Coordinating Green Clouds as Data-Intensive Computing

Abstract

Cloud providers seek to maximize their market share, so they deploy datacenters to (1)Build a capacity that will accommodate the entire compute demand. (2)Maintain geographical affinity to its customers which is vital for low-latency applications e.g. live streaming services. Achieving these goals by a single provider is unrealistic from the cost of ownership perspectives. Moreover, with the carbon emissions levels data centers place an increasing demand on grid electricity and emit carbon. We suggest a solution that will allow cloud providers to operate near their customers and reduce its operational costs by lowering the data-centers deployments per provider ratio. We leverage previous studies as a starting point that focused on the data-center levels within a cloud provider. Our solution zooms out and optimizes cross-data-centers deployment and beyond, by suggesting a cross-cloud providers collaboration. Also, we suggest operating datacenters using both green as well as brown energy sources as a fast-data system. Finally, we show how cross-cloud coordinator service can reduce the projected data-centers carbon emissions growth by 22% from the current expected trajectory.

Topics

    3 Figures and Tables

    Download Full PDF Version (Non-Commercial Use)