Resource orchestration for multi-task application in home-to-home cloud

Abstract

Along with the increasing capability of consumer electronic devices, it becomes the new trend that the home cloud computing takes household consumer electronic devices as a sharing resource pool, which can bring great convenience to consumers. However, how to find the most suitable resources or services in the home cloud is a challenge for the consumers. When offloading a complicated application to the home cloud, each task of the application can be executed individually with its own computation, storage and bandwidth requirements. The heterogeneous consumer electronic devices that contain different performance metrics affect the destination choice. This paper proposes a distributed home-to-home cloud infrastructure where volunteer users allow the resources of their consumer electronic devices to be shared with the neighboring houses. By extending the home gateways, a decentralized controlling is responsible for resource monitoring, orchestration, task offload and migration. With consideration of load balance and volunteers' opinion, the resource orchestration is formulated as an optimization problem with one objective and multi-constraints. Finally, a Particle Swarm Optimization (PSO)-based algorithm is used to obtain the approximate optimal solution. Simulation results show that the solution for all cases studied almost achieves 90% objective value in acceptable time1.