Design and Implementation of Hybrid Cloud Device-as-a-Service

Traditional Cloud Device-as-a-Service (DaaS) model for providing Virtual Desktop Infrastructure (VDI) and Virtual Mobile Infrastructure (VMI) services are being provided using each independent systems because of difference of its system features. To provide VDI and VMI service simultaneously, a clouds service providers have to build each server infrastructures that cost twice. In this paper, we propose a cost effective heterogeneous DaaS structure capable of offering VDI and VMI services simultaneously on a single cloud platform. Additionally, it presents a technique to distribute CPU loads, used for rendering and encoding, by leveraging low-cost general-purpose GPUs on cloud servers that will enable a host machine to function more efficiently by operating more guest machines, owing to elimination of CPU loads. In this paper, we introduce in detail our proposed scheme and discuss its originality; moreover, we present performance estimation results.