Some Challenges in Adaptive Fault Tolerant Computing

The objective of this short paper is to describe the interest of adaptive fault tolerance in computer systems and discuss some of the challenges to be addressed. The necessary evolution of computer systems implies adaptation at both functional (application) and non-functional (fault tolerance) levels as far as dependability is concerned. This inherent coupling raises issues both in terms of architectural design and runtime mechanisms. Several approaches based on reflection and component-based software engineering are available today to perform fine-grained off-line adaptation of the software. However, to support dynamic adaptation, runtime modelling, monitoring and assessment are required. These additional requirements create a number of open challenges for future research but also many opportunities that we discuss.