Fines data centre ecosystem

All4Green

All4Green Logo

10% - 20% energy saving

All4Green ecosystem tailoring and Simulation tool

The All4Green ecosystem tailoring tasks will specify, design, implement and tailor the All4Green ecosystem. This system shall consist of agents using the models developed in other research topics and shall be able to interoperate with legacy systems used by end-users, data centres and energy producers in the pilots. This way the All4Green ecosystem shall be capable of handling the diverse relations amongst a federated data centre, consisting of multiple data centres, their ICT end-users, data centres, and the energy producers. In first phase as detailed as possible federated data centre management models are provided. In subsequent phases, they are extended or enhanced as they are more and better grounded after validation and evaluation in the trials activities.

The first activity will be the selection of an existing agent-development framework for the definition and development of agents. Once selected, the implementation of the agent development framework will set up various agent facilities, integrate them and put them alive for validation. The following facilities will be put in place:

  • Web-Service Gateway for agents to publish and subscribe for data, capabilities or services
  • Graphical User Interface to start and stop agents
  • Distributed database to collect, process and retrieve agent state and history and for thread management
  • Agent APIs for the connecting to the different All4Green sub-ecosystem entities
  • Agent Container APIs
  • Agent communication, composition and workflow management schemes compatible with All4Green sub-ecosystem schemes presently operational
  • Support for distributed deployment

Using the agent framework previously set up, this task implements all All4Green sub-ecosystem agents using the model elements, Green-SLA policies, energy demand and supply data exchange mechanisms and negotiation mechanisms.

  • For each ICT end-user a dedicated user-agent is implemented representing and acting on behalf of this end-user. It forms an interface to the end-user to ratify and endorse a new Green-SLA accepted by the end-user and also to provide him insight to which extent the data centre is meeting this SLA. For each data centre is also implemented a first type of data centre agent looking after its concerns related to its ICT end-users, the ICT end-user specific data centre agent.
  • For each energy provider a dedicated energy provider agent is implemented representing and acting on behalf of this energy provider. It forms an interface to the energy provider to communicate to the data centre agent its planned, actual and forecasted energy supply for the upcoming periods. For each data centre is also implemented a second type of data centre agent looking after its concerns related to its energy provider, the energy provider specific data centre agent.
  • A federated data centre is an aggregation and cooperation of multiple data centres. For each data centre of the federation a third type of data centre agent is implemented, namely the data centre management agent. Depending on the legal type of federated data centre a mediator or supervisor federated data centre management agent could be implemented. If the federated data centre is owned by one stakeholder, then supervisor type seems to be more appropriate, whereas if it is owned by multiple stakeholders being only partially cooperative the mediator type seems more applicable. In All4Green project this distinction is resolved by implementing a fourth type of data centre agent, namely a data centre specific negotiating agent.

Finally, this task integrates all the previous described building blocks into the All4Green plug-in and ensures its interoperability with the existing test beds data centre management systems, end-user environments and energy producer legacy systems.

In parallel, All4Green will provide its Simulation Tool: The development of the building blocks of All4Green solution (the All4Green "plug-in") requires a suitable execution environment able to simulate all the ecosystem interactions with the data centre. Therefore these activities will design and develop a software tool to interact with the plug-in and simulating the external interactions with the other actors inside the ecosystem. Through a simple Web interfaces, this tool will allow observations of the behaviour and interactions (both automatic and through manual operations) with the plug-in software by representing:

  • data centre end-users:
    • to manage Green-SLA (and contracts) parameters
    • to monitor system behaviour and the KPIs related to Green-SLAs
    • to check service pricing and received incentives for the collaboration
  • power providers:
    • to define QoS for the power service
    • to quickly negotiate power reduction requests to the data centre
    • to report power pricing and incentives
  • data centre operators and managers:
    • to monitor KPIs and Green-SLAs to end-users
    • to check pricing and incentives
    • to support negotiation activities

The agent framework will interface this simulator and it will be possible to create an effective test suite for the solution before the integration with the real trial environments will be implemented. Moreover, the simulation tool easily allows performing demos of the All4Green solution, even in facilities and at events where the connection to trial sites is not available.

Request an implementation

Reduce Energy up to 20%