Elements of the Self-Healing System Problem Space

One of the potential approaches to achieving dependable system operation is to incorporate so-called “self-healing” mechanisms into system architectures and implementations. A previous workshop on this topic exposed a wide diversity of researcher perspectives on what self-healing systems really are. This paper proposes a taxonomy for describing the problem space for self-healing systems including fault models, system responses, system completeness, and design context. It is hoped that this taxonomy will help researchers understand what aspects of the system dependability problem they are (and aren’t) addressing with specific research projects.