In computer networks, fault diagnosis and quick restoration are very important. As a result of a research on faults that have occurred in a Campus Network of ring topology in Shinshu University, the largest cause for faults is a hang-up in nodes. In this paper, a new method of diagnosing hang-ups, an automatic restoration and reporting the result, for each node in networks of ring topology, are proposed.
Each node sends packets for checking faults periodically, and diagnoses a hang-up which occurrs in itself. Using a relay including a timer, a node can recover from failure automatically by cutting a power supply for a short time. And then, a node sends a packet for reporting the result.