Title: FAULT DIAGNOSIS IN A FULLY DISTRIBUTED COMPUTER NE WORK

Local computer networks are being installed in diverse application areas. Many of the networks em-ploy a distributed control scheme, which has advan-tages in performance and reliability over a centra-lized one. However, distribution of control increases the difficulty in locating faulty hardware elements. Consequently, advantages may not be fully realized unless measures are taken to account for the diffi-culties of fault diagnosis; yet, not much work has been done in this area. A hardcore is defined as a node or a part of a node which is fault-free and which can diagnose other elements in a system. Faults are diagnosed in most existing distributed local com-puter networks by assuming that every node, or a part of every node, is a fixed hardcore: a fixed node or