Method of and a system for autonomously identifying which node in a two-node system has failed

A method of and a system for autonomously identifying which node in a two-node system has failed are described. The system includes two nodes and a fault-tolerant communication fabric. The fabric defines a plurality of communication paths connecting the two nodes, and fault-tolerant loop-back communication in which each node can send a message to itself utilizing at least one switch structure of the fabric. In addition, each of the two nodes includes logic for performing the service; logic for testing the functionality of the respective node; logic for sending test result messages to both nodes; fault-isolation logic for analyzing test result messages from both nodes; and logic for disabling the other node from performing the service only if the fault-isolation logic determines that the respective node is capable of successfully performing the service and also determines that the other node is incapable of successfully performing the service.