Online strategies for dynamic power management in systems with multiple power-saving states

Abstract

Online dynamic power management (DPM) strategies refer to strategies that attempt to make power-mode-related decisions based on information available at runtime. In making such decisions, these strategies do not depend upon information of future behavior of the system, or any a priori knowledge of the input characteristics. In this paper, we present online strategies, and evaluate them based on a measure called the competitive ratio that enables a quantitative analysis of the performance of online strategies. All earlier approaches (online or predictive) have been limited to systems with two power-saving states (e.g., idle and shutdown). The only earlier approaches that handled multiple power-saving states were based on stochastic optimization. This paper provides a theoretical basis for the analysis of DPM strategies for systems with multiple power-down states, without resorting to such complex approaches. We show how a relatively simple "online learning" scheme can be used to improve the competitive ratio over deterministic strategies using the notion of "probability-based" online DPM strategies. Experimental results show that the algorithm presented here attains the best competitive ratio in comparison with other known predictive DPM algorithms. The other algorithms that come close to matching its performance in power suffer at least an additional 40% wake-up latency on average. Meanwhile, the algorithms that have comparable latency to our methods use at least 25% more power on average.

Networking Strategies for Structural Health Monitoring in Wireless Sensor Networks

“…The selection of the threshold has to be done with a good understanding of both node's hardware characteristics and the typical request patterns. Although many implementations have a single fixed threshold, a number of studies have designed methods for adapting the threshold to changes in the requests with machine learning techniques or the distribution of idle periods [15,16].…”

Sleep with Guilt and Work Faster to Minimize Flow Plus Energy

“…The input is the length of the period, known only when the period ends. There are several interesting results with competitive analysis (e.g., [2,11,13]). In its simplest form, the problem assumes the processor is in either the awake state or the sleep state.…”

Stochastic learning feedback hybrid automata for dynamic power management in embedded systems

“…This section of the research was inspired by the research detailed in [1]. Here, several former predictive algorithms are analyzed to be compared to the SLHA model for Dynamic Power Management.…”

“…This deterministic algorithm is presented in detail in [1]. Its strategy uses the approach of the optimal offline algorithm to determine at each time t the optimal state for the elapsed idle period.…”

“…The OPBA is the main approach investigated in [1]. This method is adapted from the Probability-Based Lower-Envelope Algorithm, which is a strategy based on the Lower Envelope Algorithm explained earlier.…”