Multi-tier Elastic Computation Framework for Mobile Cloud Computing

Federating the portability and mobility of mobile devices with the computation capacity on desktop computers have been a widely discussed computation model for the next decade. However, the mobility of the mobile devices also introduces challenges on the federation. This work developed the elastic computation framework to tackle the aforementioned challenge. The elastic computation framework federates the computation resources on wearable devices, mobile devices, nearby computers, and remote computers into a pool of computation resources. Each resource in the pool is characterized by its network delay, expected response time, and computation capability. Each mobile device is also characterized by its mobility and computation workload requirements. The elastic computation framework assigns computation resources in the pool to meet the workload requirements on mobile devices. The framework consists of resource allocation component, task scheduling algorithm, and task dispatch middleware. In the experiment, we compare the developed scheduling algorithm with other known algorithms by simulation. The results show that the developed scheduling algorithm does not complete the most tasks though, the cost/performance of system resources is the best among all the algorithms. To be specific, the cost-performance of the developed algorithm can be at least two times better than that of compared algorithms.