After two and a half years, DC4Cities has come to its end demonstrating that its application to existing and new data centres increases the use of renewable energy to up to 60%. This energy optimization, as research within DC4Cities supports, neither requires any modification to the logistics, nor negatively impacts the quality of services.
Started in September 2013 to promote the role of datacentres (DCs) as eco-friendly key players within smart cities, this European research project matches DC power demand with the available renewable energy. This setting is in compliance with a scenario in which an Energy Management Authority (EMA) of a smart city sets a minimum target of renewable energy that can be up to 80%.
To this end, DC4Cities relies on two subsystems: one in charge of providing and continuosly updating information related to energy availability, weather forecast and locally produced power data; and another to gather information about the software running inside the DC while providing directions on its expected energy behaviour.
DC4Cities schedules its workload according to local renewable energy availability and the needs of the electric grid so that it coincides with the time of day when renewable energy power consumption can be maximized. In such a way, system efficiency is optimized, and the use of renewables can increase significantly.
In addition to optimizing the power demand of a single DC, DC4Cities allows for cooperation in a geographical DC federation. By exploiting different renewable sources, DC specialization or location benefits, federation is able to push DC eco-friendliness further than individual optimization can achieve.
Given that DCs remain large energy consumers, the application of DC4Cities can contribute to managing city-scale grid stability and efficiency optimization.
The evaluation of DC4Cities was carried out in three trial locations based on different scenarios and settings including geographical DC federation. In Trento, results demonstrated the ability of DC4Cities to adapt the power demand of an application to the energy conditions. In Barcelona, the capability of DC4Cities to exploit high differences in renewable energy availability between day and night was shown: improvements next to 25% in renewable usage were achieved. Finally, concerning the HP Experiment in Milan, a dual power source approach (PVs and grid) was used. Power flexibility was achieved by focusing on infrastructure usage optimization, showing very significant improvements over the baseline.
The next step for DC4Cities is to distribute and apply this technical solution to real-world data centres.