COSMIC project

The COSMIC project is an Inria exploratory action on Coordinated Optimization of SMart grIds and Clouds (2016-2018).

The distributed nature of Cloud infrastructures involves that their components are spread across wide areas, interconnected through different networks, and powered by diverse energy sources and providers, making overall energy monitoring and optimization challenging. This project takes advantage of the opportunity brought by the Smart Grids to exploit renewable energy availability and to optimize energy management in distributed Clouds.

Results

Context

The energy drawn by Cloud data centers is reaching worrying levels, thus inciting providers to install on-site green energy producers, such as photovoltaic panels. Considering distributed Clouds, workload managers need to geographically allocate virtual machines according to the green production in order not to waste energy.

In this project, we consider a distributed Cloud infrastructure comprising several data centers geographically distributed and powered by the regular electrical grid on one side and on-site photovoltaic panels on the other side. The user management of the Cloud is assumed to be centralized as shown on the figure below.

The considered Cloud architecture
Considered Cloud system

SAGITTA: a Stochastic Approach for Green consumption In disTributed daTA centers

As a first approach, we consider the problem of scheduling workload across multiple data centers for minimizing renewable energy loss. To solve this issue, we propose SAGITTA: a Stochastic Approach for Green consumption In disTributed daTA centers. SAGITTA uses a stochastic approach for estimating renewable energy production, and a greedy heuristic for allocating resources to the incoming user requests. Our simulation-based results show the efficiency of SAGITTA compared to classical allocation approaches. Indeed, compared to the optimal solution, SAGITTA consumes 4% more energy overall, and wastes only 3.14% of the available green energy, while a classical round-robin solution consumes 14.4% more energy overall than optimum, and wastes 28.83% of the available green energy. More details can be found in [CDO17].

Co-simulation of FMUs and Distributed Applications with SimGrid

SimGrid is a versatile platform for the simulation of distributed systems which embeds a set of rigorously validated IT models (e.g. CPU, IP network, disk, energy consumption). The platform manages the co-evolution and the interaction between these models and the different concurrent processes that compose a distributed application code. Since 2010, the Functional Mock-up Interface (FMI) standard of the Modelica Association is becoming an essential solution toward co-simulation. It offers a unified framework and an API to control equation-based models of multi-physical systems (e.g. electrical, mechanical, thermal systems). The FMI-compliant tools enable (1) to design a model and export it as an FMU (Functional Mock-up Unit) –i.e. a simulation unit compliant with FMI– (2) and/or to import and use an FMU as a component in their modeling environment. In this work, we specify an FMU as a SimGrid model to ease the simulation of distributed data centers including their Computer Room Air Handler (CRAH). More details can be found in [COQ18].

Publications

People