Fines data centre ecosystem

All4Green

All4Green Logo

10% - 20% energy saving

Objectives

To improve the operational efficiency of an IT services ecosystem that comprises data centres (stand-alone and federations), energy producers and the ICT users All4Green will focus on the following research challenges:

  • How to shape the energy consumption of data centres in order to
    • Schedule load so as to use mid-term predictable energy overproduction and short-term energy production peaks.
    • Avoid sudden peaks of energy power consumption, requests to lower consumption to help the power provider with short term peaks.
  • How to create energy consumption profiles for data centre operations from the experience and knowledge pool so that energy providers can rely on them when integrating renewable energy sources.
  • How to design Green-SLAs between:
    • Data centres and end users.
    • Data centre and energy providers shaping energy consumption and how to make these Green-SLAs attractive to data centres and end users.
  • How the results of the aforementioned research questions change when federations of data centres enter the scene:
    • What modifications or extensions to the technical energy profiling are necessary or possible,
    • How does the system of Green-SLAs have to be adapted.
  • How will the change of paradigm that comes with the All4Green concept influence the currently emerging data centre market?
    • “Maximising efficiency under the constraint of given resources”

Together these research challenges lead to the complex task of identifying and tuning the parameters of IT services ecosystems to induce an economical and environmental friendly use of energy including the very source of energy generation. This leads to the following technical approach:

  • Identify optimal energy-related communication structures within the ecosystem (data centre, energy provider and end user).
  • Define a mechanism to share, update and adapt energy consumption and production plans between DC and energy provider that is real-time, reliable and robust/resilient/fault tolerant.
  • Create an agent-based negotiation scheme for the collaboration mechanisms between the different actors, this will need to:
    • Define an agent framework allowing actors to interoperate.
    • Define robust coordination schemes for real-time job scheduling, load distribution and resource allocation across federated DCs.
    • Define agent-based interactions schemes for increasing the flexibility of Green-SLAs.
  • Develop energy profiles for energy-aware data centre operation and define a set of DC energy related enforcement rules using the energy aware information and monitoring system
  • Design a system of vertically synchronized Green-SLAs with resources constraints between
    • ICT end-users and DC
    • DC and energy providers
  • Define a clear mapping of Green-SLA to Technical SLA:
    • Monitored KPIs, applicable control rules, ranges of energy related Quality of Service (QoS) indicators
    • Prices related to energy related QoS, incentives and penalties
  • Evaluate the effectiveness of the deployed technical solution first through simulations and then through pilots at trial sites.

Request an implementation

Reduce Energy up to 20%