Fines data centre ecosystem

All4Green

All4Green Logo

10% - 20% energy saving

All4Green building blocks

All4Green envisages energy savings in the Data Centre and related ecosystems as the synergistic effect of three new energy optimization directions, enabled by the research scope of the project:

  1. Data centres offer services to ICT Users, respecting Service Level Agreements (SLAs) typically centred on performance and availability. In the proposed approach, Green-SLAs allow data centres to relax certain constraints (in a regulated way) to achieve high energy savings with minimal impact on the performance and availability of the system. An example is defining the performance based on the effective request rate, allowing a drastic reduction of ICT resources during low load periods. All4Green Green-SLAs between data centres and ICT users will give also data centres the freedom to react to requests from the energy provider to temporarily reduce its energy consumption e.g. by temporarily putting low-priority jobs on hold or abstaining from some degrees of cooling.
  2. Data centres are high energy consumers. Sharing mid-term (days) power need plans with the respective electrical power provider will allow the provider to align energy production and usage. Moreover, the data centre, based on Power Provider requests, can contribute to short term (minutes) energy consumption reduction, to help handle power consumption peeks. The benefits for the data centre implied by this collaboration are a reduced risk of power shortages, especially useful also when the power provider is using non-constant power sources, like wind and sun; and a mechanism to reward the collaborative energy consumers with economical reduced rates and incentives (indirectly, this new economical model will allow the data centre to share the revenues with the cooperating end users, i.e. the ones accepting Green-SLA).

    Energy providers also have several benefits. They can reduce unnecessary spending, through last minute energy buying on the marketplace. They can also reduce emissions due to the possibility – for the energy provider - to lower the proportion of polluting energy sources, potentially even to zero. This in turn can reduce the need to use inefficient energy sources, and can reduce the transport of energy from production locations situated far from the consumption area, therefore avoiding the energy loss/inefficiencies of the transportation.

  3. Data centres can federate and collaborate to exchange load between them, to meet short time power reduction request from energy providers, and to collaboratively develop longer term plans that capitalize on the price and quality attributes of the energy offered by different providers. Combining these collaboration opportunities in an integrated way in one ecosystem leads to a situation where all the actors, not only the energy provider, are responsible for detecting and reacting to peak energy demands or unanticipated supplies of renewable energy. Instead, the energy provider triggers a chain reaction triggered by its request to the data centre to lower power consumption. The data centre reacts to this request by applying a set of mechanisms that temporarily reduces its energy demand, for the short term.

Some of these mechanisms focus on the temporary reduction of work load on the data centre infrastructure by delaying jobs (time shift) or relocating jobs to federated data centres (location shift). These actions require cooperation between data centres and their customers. This cooperation can be achieved by the design of Green-SLAs that create an incentive for the data centre customer to accept temporary compromises in traditional KPIs.

Other mechanisms for data centre to reduce its energy consumption are based on the principle of building up energy/cooling reserves in the data centre (by storing energy directly in a UPS, or by pro-actively reducing the temperature in the data centre beyond average levels, when a surplus of renewable energy is available), which in times of power shortage can replace the energy from the public electricity grid. Thermal “debts” can be created in the data centre by allowing the temperature to rise up to a certain extent – these debts can then be paid back with renewable energy.

Request an implementation

Reduce Energy up to 20%