Over the past decade, the demand for computing power has increased tenfold with the progressive digitization of the industry. At the same time, the energy requirement per gigabit in data centers is now twelve times lower than it was ten years ago. A study by the Borderstep Institute and eco – the Internet Industry Association – also concludes that a 30% reduction in CO2 emissions from data centers can be expected by 2030. This development is based on virtualization and the optimization of heating, cooling, and waste heat management, as well as the use of energy-saving technologies. To ensure that data centers are optimally positioned, both now and in the future, it is necessary to incorporate modern approaches in virtually every area involved.
1. Secure Power Supply and UPS: Unlocking Sustainability Potential.
No data center is complete without an uninterruptible power supply (UPS). In the event of a mains power failure, the UPS system's batteries temporarily take over until the diesel generators kick in. New concepts are now being applied to batteries and battery rooms compared to traditional systems: modern, maintenance-free batteries are continuously climate-controlled using intelligent charging electronics. In the event of a total external power failure, they can supply the data center with full-load power for approximately ten minutes. During this time, the diesel generators start up, ensuring a longer power supply.
Saving Energy with DC Power:
Hyperscale companies, in particular, are implementing an interesting strategy in this area related to the concept of sustainability. The UPS system converts AC power from outside into DC power to charge the emergency power supply batteries. The positive effect is that using DC power significantly reduces heat emission and energy consumption. Conventionally, DC power is converted back to AC power to supply the computer hardware in case of a problem. However, new concepts are based on using only DC power in the data center and eliminating the energy-intensive conversion process. The Open Compute Project (OCP) and the Open19 initiative aim to help eliminate data center inefficiencies with open specifications and the supply of DC power to the racks.
2. Security, Fire Protection, and Environmental Control Systems:
The strictest security standards are applied in the data center, not only for defense against cyberattacks but also for structural security, access control, and system redundancy. Fire protection plays a crucial role, as there is almost nowhere else where energy is as concentrated as in a data center. Therefore, a dense network of detectors and an early fire detection system are just as important as a fire suppression system. The air in the data center is constantly extracted, and the concentration of particles can be measured using laser light. In the event of a fire, the suppression system floods the data center, smothering the flames. Simultaneously, it typically triggers an automatic alarm at the fire department, and efforts are made to maintain overall operations.
Sustainable Climate Control and Exhaust Heat Recovery:
Air conditioning also plays a crucial role, as hardware emits almost 100% of the energy it needs as heat. Furthermore, excessive temperature increases can damage sensitive hardware. Hot and cold aisle containment has become standard practice. New concepts, such as waste heat recovery, aim to utilize the heat generated by data center hardware to, for example, heat other parts of the building.
3. Data Center Interior: Server Racks in Smart Architecture:
Heat reduction is often achieved by constructing a raised floor in the server room. This is a simple principle. Hot air exiting the servers is drawn in by a recirculating air cooling unit, cooled by a heat exchanger, and then returned to the racks. This air circulation ensures a suitable ambient temperature in the heart of the data center, where servers, storage, and network systems are housed in standardized 19-inch racks. Modern racks are designed to be as flexible and space-saving as possible, supporting high port density, individually adaptable to IT applications, and offering ample space for cabling. This is especially important given the increasing use of "Spine-Leaf" architectures, which require significantly more cabling than conventional approaches. The racks' ability to quickly and easily replace IT hardware for new applications, as needed, also provides future-proofing.
4. Data Center Cabling: Structured and Fail-
. Passive network components, such as cables and connectors, which redundantly connect all systems to each other and to the internet, must allow the use of both copper and fiber optic cabling networks. Structured, fail-safe cabling ensures the success of a data center concept and seamless operation, while also complying with current standards. Data center technology is constantly evolving. Therefore, cabling must be compatible with all current and future communication systems and be neutral, not only with respect to the transmission protocol but also with respect to the end devices. Otherwise, future changes to the data center infrastructure risk incurring considerable costs.
Ready for New Data Transmission Speeds:
The type of data transmission, transmission distance, fiber, and connector types must be selected based on individual needs. Large data centers, in particular, face an increasing "need for speed." New cabling systems, such as PreCONNECT® SEDECIM, support the new 400GBASE-SR8 transmission speed, enabling transmissions of up to 100 meters at 400 Gbit/s.
These particularly high reliability requirements demand top-quality connectors. This includes a special surface treatment that ensures the connectors are dirt-repellent and self-cleaning, preventing connection problems.
