In the era of energy efficiency, moving towards agile and responsible data centers

The size of the digital universe doubles every four years. The approximately 2.5 billion GB of data created every day must be processed, stored, and delivered to users who are increasingly demanding regarding quality of service. Data centers, which host IT services, must therefore be designed and operated to minimize the risk of downtime. To this end, operators have long prioritized equipment redundancy and facility security. The main corollary of this has been the oversizing of cooling systems, which alone account for nearly 40% of a data center's total energy bill (2). This oversizing, which continues to be observed in the majority of IT rooms, leads to electricity consumption far exceeding actual needs. While kWh prices in France are on average 25% lower than in the rest of Europe, this usage is not responsible from an ecological perspective.

Aware of these economic and environmental challenges, data center operators began working on the issue of energy efficiency several years ago. They are now seeking to master the entire "energy cascade," meaning reducing consumption at all levels: IT components, technical equipment, layout, and the integration of IT rooms into their geographical ecosystem.

The first area of optimization concerns IT components. The development of superconducting materials, particularly for processors, allows for better heat distribution and operation at higher temperatures—the temperature commonly accepted today in the IT room by servers is 23-24°C, compared to 16°C in the late 1980s. Furthermore, in many cases, servers are used at less than 50% of their capacity while consuming almost as much energy as a server at 100% load. Better sizing, coupled with the use of technologies like virtualization, optimizes their use and efficiency. Finally, it is important for procurement policies to integrate the energy weight of IT equipment: older units must be replaced by new-generation hardware that is less energy-intensive, dissipates less heat, and can withstand higher temperatures.

Second area: optimizing the most energy-consuming technical equipment, namely those that provide cooling and power supply for the data center. Opting, for example, for new-generation cooling production or electrical distribution systems allows for both better equipment efficiency and reduced energy consumption and losses.

The architecture of the computer center is the third key parameter in reducing energy consumption. Take the case of a data center operating at 75% of its IT load level. Its operating cost is six times lower than that of an IT room with a 10% load rate. However, occupancy rates are not always at their highest, and hosting needs fluctuate both upwards and downwards. To respond to this volatility, data centers are increasingly designed modularly, with independent and adaptable components put into production as the load increases. For existing data centers, significant energy and operational gains can be achieved through room re-urbanization operations, such as server densification per m² (increasing production within the same footprint) or hot and cold aisle containment. In this regard, bulk air blowing is an aberration, as fresh air is not directly supplied to the IT equipment to be cooled. To maximize the efficiency of cooling systems, it is essential to separate thermal flows to channel the cold air and blow it as close as possible to the equipment based on its load level.

Finally, the last element of the energy cascade: the geographical location of the data center and its integration into the target environment. Several parameters must be studied, including temperature, humidity levels according to the seasons, and atmospheric pollution. Most of the time, systems that use so-called free energy (air, water) to cool data centers significantly reduce the energy bill. However, they must be coupled with humidifiers and filtering systems. It is therefore necessary during the design phases to properly evaluate the needs and constraints related to the future operation of the room to make the most appropriate choices.

The data center sector is currently going through a transition phase. This transition is leading operators to seek energy control, flexibility, and agility in the operation of their rooms. Innovation and evolving technologies are opening the way to new sources of energy savings.

Due to the ability of hardware to withstand higher temperatures, it is now feasible to more systematically use cooling systems that utilize outdoor air temperature (free-cooling, free-chilling), groundwater, or rivers (geo-cooling). Furthermore, the potential of Data Center Infrastructure Management (DCIM) solutions is still under-exploited for monitoring IT rooms and detecting energy-intensive areas. Initiatives to better utilize dissipated heat by recovering it to meet the heating needs of the rest of the building, or even the immediate neighborhood, offer great prospects. Finally, installations of alternative energy sources (photovoltaic panels, etc.) to power part of the data center equipment are multiplying worldwide, with positive results in terms of energy efficiency.

In conclusion, many proven solutions already exist to lower the energy bill of data centers at all levels, without compromising service continuity and without requiring excessive investment. Every "small" optimization operation is a stepping stone for the ecological and economic structure. It is a matter of finding the right combination based on existing conditions.

(1) Source: RTE – Réseau de Transport Electrique
(2) Source: Gimélec