Flow Network-Based Real-Time Scheduling for Reducing Static Energy Consumption on Multiprocessors

Abstract

The energy management for embedded real-time systems is crucial due to their restricted power supplies. With the advancement of technologies, the static energy consumption of the embedded systems that is caused by their leakage power is growing. Thus, a number of research works have started focusing on reducing the static energy consumption by making the systems transit into low-power states, wherein some hardware components are temporarily shut down. Specifically, when a processor is idling, they attempt to set the processor into one of several low-power states. To make a processor remain in the low-power state as long as possible to minimize the energy consumption, the idle time should be maximally clustered. At the same time, in order to satisfy the real-time constraints of the tasks, the length of the clustered idle time should be estimated accurately. To achieve our objective, we propose energy-efficient real-time scheduling algorithms on symmetric homogeneous multiprocessors with a dynamic power management scheme for periodic real-time tasks. The proposed algorithms rely on a flow network model that effectively helps to cluster the idle time while respecting the real-time constraints. In our experimental evaluation, the proposed algorithms consume a comparable static energy to an existing off-line scheme that is the only suitable existing algorithm in the problem domain. Furthermore, we show that the proposed algorithms consume less static energy than the existing one in a case where the total workload of the given task set is low and the task completion is earlier than expected.INDEX TERMS Dynamic power management, energy-aware algorithm, flow network problem, multiprocessor unit, micro-controller unit, real-time scheduling.

A Lightweight Optimal Scheduling Algorithm for Energy-Efficient and Real-Time Cloud Services

“…At every boundary, the scheduling algorithm is invoked to reserve the execution time for all active jobs. Here, we formulate an optimization problem with a combined flow network, as in [12]. For convenience of description, we define four sets at the current time interval as follows:…”

“…Specifically, the solver algorithm clusters the idle time close to the end of either the last boundary (b k ) or the current time (b 0 ) when the current mode is CB or CF, respectively. On the other hand, the previous algorithm in [12] assigned a parameter cost based on the current mode to each edge to prioritize a certain flow over the flow network, and its feasible solution was found by using the min-cost-max-flow solver algorithm. However, because of the high complexity of the min-cost-max-flow solver, it is difficult to dynamically schedule real-time tasks during system runtime.…”

“…Thus, we apply a method that does not use the cost to the proposed solver algorithm to find a feasible solution close to the solution from the min-cost-max-flow problem by the previous algorithm in [12]. In the combined flow network with idle time shown in the previous subsection, a feasible solution can be found in any order for all edges, so there is no need to control the flow using cost.…”

“…Thus, ensuring the optimality of real-time scheduling, which can maximize the degree of SLAs, is a valuable contribution to cloud services. In the context of this paper, we define the optimality of the real-time schedule as the same as in [12] for ease of understanding.…”

“…In this study, we significantly reduced the computational complexity of the scheduling algorithm compared with the latest works [12]. Because the complexity of the previous algorithm was Ω(N 3 logN ), where N is the number of realtime tasks, it is difficult to dynamically schedule real-time tasks during system runtime.…”