Sleep-aware mode assignment in wireless embedded systems

Abstract

Minimizing energy consumption is a key issue in designing wireless embedded systems. While a lot of work has been done to manage energy consumption on single processor real-time systems, little work has been done in network-wide energy consumption management for real-time tasks. Existing work on network-wide energy minimization assumes that the underlying network is always connected, which is not consistent with the practice in which wireless nodes often turn off their network interfaces in a sleep schedule to reduce energy consumption. Moreover, existing sleep scheduling techniques are unaware of computation status and often lead to unnecessary wakeup overheads. In this paper, we propose solutions to minimize network-wide energy consumption for real-time tasks with precedence constraints executing on wireless embedded systems. Our solutions jointly consider the radio sleep scheduling of wireless nodes and the execution modes of processors. Based on different wireless network topologies, we propose energy management schemes to minimize energy consumption while guaranteeing the timing constraint and precedence constraint. When the precedence graph is a tree, our solution gives optimal result on energy management. The experiments show that our approach significantly reduces total energy consumption compared with previous works.

Introduction

Recent years have seen the deployments of wireless embedded systems, e.g., wireless sensor network systems, in a number of mission-critical real-time applications including manufacturing, battlefield monitoring, security surveillance and so on. These systems are often composed of computation nodes connected wirelessly and collaborating in finishing a set of real-time tasks. Minimizing energy consumption is a key issue in designing such wireless networked embedded systems. Despite the significant amount of work on minimizing energy consumption on stand-alone systems as well as network communication energy, little work has been done to minimize network-wide energy consumption as a whole. This paper proposes solutions to minimize network-wide energy consumption for real-time tasks in wireless embedded systems.

Reducing energy consumption for stand-alone systems has been extensively studied. By reducing processor supply voltage and clock frequency, we can reduce energy consumption at the cost of performance. Various embedded processors such as the Intel XScale processor [9], the Transmeta Crusoe processor with LongRun power management technology [21], and AMD’s mobile processor with AMD PowerNow! technology [1] are all able to effectively reduce dynamic power consumption by supply voltage scaling. To reduce CPU energy consumption on a single processor architecture with independent tasks, dynamic voltage scaling (DVS) solutions were first proposed in [22], [13]. Supply voltage and clock frequency is typically scaled down to achieve energy reduction with a performance trade off. Precedence relationships among tasks have been studied in [2], [7]. However, the above approaches reduce processor energy consumption only. Considering other components in a system together with the processor, research has been conducted to reduce total energy consumption in a system [25], [10]. Recent studies in [11] show that besides using dynamic voltage scaling to reduce dynamic power consumption, adaptive body biasing can be used to reduce leakage power. Chen et al. [3] propose leakage-aware scheduling for real-time tasks in multiprocessor systems. All the above work only considers energy consumption for a single system.

For networked embedded systems, it is important to consider energy reduction for all the nodes in the network. Some work has been done in this area to consider CPU energy for each node. In [12], slack distribution schemes are introduced for scheduling distributed real-time embedded systems while considering admission control and service demands. In [15], static slack is assigned based on the degree of parallelism in the schedule.

Besides the CPU, the radio is another major source of power expenditure in wireless embedded systems. An effective approach to conserve radio power is to schedule radios to be turned off when not in use. A number of sleep scheduling schemes have been implemented in practice [8], [23], [17]. Sleep scheduling introduces new challenges for minimizing the energy consumption in networked embedded applications. For example, it is not suitable to simply assume that the underlying network is always ready. The execution of distributed real-time tasks thus must be jointly scheduled with the radio activity of nodes in the network. When a computation node finishes execution, the wireless radio of the next node in the task execution chain may still be in sleep mode. Communication can only take place when both nodes’ wireless radios wake up. Scheduling real-time tasks incognizant of radio status may not only lead to more energy consumption but also deadline misses. At the same time, most of the current sleep scheduling schemes are unaware of the computation status and often lead to unnecessary wakeups which waste energy.

While the above work consider CPUs and communications in the network embedded systems, other work has been done to consider network-wide energy minimization of both CPU and the communication in wireless sensor networks [19], [20], [24]. In [24], an Integer Linear Programming(ILP) formulation and a polynomial time heuristic algorithm are proposed to find energy-balanced task allocations in WSNs. The allocation is based on homogeneous sensor nodes for balancing energy dissipation. In [20], solutions are proposed for mapping and scheduling the tasks of the application with minimum schedule length subject to energy consumption constraints. This paper assumes task allocation and scheduling has already been done previously, and it assumes a heterogeneous network. The goal of this work is to assign the appropriate execution mode on each CPU to minimize the network-wide energy consumption.

In this paper, we propose solutions to minimize network-wide energy consumption for real-time tasks with precedence constraints executing on wireless embedded systems. First, based on different network topologies, we propose several static energy management schemes to minimize energy consumption while guaranteeing the timing constraint and precedence constraint. Then, we propose an efficient dynamic energy management scheme which will be executed at run time to further improve energy consumption. Optimal solutions are proposed in this paper for static energy management when the given task precedence graph is a tree. The experiments show that our approach reduces total energy consumption significantly compared with the previous work.

The remainder of this paper is organized as follows. Section 2 introduces basic concepts and definitions. Theorems and algorithms are proposed in Section 3. Experimental results and concluding remarks are provided in Sections 4 Experiments, 5 Conclusion, respectively.