This welcome new addition to the study of real-time embedded systems should satisfy both generalists and researchers looking for current theories, algorithms, and a comprehensive analysis. The authors provide detailed approaches and implementation results from their own research while also exposing cutting-edge issues. A common theme throughout the book concerns the many dimensions of scheduling for resource allocation and processor optimization, while recognizing the constrained embedded computing system’s cooling and power environment. For example, because new mobile devices have limited memory and battery power, the scheduling and optimizing of both on-board and off-board memory is impacted by the power required, the battery or memory type (such as phase change or scratchpad memory), and whether the battery will be recharged or have a single lifetime charge. This battery-aware scheduling is especially suited for wireless sensor networks (WSNs).
As embedded systems expand into consumer electronics, home applications, office equipment, and automobiles, the need for guarantees is paramount and potentially lucrative. This book has a number of excellent chapters that build upon each other. The authors’ foundation--optimization, synthesis, and networking--starts with the more traditional real-time issues of optimization, time management, and scheduling techniques. They then apply the techniques from the embedded real-time domain all the way up to cloud computing.
The final chapters address cutting-edge cloud computing and WSN issues--again, the theme is resource scheduling. The authors address the unique challenges of analyzing ad hoc and wireless networks, and include what they believe is unique research on battery behavior modeling. They deal with managing the power available on the processor, energy limits, and devices with limited power, and they present their task scheduling and energy-aware online (TSEO) algorithm.
Unique embedded processing issues deal specifically with scheduling. Basically, can the processor manage this while dealing with its own limited power consumption? This is the hurdle that next-generation designs must overcome as the trend toward portable, scalable, and sophisticated sensor networks becomes pervasive. Cloud computing is also analyzed as it pertains to scheduling computing resources. The authors lay the groundwork for scheduling in the cloud. This includes dealing with such issues as the speed to copy a disk image, swapping images, preemptive resource allocation, and basic autonomic resource contention. This runs the gamut of real-time applications, but taken into a virtual cloud physically executing all over the network. The resulting infrastructure as a service (IaaS) is the next level of computing.
The authors leave readers with a vision of embedded on-board processing in this new area: sensors on unmanned vehicles will perform their own video analytics and radar track fusion to reduce network traffic through better information fidelity. To implement this vision, these systems will have to perform the necessary scheduling and resource optimizations. The tools and techniques described throughout this book lead to the next generation of portable, scalable, and sophisticated sensor networks.