Hybrid computation offloading for smart home automation in mobile cloud computing

Smart home automation enables the users to realize the access control of the in-home appliances by the mobile devices. With the rapid development of mobile cloud computing, offloading computation workloads of the home automation applications to nearby cloudlets has been treated as a promising approach to overcoming inherent flaws of portable devices, such as low battery capacity. The computing capacity of cloudlet is limited compared with the distant public cloud whose elastic computation resources are almost infinite. Therefore, some mobile services should wait for the occupied computation resources in the cloudlet to get released, which is less energy-efficient. In view of this challenge, we model the waiting time spending in the cloudlet as a M/M/m/infinity queue and propose a hybrid computation offloading algorithm for home automation applications to minimize the total energy consumption of the mobile devices within a given constant deadline. The proposed algorithm combines cloudlet with public clouds, providing a more energy-efficient offloading strategy for home automation applications. Technically, a particle swarm optimization (PSO)-based heuristic algorithm is implemented to schedule mobile services. Comprehensive experiments are conducted to demonstrate the effectiveness and efficiency of our proposed algorithm.