Interactive Self-Healing for Black-Box Components in Distributed Embedded Environments

As self-management features of computer systems depend on self-knowledge we have to cope with the issue that most large systems are not entirely known to the self-management software. Especially when system parts are supplied by external companies which want to preserve their intellectual property we have to cope with black-box components. This paper presents a novel approach for self-healing in distributed embedded systems containing black-box application software. The interactive self-healing process is based on well defined system knowledge and enables the construction of a functional dependency graph which serves as basis for rule-based root cause analysis and self-healing.