Enhanced reliability-aware power management through shared recovery technique

Abstract

While Dynamic Voltage Scaling (DVS) remains as a popular energy management technique for real-time embedded applications, recent research has identified significant and negative impact of voltage scaling on system reliability. For this reason, a number of reliability-aware power management (RA-PM) schemes were recently proposed to preserve the system reliability when DVS is used. In this paper, we propose a new approach, called the shared recovery (SHR) technique, to minimize the system-level energy consumption while still preserving the system's original reliability. The main idea of the SHR technique is to avoid the offline allocation of separate recovery tasks to the scaled tasks by assigning a global/shared recovery block that can be used by any task at run-time. Our simulation results show that, compared to the existing RA-PM schemes, our scheme can achieve up to 35% energy savings. Further, this performance is shown to be comparable to the maximum energy savings that can be achieved by any algorithm. Interestingly, our extensive evaluation indicates that SHR offers also non-trivial gains over the previous algorithms on the reliability side. Further, a dynamic extension is proposed to improve energy and reliability management at run-time by reducing the size of the recovery block and re-using the slack that arises from early completions.

Low-Energy Standby-Sparing for Hard Real-Time Systems

“…We focus only on transient faults as it has been observed that transient faults are the major source of concern in safetycritical applications [27][29] [31]. In this paper, we consider frame-based real-time applications [29][31] with hard timing constraints where a set of dependent tasks, e.g., Γ={T 1 , T 2 , T 3 , …, T n }, is executed within each frame and must be completed before the end of the frame (deadline) [31]. The LESS system does not need any dedicated scheduler.…”

“…Note that, as discussed in Section II, here we focus only on transient faults. The occurrence of transient faults in digital systems usually follows a Poisson process [6][7] [29]. It has been observed that in DVS-enabled systems the rate of transient faults increases exponentially as the supply voltage decreases, so that the fault rate can be expressed as [32]: …”

“…To consider these five issues, we revised the analytical energy models and the energy management algorithm. 2) By conducting new simulation experiments, we have compared the LESS system (with the revised energy management algorithm) with a recent low-energy timeredundancy system (presented in [29]) with respect to the energy consumption and reliability. 3) We have developed an analytical reliability model for the LESS system.…”

Generalized Standby-Sparing techniques for energy-efficient fault tolerance in multiprocessor real-time systems

“…The recovery job is executed at the maximum speed only when its scaled primary job incurs transient faults. The shared-recovery based RAPM scheme has also been studied where several tasks with scaled frequencies may share one recovery task for reliability preservation [22]. More recently, considering such negative effects of DVS on system reliability, the recovery allowance based scheme that can achieve arbitrary reliability levels for each periodic task is further studied [23].…”

Towards fault-tolerant embedded systems with imperfect fault detection

“…Perfect fail-silent behavior is one assumption that is often used in literature. It is assumed that all faults are detected within a certain time interval and the fault-detection overhead is contained in the tasks' WorstCase Execution Times (WCETs), e.g., in fault-tolerant task scheduling [8,15,4,17,6,5], in reliability-aware energy management [16,20,22] and in error-aware system design [9,10]. With this assumption, each task will produce either a correct output or no output at all.…”