During an Uptime Institute Leadership Conference in Perth, Rana Qadir, VP of Sales for Australia, New Zealand, and Greater India, pointed out that despite the AI boom, sustainability has been neglected.

In addition, he explained that some large collocation providers and operators have once made aggressive public statements and commitments about net zero commitments yet have been unable to explain how unfolding tech trends like AI would impact these targets in the future.

Indeed, this is a question the DC industry needs to address as AI readiness becomes increasingly important.

Why Sustainability Matters

DCs are major consumers of electricity – estimated to have taken up an estimated 1% to 1.3% of global electricity demand in 2022, and this is expected to increase as DCs start to ramp up rack density in response to the demands of big tech trends like AI, IoT and edge computing, and data-heavy applications.

Moreover, higher rack density will require higher cooling capacities, contributing to the energy demand in DCs. The International Data Corporation (IDC) recently forecast that AI DC energy consumption will grow by a CAGR of 44.7% by 2027, with AI workloads consuming an increasing portion of all DC electricity.

As part of the same report, IDC also predicted that global DC electricity consumption would double by 2028.

These will affect the following aspects of data center operations:

Environmental Impact: DCs' carbon footprints, especially those utilizing fossil fuel-powered grids, will increase along with their contribution to greenhouse gas emissions.

Operational Costs Rising: In addition to a rise in utility costs, DC operators must also ensure that their facilities have uninterrupted power supplies to reduce downtime. This means investing in upgrades for existing backup generators or acquiring new power sources is necessary.

Peak demand charges may also apply when a DC's racks are engaged in more intensive workloads during certain hours of the day, adding more costs to the facility's overall expenses.

Regulatory Compliance: While APAC's data center regulations vary, those that aim to standardize DC practices for sustainability are on the rise.

At time of writing, Singapore, Japan and Australia have implemented green certifications that encourage energy efficiency (Singapore) or promote the inclusion of renewable energy sources (Japan and Australia). Also, South Korea and Australia have established renewable energy goals and/or carbon neutrality pledges impacting how DCs source energy and manage waste.

Furthermore, DCs may be subject to other environmental regulations, such as those governing energy-efficient advanced cooling solutions and water usage, or those mandating infrastructure resilience, disaster recovery, and business continuity planning.

Deploying Real-Time Monitoring & Smart PDUs for Sustainability Planning

Sustainability planning must begin with an understanding of the current energy demands of a facility and where improvements can be made to increase efficiency. It starts with real-time monitoring and metrics, and smart sensors feeding live data into a facility's DCIM for analysis and actionable insights.

Below are a few examples of how data and monitoring can help a team pinpoint where to begin:

·         Temperature: Cold and hot spots detected may indicate an ineffective cooling system or the need to upgrade HVAC or CRAC equipment.

·         Humidity: The relative humidity of an environment can affect equipment's operating conditions and life expectancy. When coupled with temperature data, it can also indicate over-cooling.

·          Differential Airflow: Properly managed differential airflow ensures that cold air is distributed effectively to racks and hot air is effectively expelled. The presence of abnormal readings in either of these could indicate uneven cooling, which can damage equipment.

·         Air Pressure: Maintaining ideal air pressure ensures that air flows properly from cold to hot aisles, or where it is needed to cool racks and equipment. For facilities sensitive to airborne pollution, positive air pressure will also prevent dust and contaminants from entering from outside.

·         Dust Levels: It is possible for dust accumulation to cause static electricity buildup in equipment, clog vents, and restrict airflow, leading to equipment failure due to overheating. Keeping track of this will allow teams to schedule preventive maintenance and extend the life of infrastructure.

Monitoring the facility environment with sensors allows teams to optimize cooling strategies and identify less efficient cooling systems that can be replaced and phased out in order to reduce power consumption and carbon footprints.

Besides sensors, smart PDUs provide real-time power monitoring capabilities that can assist DC teams in understanding power usage and optimizing Power Usage Effectiveness, an international standard that measures DC energy efficiency. For example, Raritan's intelligent rack PDUs, PX4 provides real-time insights with ±0.5% metering accuracy on power quality, branch circuit monitoring, and total harmonic distortion along with other metrics.

Total harmonic distortion is an important metric for teams that are concerned with sustainability - it measures how much the rack's electrical load distorts the facility's power input. Waveform Capture, supported by the Xerus technology platform, allows teams to visualize the power being fed to the PDU or distributed to the PDU's outlets, enabling data analysis, trend prediction, and data anomalies.

On top of that, smart PDUs like the PX4 also have load balancing capabilities that optimize power distribution across multiple outlets, spreading electrical demand more evenly across available circuits, reducing the risk of overloaded circuits and maximizes use of available power. Over time, this supports more efficient power use and reduces overall energy consumption for the rack.

Conclusion

Using both sensors and smart PDU, teams can achieve better oversight over their facilities’ current power usage and environmental conditions, project power requirements needed for future retrofits and expansion, and identify areas where power supply and existing infrastructure need to be replaced or reconsidered for smarter power management.

Over time, the data collected about power consumption and environmental needs up to the rack level will help teams make data-driven decisions and predictions as well– when to schedule maintenance, where and when to adjust temperature controls, and which parts of the infrastructure require repairs or replacements, resulting in a more efficient facility with lower operational costs, less downtime,  and best of all, a lower environmental impact that can meet sustainability goals of the country it operates in.