Abstract
The massive development of the cloud marketplace is leading to an increase in the number of the Data Centers (DCs) globally and eventually to an increase of the CO related footprint. The calculation of the impact of Virtual Machines (VMs) on the environment is a challenging task, not only due to the technical difficulties but also due to the lack of information from the energy providers. The ecological efficiency of a system captures the relationship between the performance of the system with its environmental footprint. In this paper we present a methodology for the estimation and prediction of the ecological efficiency of VMs in private cloud infrastructures. We specifically focus on the information management starting from the energy resources in a region, the energy consumption and the performance of the resources and finally the calculation of ecological efficiency of a VM. To this end, we have designed and implemented a framework through which the ecological efficiency of a running VM can be assessed and the ecological efficiency of a VM to be deployed can be forecasted. The presented framework is being evaluated through several private cloud scenarios with VM deployments in hosts located in Germany.
1. Introduction
The rapid growth of ICT application services goes along with an increase in number and size of data centers (DCs) that host these services. Because data centers are massive energy consumers, the carbon footprint of application services is moving more and more into focus. It is considered that ICT presently accounts for approximately 2% of global carbon emissions (more than 830 million tons of carbon dioxide) [1,2]. The advent of the cloud computing paradigm gave an enormous boost to the ICT services sector and the prediction towards 2020 is that the market (SaaS, PaaS, IaaS) will quadruple from what it is today.
5. Conclusions
In this paper we investigated the environmental impact of cloud resources and specifically about techniques to measure and monitor the CO2 footprint of virtual machines in cloud computing infrastructures. We identified ecological efficiency of VMs as the metric that effectively defines such impact and we modeled this metric using energy and performance related parameters of the hosting system. To this end, we proposed a service architecture which allows monitoring the eco-efficiency of VMs in private cloud scenarios. The designed system aggregates information from public energy data-stores and historical data for each country in order to define the CO2 emission factors for each region of interest. Through this methodology we are able to calculate the eco-efficiency of VMs in a location and time specific manner.
