Energy-aware semantic modeling in large scale infrastructures

Including the energy profile of the computing infrastructure in the decision process for scheduling computing tasks and allocating resources is essential to improve the system energy efficiency. However, the lack of an effective model of the infrastructure energy information makes it difficult for the resource managers to provide up-to-date energy metrics of underlying environment. We modeled the relationship between energy components and infrastructure components and defined energy description elements and energy metrics using semantic web technologies, and developed a description model called Energy Description Language (EDL).
Based on the EDL, we prototyped an basic knowledge system called Energy Knowledge Base (EKB) that provides contextbased information retrieval support for applications, in order to make energy aware decisions on scheduling and resource allocation. We validated the logic and effectiveness of the energy ontology by testing the performance of the EKB.