ECOGreen is a system that can anticipate resource requirements, having characterized current applications, and will attribute a set of jobs to a particular data center, balancing all the options that are available in the energy market.

There is a template that allows the modelling of a data center, in terms of its network behavior, computer behavior, cooling and other factors. Then there is a model for the energy supply system of each country and even for geo-distributed data centers, so that tasks can be assigned to different data centers throughout the world.

A manager can then say, ‘As far as is possible, we will configure our system to use data centers where a maximum of renewable energy is available. Then we’ll set it to migrate tasks to other data centers as the situation develops.’

In this way we can let them match the high level of their expectations to the potential performance of the data centers.

This highly developed knowledge-based system is therefore tailored to the needs of each data center, providing suggestions that take into account many different factors: the cost, quality and origin of various potential sources of electrical energy, power loss in battery banks due to aging and charging sequences, quality of service constraints, while monitoring the situation in real time. Lab tests have predicted savings of up to 71 % in electricity cost, compared to the state-of-the-art.

Related Publications

ECOGreen: Electricity Cost Optimization for Green Datacenters in Emerging Power Markets
Pahlevan, Ali; Zapater Sancho, Marina; Coskun, Ayse K.; Atienza Alonso, David
2021IEEE Transactions on Sustainable Computing (T-SUSC)Publication funded by Compusapien (Next-gen computing systems inspired by the human brain)Publication funded by RECIPE H2020 (REliable power and time-ConstraInts-aware Predictive management of heterogeneous Exascale systems)