Escalonamento de tarefas em grid no cenário de TV digital interativa

Abstract

The popularization of mobile devices and advances in the development wireless technologies and Grid Computing, has motivate the use of these devices for distributed computing. Grids appear as a low cost solution to provide computing power. However, the task scheduling in grid presents challenges related to the dynamic and heterogeneous characteristics of the application scenarios of Grid computing paradigm. Infrastructure grid with mobile devices interoperation such as resource providers, scheduling is more difficult, due to the wireless communication and consumption of energy stored in batteries of these devices. This document presents the task scheduling algorithm for service-oriented grid, with interoperation of mobile devices such as resource providers, in home area networks in the context of digital TV. The algorithm, called PUTS (Power-aware User-preference Task-size based Scheduling), performs balancing of the energy levels of the batteries of mobile devices, in order to keep them longer in the grid, and allows the user-preference settings that influence assignment of tasks, which depending on the number of devices on the grid, reduces the number of tasks on the device. Two case studies were conducted and the results demonstrate the effectiveness of the algorithm and performance impact due to the configuration of user preferences.