Much emphasis in real-time research is put on timing issues. Although timing is what characterizes a real-time system, increasingly important aspects are energy and power consumptions. Lower power consumption yields lower fabrication cost - cheaper package, cheaper power sources and higher reliability. Lower energy consumption yields lower operational costs for battery operated systems or longer lifespan. These two issues are important in all types of applications, but especially for mobile computing & communication, deep-space exploration and medical implants.
The typical approach to power optimization in industry today is centered on “performance-power consumption trade-off”. High-performance systems are usually seen as energy and power eager, while low-power systems are perceived as being of low performance. Although performance is important, real-time applications mostly require predictability while minimizing other aspects influencing fabrication and operational cost. Resource cost, power/energy consumption together with reliability and maintainability are the factors which enforce the product cost.
In this context, there is a need for methods which target power and energy consumption minimization, while fulfilling certain timing constraints. Moreover, these methods are required throughout the whole design process. Approaches for power and energy reduction at lower abstraction levels have already been developed. At higher levels, such as system-level, there are rather few methods addressing these issues, and even fewer which take into account the specific features of real-time systems.
The goal of this project is to develop methods for designing real-time systems with low energy and low power consumption. The design methodology is focused from the start, at system level, on energy and power issues, while fulfilling the timing requirements.