In the data center space, circularity efforts typically focus on scalability, refurbishment, reuse, and disposal of end-of-life IT products. Heat reuse – the process of capturing waste heat from data center operations for heating purposes in other facilities – is also part of the circularity realm.
IT teams typically refresh their computing hardware on schedules that vary from one to eight years, with typical refresh rates of three to five years. Refresh rates slowed from 2015 to 2020, but are expected to increase again after the next generation of more energy-efficient data center computing equipment begins to hit the market in 2021.
Operators are expected to manage end-of-life equipment responsibly, maximizing refurbishment and reuse of servers, storage, and networking products; reuse of components for spare parts; and recovery and recycling of critical materials.
The best practice in equipment management is to upgrade or refurbish equipment for reuse in a carrier’s own data centers. Efforts should be made to minimize the percentage of end-of-life equipment sent to landfills. Most countries, states and provinces have laws designed to encourage or enforce recycling and reuse.
Data center operators should understand and adhere to local regulations governing end-of-life products in their data center facilities. They will also need to do their due diligence to verify that the chosen product recycler(s) are handling the products and materials as promised and that the equipment does not end up in a landfill. Data center operators who fail to verify the responsible management of their discarded equipment have faced heavy fines and damage to their reputation.
The reuse of waste heat recovered from data center cooling systems has become a topic of intense interest in many jurisdictions, particularly in the European Union. Heat reuse is classified as a topic of circularity, as the heat generated during data center operation can be captured, recovered and used for beneficial purposes to heat other facilities, such as offices, swimming pools and greenhouses, or supply heat to a district heating system. .
Heat reuse systems are more efficient and economical when used in conjunction with a direct liquid cooling system. The return water temperature will generally be much higher than can be achieved in a chilled water or refrigerant-based cooling system.
An air-cooled IT equipment/chilled water heat removal system can power a heat reuse system, but it will require more equipment and energy to concentrate the heat. In either case, a plate and frame heat exchanger or heat pump will be needed to concentrate the heat before transferring it to a heating supply loop outside the data center.
In general, the amortization or return on investment of heat recovery systems is low, especially in small data centers. Innovations in data center cooling techniques will be needed to improve the economics of heat recovery. These innovations could have the ancillary benefit of reducing or even eliminating water consumption in data centers – a significant benefit for a new cooling technique.
Measures, Metrics and Goals
Data center operators should communicate metrics and set targets for the reuse and recycling of end-of-life products and components and for the percentage of end-of-life products and components sent to landfill.
To validate this, data center operators must maintain an inventory of end-of-life products sent to their recyclers or product reclaimers. Operators should also verify that recyclers and reclaimers track the final disposal of refurbished products and components.
For example, an operator sends its laptops, workstations, and servers to an organization that refurbishes and redeploys the equipment for use in the education sector in other countries.
This is a good destination for second use, but the operator should clarify where equipment that cannot be repaired or refurbished is sent, and how receiving organizations manage their own equipment in end of life. The operator must ensure that end-of-life equipment is properly managed after the end of its “second” life.
A metric and target for average computing life or refresh rate can be set. Refresh rates are usually set for reasons of reliability or availability of more power-efficient next-generation technology equipment (as measured by workload delivered per unit of power consumed). Longer refresh rates should result in lower environmental impacts.
A potential approach to increasing the average refresh time is to set refresh rates for different groups of servers based on performance and workloads, as well as the need for and benefits of new technologies.