DOLFIN aims to significantly contribute towards improving the energy efficiency of Data Centres and stabilizing of Smart Grids, a holistic approach, across networks of Data Centres and Smart Grids. Today, most DCs are part of computing and storage clouds, offering their customers Virtual Machines (VMs) as a virtual operating environment. DOLFIN will model, monitor, and measure energy consumption and enable seamless, autonomic migration of VMs between servers of the same DC or across a group of Energy-conscious, Synergetic DCs, aiming to:
- optimize the overall energy consumption by dynamically changing the percentage of active versus stand-by servers and the load per active server in a DC, and
- stabilize the Smart Grid energy distribution, under peak load and increased demand, by dynamically changing the energy consumption/ production requirements of the local DCs.
The DOLFIN follows a three levels hierarchical approach for optimizing energy consumption
Level 1- Optimization inside a Data Centre
Optimizing the energy consumption within the limits of a single DC, based on system virtualization and the optimal distribution of VMs. This will be coupled with the dynamic adaptation of active and stand-by servers and the load optimization per active server. DOLFIN will utilize a new monitoring framework to measure the energy consumption per server module/networking component and activate low-power states on devices. DOLFIN will complement these energy management mechanisms by prototyping a system for heat absorption from ICT servers, through the use of cold water heat management systems.
Level 2- Group optimization of Synergetic Data Centres
Optimizing the cumulative energy consumption in a group of DCs, based on optimal distribution of VMs across all of the servers that belong in the group of DCs using policy-based methods. DOLFIN will measure the energy consumption on the DC level and deploy policies and solutions that will achieve a decreased cumulative power consumption across the whole group of DCs.
Level 3- Optimization at Smart City/Smart Grid Stabilization
Optimizing the energy consumption at the smart city level and stabilizing the Smart Grid, based on distribution of VMs across a group of DCs, following an electricity demand-response approach. In order to stabilize the Smart City Energy consumption, DOLFIN will dynamically change the percentage of active versus standby servers per DC and optimize the load per active server. Techniques to feedback power to the electricity network will be employed, either by utilising electricity produced by in-house sources or by restoring electricity from charged backup batteries.
DOLFIN has a clear industrial focus targeting solutions with high potential impact. Backed by industrial companies such as Interoute, PSNC, Wind and Siemens, DOLFIN is targeting the wide majority of the already deployed DCs and the ones that cannot directly utilize renewable energy sources due to their location.