks to the technology:

1. 20% Reduction in Cooling Energy Consumption

   This is not a marginal improvement. In an industry where energy expenses may represent 40% to 50% of operational costs, a 20% reduction in cooling consumption has a direct and significant impact on the bottom line.

2. Eight-Month Return on Investment (ROI)

   In an industry accustomed to infrastructure payback periods of 10 to 15 years, a project that pays for itself in less than three quarters is nothing short of exceptional.

This accelerated ROI is driven by the low implementation barrier, just three days for deployment, combined with substantial and immediate operational savings.

It is arguably one of the highest-impact, lowest-risk efficiency investments available to data center operators today.

Sustainability: Turning Intent into Action

The global spotlight on data center energy consumption has never been brighter.

Regulators, customers, and investors increasingly expect operators not only to improve efficiency but also to prove it with transparent, auditable data.

Solutions such as EkkoSense enable colocation providers to deliver real sustainability reporting to their customers.

Rather than allocating overall facility consumption through rough estimates, operators can demonstrate how software-driven optimization actively reduces the emissions associated with each customer’s IT infrastructure.

Conclusion: Software Is the New Hardware

For decades, the industry has attempted to solve data center challenges by adding more hardware, more cooling units, more containment systems, and more infrastructure.

While hardware remains essential, we are reaching a point of diminishing returns.

The next frontier of efficiency lies not in the compressor, but in the algorithm.

The ability to understand, in real time and with surgical precision, exactly what is happening to airflow inside the data hall is becoming the key to efficient data center operations.

This 230 kW project demonstrates that digital transformation does not have to be a multi-year journey.

Three days to deploy. Twenty days to understand. Eight months to achieve payback.

That is the pace of the data center of the future.

Implementing thermal intelligence is not merely a smart financial decision. It is an operational responsibility.

In a world that runs on data, we can no longer afford to manage the infrastructure supporting it with tools from the last century.

The time for assumptions is over.

It is time to replace guesswork with certainty.