Altocumulus: Harvesting Computational Resources from Devices at the Edge of the Cloud

Jung, YoungHoon; Szczodrak, Marcin K.; Neill, Richard W.; Carloni, Luca

The utilization of smart devices located at the edge of the cloud networks continues to accelerate, leading to new forms of heterogeneous cloud computing models. This trend has been influenced by two factors. First, the number of intelligent embedded devices in the cloud, such as smartphones and set-top boxes, has significantly increased. Second, cloud application executions like data processing and analytics are becoming more delegated from the clouds to the edge, thus making the transactions and communications in the cloud more local. We call Altocumulus a new class of heterogeneous cloud computing which occurs at the edge of the cloud where it harvests the resources of edge devices, typically embedded systems, to enhance application capabilities and performance. By reporting on a number of prior studies, we claim that Altocumulus is well positioned to support all main cloud-computing service models: IaaS, PaaS, and SaaS. We present our vision on how Altocumulus could develop in the context of other cloud computing technologies by discussing different models of computation delegation. We conclude by presenting possible use-case scenarios and open areas of research.



  • thumnail for jung_altocumulus_edge.pdf jung_altocumulus_edge.pdf application/pdf 1.04 MB Download File

More About This Work

Academic Units
Computer Science
Department of Computer Science, Columbia University
Columbia University Computer Science Technical Reports, CUCS-015-13
Published Here
September 19, 2013