Fog paradigm for local energy management systems

Abstract

Cloud Computing infrastructures have been extensively deployed to support energy computation within built environments. This has ranged from predicting potential energy demand for a building (or a group of buildings), undertaking heat profile/energy distribution simulations, to understanding the impact of climate and weather on building operation. Cloud computing usage in these scenarios have benefited from resource elasticity, where the number and types of resources can change based on the complexity of the simulation being considered. While there are numerous advantages of using a cloud based energy management system, there are also significant limitations. For instance, many such systems assume that the data has been pre-staged at a cloud platform prior to simulation, and do not take account of data transfer times from the building to the simulation platform. The need for supporting computation at edge resources, which can be hosted within the building itself or shared within a building complex, has become important over recent year. Additionally, network connectivity between the sensing infrastructure within a built environment and a data centre where analysis is to be carried out can be intermittent or may fail. There is therefore also a need to better understand how computation/analysis can be carried out closer to the data capture site to complement analysis that would be undertaken at the data centre. We describe how the Fog computing paradigm can be used to support some of these requirements, extending the capability of a data centre to support energy simulation within built environments.