Autonomic Software Systems

GREATER demand for complex functionality, particularly in the aerospace industry, necessitates greater degreesof autonomy in our computer software and systems. Autonomic computing [1] and related initiatives aimed atachieving self-properties, self-healing, and ultimately self-governance are becoming more important and relevant asa result.However, there is still much to be done to achieve the vision of self-managing systems [2]. This is particularly true inembedded systems, as highlighted by Nordstrom et al. in their article “Autonomic Healing of Model-Based Systems”.The authors highlight the need for resilience and availability of systems in complex environments, where models maynot necessarily provide all the information needed for informed decision-making regarding future behavior, leavingit to be a relatively human-intensive process. The authors describe their work on facilitating self-healing, one of theoriginal self-CHOP properties and a vital component of autonomic systems, in model-based development.We must also have assurance regarding the decisions made in self-managing systems. The potential implicationsof incorrect decisions are great, but decisions must often be made by autonomous components of ground-controlledsystems simply because of time constraints. In their paper “Game-Based Model Checking for Reliable Autonomyin Space”, Bakera et al. describe the game-based model checker GEAR, a useful tool for automatically provingproperties of autonomic systems for increased levels of assurance. They illustrate its usage with respect to a casestudy involving the ExoMars Rover.Finally, the longer-term vision of creating Autonomic Software Systems is discussed with a return to biologyfor further inspiration in creating a holistic self-managing environment in Next Generation Autonomic SystemEnvironment.