| Summary: | 碩士 === 國立屏東大學 === 資訊科學系碩士班 === 107 === Recently, the Internet of Things has been developing rapidly. Mobile devices, which are dealing with complicated tasks (such as speech recognition and human detection and counting) need higher computational ability. Mobile devices with limited computing and battery power are not enough to cope with the resource-hungry computing tasks mentioned above. Therefore, tasks should be offloaded to the cloud and the cloud will return the result to the user. Then the cloud system will bearing more computation load. In addition, since the cloud system is far away from users’ mobile devices, it would cause a higher network latency and cost increased. Therefore, Fog Computing has risen in recent years. Fog computing deployed many micro cloud servers in the proximity of users’ mobile devices, forming what are known as Fog Nodes. There are now a great many Fog Nodes for mobile device to choose from for conducting task offloading. This dissertation proposes a hybrid algorithm to synthesize time delay, energy consumption, offloading costs, and server class, to be the important parameters for task offloading decisions. Simulation results demonstrate that the proposed hybrid algorithm is better than others in average delay, energy consumption and cost.
|
|---|
