S. Agrawal, R.S. Yadav, and R. Vijay
Dynamic power down, dynamic voltage scaling (m, k) model,scheduling, weakly hard real-time system
This paper presents a system energy-aware scheduling algorithm for weakly hard real time systems modelled with (m, k) constraint. The system consists of a DVS processor (frequency dependent) and peripheral devices (frequency independent) components. The energy minimization is done in two phases. The first phase uses a static partitioning strategies followed by a greedy-based speed assignment at task level. A theorem is derived to ensure the correctness of the modified partitioning strategy. The second phase performs speed fine tuning to refine the feasible schedule received from first phase by considering each job individually. The simulations for synthesized as well as real-world applications illustrate that our approach can reduce approximately 27% system energy consumption (especially for systems with higher utilizations) while guaranteeing the (m, k) at the same time.
Important Links:
Go Back
