PROJECT TITLE :

A Multi-task Oriented Framework for Mobile Computation Offloading

ABSTRACT:

In recent years, computation offloading has gained popularity due to the fact that it is an efficient way to lower the amount of energy consumed by smartphones while simultaneously improving their overall performance. We propose and implement a lightweight offloading framework that supports offloading of compute-intensive tasks and efficiently deploying the server in order to deal with the heterogeneous architectures that exist between the smartphone and the server, as well as to simplify the process of deploying the server. This allows us to simplify the deployment process. Offloading services, both generic and those that have been customized by the developer, could be provided for a variety of third-party applications by utilizing this framework. In addition, we develop a multi-task offloading strategy for the framework so that it can manage the extensive offloading requests that come from a variety of mobile devices. When an offloading request is received, the master node on the server side decides whether or not the task in question should be offloaded, as well as which virtual machine (VM) should be responsible for handling it. This ensures that both the overall execution time and the amount of energy consumed are optimized. We put this framework into action and evaluate it by analyzing the difference in running time, energy consumption, and percentage of CPU utilization between three different execution modes and three different applications. Experiments using the multi-task offloading strategy are also carried out in simulation environments by our team. The results of the experiments show that this framework is effective at lowering the amount of energy consumed and increasing the performance of compute-intensive tasks. Furthermore, the multi-task offloading strategy is suitable for intensive offloading requests.