Abstract—The paper presents the state-of-the-art study on the recently published literatures subjected to power aware design in various industrial applications. The basic conception of power-aware design is placed at first and then the recent progress is investigated. For each paper, a brief summary is given to introduce the related power awareness design technologies and applications. Index Terms— power awareness, embedded system, energy management I. INTRODUCTION With modern electronic products becoming faster, smaller, and cheaper in the past half century, the power management emerges as a focal point in many commercial systems. In high performance systems, power-aware design techniques aim to maximize performance under power dissipation and power consumption constraints such as in portable equipment. In addition, a power-aware design can also help to reduce the energy cost. P. Ranganathan, a distinguished technologist in the HP lab, wrote an article about the concept, current research point, and future work in Power aware computing [4] recently. It is obvious that there is a minimum amount of energy needed to perform certain tasks and a corresponding minimum amount of heat extracted to avoid thermal problems. However, designs often introduce additional inefficiencies above the actual energy. These inefficiencies are often introduced when the system design must reconcile complex trade-offs that are difficult to avoid. They are often designed for the most aggressive work load performance and worst-case risk tolerance. Such designs can lead to resource over provisioning to better handle transient peaks and offer redundancy in the case of failure. Further, sometimes the traditional design focused on the small performance improvement by the cost of power and heat extraction. How to reduce these inefficiencies is called power aware strategies. Power aware strategies can be activated either in hardware or in software. To use the more power efficient component in the design is the most direct and efficient way. Another efficient method is to create energy proportionality by scaling down energy for unused resources, which requires algorithms respond to the consequences of turning down a system. Other than having the resources adapt when not fully utilized for a given task, one can also match tasks to the resources most appropriate to the size of the task. In cluster or multi-core environments, the designs might need to explicitly introduce multiple operation modes with different power-performance trade-offs. Overlap energy events seek to combine multiple tasks into a single energy event. Decomposing system functionality into smaller subtasks can help increase the benefits from an overlapping energy event by avoiding duplication of energy consumption for similar subtasks across different larger tasks. The other less common methods include broadening the scope of the solution space, trading off other metric for energy, and spending power to save power. In this paper, the basic conception of power aware design is placed and then the recent published literature about the power aware computing in various system architectures such as computing clusters, battery power systems, hand hold devices, distributed real time systems as well as wireless sensor network systems [1-3] will be investigated. II. RECENT PROGRESS In this section recent developments in power aware systems will be presented which cover energy, power and thermal management approaches in various system architectures such as computing clusters, battery power systems, hand-hold devices as well as distributed real-time systems. A. Computing Cluster Clusters of commodity class PCs are widely used. To reduce power, cluster power management mechanisms have been proposed in both homogenous and heterogeneous systems. There are more challenges in a heterogeneous system. For a given workload of a heterogeneous cluster, the front-end PM needs to decide how many and which back-end servers should be turned on and how much workload should be distributed to each server. Since the request rate of cluster changes from time to time, these decisions have to be reevaluated and modified regularly. Heath et al. [5] argued that designing efficient servers for heterogeneous clusters requires defining an efficiency metric, modeling the different types of nodes with respect to the metric, and searching for request distributions that optimize the metric. This approach is time consuming and is not designed for real-time clusters. In paper [6], the author presents a threshold based method for efficient power management (PM) of heterogeneous soft real time clusters. The mechanism of the algorithm is built on a sophisticated offline analysis and provides an efficient threshold-based online strategy. The offline analysis divided Recent Advances in Power Aware Design Tiantian Xie, Bogdan Wilamowski Department of Electrical and Computer Engineering, Auburn University, Auburn, USA Email: {tzx0004, wilambm}@auburn.edu 978-1-61284-972-0/11/$26.00 ©2011 IEEE 4632the maximum request rate of workload range into small ranges using many thresholds. The power is controlled and the workload distribution decision is made for each range offline. The system just needs to measure the actual request rate of the cluster and decide which small range it falls into online, and then the corresponding PM is decided. In this approach, a PM algorithm makes three design decisions on an ordered server list, server activation thresholds and workload distribution. To decide the server list, the servers are ordered by their power consumption efficiency. The PM mechanism usually turns on a server when needed or when it leads to reduced power consumption to make sure an active server usually works under a high workload. To decide the server activation thresholds, the author combines capacity and optimal-power thresholds. At last the algorithm proposes a strategy to optimally distribute the workload among active servers. The result shows that it incurs low overhead and leads to optimal power consumption. B. Battery Power System Many modern consumer electronics use batteries as a power supply[7-8]. The batteries have the finite capacity which determines the time during the device can be used. Several batteries are usually provided in sequential order to extend the usability of such device, but it is not the efficient way. People found that using multiple batteries in a