ECO₂Clouds - Experimental Awareness of CO2 in Federated Cloud Sourcing

Ecological implications of cloud-based IT infrastructures are creating a critical gap in the current state of the art in research and business. Therefore, the ECO₂Clouds project investigated strategies that can ensure effective application deployment on the cloud infrastructure, reduce energy consumption and by association CO2 emissions and furthermore the costs of the whole execution. The need for novel deployment strategies becomes more evident when an application spans multiple clouds. Cloud providers operate under different regulatory frameworks and cost structures in relation to environmental policies and energy value-chains. In addition, optimizing the key assets like deployed virtual machines, applications and databases is constrained by a set of requirements such as quality, privacy and cross-platform service-level agreements.
ECO₂Clouds provided a challenging and innovative approach to cloud computing service delivery by:

• developed extensions and mechanisms for cloud application programming interfaces to collect eco-efficient data at the physical cloud infrastructure and virtual machine levels

• investigated the key environment and the needed quality and cost parameters in order to deploy virtual machines in multi-cloud environments

• developed evaluation mechanisms and optimization algorithms to assess different parameter configurations and their influence in energy-efficient cloud sourcing and application deployment strategies
  
  

For a proof of concept execution, the carbon-aware mechanisms have been integrated into the Future Internet Research and Experimentation (FIRE) facility BonFIRE to test, validate and optimize the eco-metrics, models and algorithms developed. Furthermore, the availability and provisioning of eco-efficient deployment strategies improved the BonFIRE service offering. Using those models and tools, the overall energy and CO₂ consumption could be decreased by 40%. As a consequence, these promising results will form the basis for future research.