Towards context-awareness and self-configuration of embedded systems: resource usage and real-time characteristics
Mariusz Pelc, Richard John Anthony, Paul Ward, James Hawthorne · 2009
Embedded systems are becoming truly ubiquitous in society, with new applications and usage domains appearing at a rapid rate. There is an increasing interest in, and need for, solutions which give these systems the ability to adjust their behaviour to the current environment conditions. Although context-awareness and self-configuration features are obvious responses to the needs mentioned, each potential implementation of them is inevitably a compromise between level of flexibility achieved and the accompanying resource costs. This balance is especially vital in the case of resource constrained embedded systems, which dominate in many application domains. In this situation it is extremely important to provide solutions for which the resource configuration can be accurately predicted for a given scenarios, and thus validation can be extended beyond behavioural correctness, to include real-time performance characteristics and resource usage such as memory, even though the precise operating circumstances may be unknown in advance. In this paper we present a middleware architecture and accompanying software component model supporting self-configuration, context-awareness and dynamic decision making within embedded systems. The solution is scrutinised from the point of view of resource usage and efficiency and real-time performance, using typical hardware and operating platforms.