Software architecture for self-healing of concurrent and distributed components

This paper describes a self-healing component architecture that is able to detect object anomalies in the component, reconfigure inter-component and intracomponents before and after repairing the sick object, repair it, and then test the healed object. A distributed and concurrent component is structured to tasks (active objects), connectors, and passive objects accessed by tasks. Each self-healing component is designed with two layers: the service layer and healing layer. The service layer of a component provides functional services to the other components, whereas the healing layer monitors objects in the service layer and repairs the sick objects detected. In addition, this paper describes constituents of the healing layer of a concurrent and distributed component, which include the Component Repair and Reconfiguration Controller (CRRC), Component Monitor, Component Repair and Reconfiguration Plan (CRRP), and Component Repair and Reconfiguration Executor. To illustrate this approach, the elevator system is considered.