安全关键嵌入式系统的能耗测量分析与实时调度研究

发布时间：2018-08-26 09:19

【摘要】：嵌入式系统正深刻地推动着社会的信息化、智能化发展，并且在国民经济中扮演着举足轻重的作用。然而，嵌入式系统由于通常工作在复杂严苛的受限环境中，尤其对于那些采用电池供电的设备，因而嵌入式系统面临着严重的能耗约束。不仅如此，随着互联网的进一步发展以及物联网产业的逐渐兴起，嵌入式系统面临的安全威胁也日益严峻。低能耗、高安全性的嵌入式系统设计正面临着前所未有的挑战。低功耗嵌入式系统设计，离不开对系统能耗的测量分析。本文针对嵌入式软件传统能耗测量方法的不足，从专业测控领域寻找解决办法，设计实现了一种基于LabVIEW的精确能耗测量手段。该方法对嵌入式程序软件所运行的目标设备直接进行电压和电流信号的测量与采集，并采用一种基于离散数据的近似能耗计算方法，求得有效能耗数据。通过实际的对比测量，，证明了本方法的正确性与有效性。要深入地研究安全与能耗之间的联系，对安全算法的能耗特征分析就显得尤为重要。为了获得相关安全算法的精确能耗和性能数据，基于本文提出的嵌入式软件能耗测量方法，搭建了实际的物理测量平台。在该平台上测得了典型安全算法的能耗数据，主要包括：对称密码算法、非对称密码算法、哈希算法。根据所测数据，本文提出一种多维的分析框架，从功率、速度、单位能耗成本这三个角度揭示了安全算法的能耗特征。为了帮助解释和研究这些能耗特征，本文还构建了相关的数学模型来进行归纳和扩展。高效率的安全与能耗感知的实时任务调度算法，对于降低系统能耗、提高系统安全性有显著作用。本文针对异构环境下的实时嵌入式系统，提出了安全与能耗感知的调度长度和能耗权重可调的WTSE（Weight Tuning for Schedule Length andEnergy Consumption）实时任务调度算法。该算法充分考虑了网络通信、安全服务、系统能耗、任务实时性等多种因素，并通过作业调度长度和能耗权重可调的线性判别函数来筛选最优结点。此外，设计了一种任务松散时间利用率机制，以确保安全服务不会占用过多能耗和处理机时间。通过相关算法的对比实验，证明了该算法在有效提高系统安全性和降低系统能耗的同时，可进一步保持较短的作业调度长度和较高的加速比性能。
[Abstract]:Embedded system is promoting the development of information and intelligence, and plays an important role in the national economy. However, because embedded systems usually work in complex and severe restricted environments, especially for those devices powered by batteries, embedded systems are faced with severe energy constraints. Moreover, with the further development of the Internet and the rise of the Internet of things industry, embedded systems are facing increasingly serious security threats. Low energy consumption, high security embedded system design is facing unprecedented challenges. The design of low-power embedded system is inseparable from the measurement and analysis of system energy consumption. Aiming at the deficiency of the traditional energy consumption measurement method of embedded software, this paper looks for the solution in the field of professional measurement and control, and designs and implements a kind of accurate energy consumption measurement method based on LabVIEW. In this method, the voltage and current signals are measured and collected directly for the target equipment running in the embedded program software, and an approximate energy consumption calculation method based on discrete data is used to obtain the effective energy consumption data. The correctness and validity of this method are proved by the actual comparison and measurement. In order to study the relationship between safety and energy consumption, it is very important to analyze the energy consumption characteristics of security algorithm. In order to obtain accurate energy consumption and performance data of relevant security algorithms, a practical physical measurement platform is built based on the embedded software energy consumption measurement method proposed in this paper. The energy consumption data of typical security algorithms are measured on this platform, including symmetric cipher algorithm, asymmetric cipher algorithm and hash algorithm. Based on the measured data, this paper presents a multi-dimensional analysis framework, which reveals the energy consumption characteristics of the security algorithm from three aspects: power, speed and unit energy consumption cost. In order to help explain and study the characteristics of energy consumption, a mathematical model is constructed to sum up and expand. Efficient security and energy-aware real-time task scheduling algorithms play a significant role in reducing system energy consumption and improving system security. In this paper, a WTSE (Weight Tuning for Schedule Length andEnergy Consumption) real-time task scheduling algorithm with adjustable power consumption and safety aware scheduling length is proposed for real-time embedded systems in heterogeneous environments. The algorithm takes into account many factors such as network communication, security service, system energy consumption, real-time task and so on. The optimal nodes are selected by linear discriminant function with adjustable job scheduling length and energy consumption weight. In addition, a task loose time utilization mechanism is designed to ensure that security services do not consume too much energy and processor time. It is proved that the proposed algorithm can effectively improve system security and reduce system energy consumption, while maintaining shorter job scheduling length and higher speedup performance.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP368.1

【参考文献】