Lightweight embedded systems | Chapter individual record

© 2008 by Taylor & Francis Group, LLC. Lightweight embedded systems are often low-profile, small, unobtrusive, portable processing elements with limited power resources, which typically incorporate sensing, processing, and communication. They are often manufactured to be simple and cost effective. Despite their low complexity, computationally intensive tasks impede lightweight embedded systems from being deployed in large collaborative networks in large quantities. Their sensing capabilities allow their seamless integration into the physical world, while their general-purpose architecture design yields notable advantages such as reconfigur-ability and adaptability to various applications and environments. Lightweight embedded systems are gaining popularity due to recent technological advances in fabrication, processing power, and communication. Despite these advances, there are still significant scientific challenges for researchers to overcome in terms of power management, reliability, fault handling, and security. Unexpected or premature failures raise reliability concerns in mission-critical embedded applications. Failures often erode manufacturers’ reputations and greatly diminish widespread acceptability of new devices. In addition, failures in critical applications, like medical devices, often cause unrecoverable damage. The limited resources of an embedded system in terms of processing power, memory, and storage can often be mitigated by efficient communication that reduces the processing and storage load on an embedded device. In addition to effective communication, the challenge of fault handling is difficult under demands for distributed processing and real-time input from the physical world. Especially in wireless communication, interference from environmental noise and channel collisions greatly affects system performance. Often, power optimization techniques are essential in wireless communication to mitigate power loss from retransmissions. Likewise, security also poses a great challenge for lightweight embedded systems. These systems may be employed for critical applications where user or data security is a major concern. Owing to the limited onboard processing capabilities and low energy consumption, lightweight security protocols are required. These protocols provide lightweight authentication that protect against malicious, coordinated attacks. In this chapter, we first introduce a new algorithm for online dynamic voltage scaling (DVS) for discrete voltages. Minimizing energy consumption in embedded systems is a critical component of extending battery life, and in the case of distributed embedded systems, it extends to the entire lifetime of the system. Advancements in battery technology are being far outpaced by the evolution of instruction count (IC) technology. As a result, system level solutions reduce the burden on batteries by intelligently scheduling the execution of tasks. These methods bridge the gap in meeting the growing energy requirements of an embedded system. A power minimization for online algorithm is presented that carries out DVS and tailors the algorithms to todays real-world processors. The online algorithm schedules tasks without information about future tasks, which is often the case in real-time systems. It utilizes discrete voltage values, as found in today’s dynamically variable voltage processors. The algorithm has a competitive ratio between four and eight, according to our power model. The quality of the algorithm is verified experimentally against processors that do not use DVS. The voltage and frequency settings of commercially available processors, Transmeta Crusoe Model TM5500 and AMD-K6-IIIE + 500 ANZ, are reduced by 20% and 15% by the power minimization algorithm. The algorithm is 16% and 31% less energy-efficient than continuous average rate algorithms, and it can use any voltage value. This chapter concludes with a review of reliability concerns in real-time embedded systems and summarizes the recently proposed reliability optimization techniques. Finally, security requirements of a networked, lightweight embedded system are discussed along with a number of attacks and defenses.

author list (cited authors)
Dabiri, F., Massey, T., Nahapetian, A., Sarrafzadeh, M., & Jafari, R.
publication date
CRC Press Publisher
citation count