Predictive strategies for low-power RTOS scheduling

Limiting the power consumption of real time embedded systems is an important aspect, especially in portable systems (laptops, cellular phones) with tight power constraints. In this paper, we present a power-saving prediction strategy that exploits the fixed priority scheduling of the real-time tasks running on these embedded systems. Power reduction is achieved by developing an efficient low power scheme with prediction of the expected execution time of real time tasks and making use of the idle time of system for scheduling these tasks in low power modes. In the process there may be few tasks missing their deadlines. This results in a tradeoff between power saved and deadlines missed. Our simulation results for different applications show that the proposed prediction mechanism achieves a high degree of power conservation with a very small penalty of missed deadlines. Our mechanism is simple and can be implemented in most of the real time operating systems.