Adaptive reflex autonomicity for real-time systems

Abstract

It may appear that for software systems that require strict real-time behavior, the idea of incorporating self-management (and specifically concepts from Autonomic Computing) may add the burden of excessive additional functionality and overhead. However, our experience is that, not only does real-time software benefit from autonomicity, but also the Autonomic Computing initiative (like other initiatives aiming at self-management) requires the expertise of the real-time community in order to achieve its overarching vision. In particular, there are emerging classes of real-time systems for which incorporation of self-management is absolutely essential in order to implement all of the requirements of the system, and in particular the timing requirements.

Autonomic architecture for fault handling in mobile robots

“…The Self-aware layer deals with daily self-managing Developing the AIFH architecture from the MAPE-K [8] and IMD models [11] activities, when processing capacity becomes available. The Autonomous layer is the top tier, system high-level objectives are directed in this layer, and it often includes reflection [13].…”