Journal-Archives-Electronic Sciences & Systems


TITLE :	POWER OPTIMIZATION IN WIRELESS NETWORKS USING DYNAMIC VOLTAGE SCALING ALGORITHMS
AUTHORS :	T Gavaskar Vincent1, Sasipraba.2
ABSTRACT :	Execution of tasks to be completed by specified deadline is paramount important for any scheduling algorithms. Most DVS algorithm doesn't consider real-time constraints and are based on solely average computational throughput. Since they use a simple feedback mechanism such as detecting the amount of idle time on the processor over a period of time and adjust the computational load, they cannot provide any timeliness guarantees and task may miss their execution deadlines. To alleviate the difficulty, we propose a simple mechanism for providing voltage scaling while maintaining real-time schedulability. In this mechanism we select the lowest possible operating frequency that will allow the RM or EDF scheduler to meet all the deadlines for a given task set. This frequency is set statistically and will not be changed unless the task set is changed, designing of SimDVS, a unified simulation environment for evaluating dynamic voltage scaling (DVS) algorithms. At present it evaluates two static RT-DVS algorithms and three dynamic RT-DVS algorithms. All RT-DVS algorithm's task execution module differs with each other. The SimDVS has been designed such a way that any new DVS algorithm can be evaluated easily by adding new task execution module to SimDVS. It also supports to add new machines specifications to in it.
	Download Full Paper