As the need for cloud services has been growing steadily, the size and energy consumption of datacenters have increased significantly over the past years. Due to economic and environmental constraints, energy efficiency in datacenters and greenhouse emissions have become a major concern. Renewable energy is widely seen as a promising solution to supply datacenters using local energy, without greenhouse gas emissions. However, the intermittent power generation resulting from the use of renewable energy imposes a paradigm change in the way energy and computation activities are managed. On the one hand, service placement and scheduling may be used on the IT (information technologies) side to adapt to the available power. On the other hand, the storage units may be used to lessen power generation variations. Existing literature and actual deployment mainly design optimization algorithms including the entire system (from cloud service to electrical management, the latter often being neglected or simplified). Conversely to these approaches, we propose a solution where each side optimizes its own objectives, both interacting through a negotiation loop process to reach a common agreement. In this paper, we present DATAZERO, a project developing this idea to ensure high availability of IT services, avoiding unnecessary redundancies, under the constraints due to the intermittent nature of electrical and cloud services flows.
