A CENTRALIZED FAULT DIAGNOSIS AND CORRECTION SYSTEM

A CENTRALIZED FAULT DIAGNOSIS AND CORRECTION (CFD&C) SYSTEM HAS THE BASIC REQUIREMENTS OF MONITORING THE HEALTH OF THE TARGET SYSTEM, DETECTING AND DIAGNOSING ANOMALIES AND RECOVERING FROM THE ANOMALIES. THIS PROPOSAL DESCIRBES THE DESIGN OF A CFD&C WHICH WILL HAVE THE CAPABILITY TO PERFORM ON-LINE DIAGNOSTICS AND CORRECTION FORM A CENTRAL LOCATION WITHIN A MULTI-NODE CONFIGURATION OF ELEMENTS. THE CFD&C WILL BE DESIGNED TO BE FLEXIBLE, DYNAMIC, AND ADAPTIVE. ISSUES SUCH AS DATA MANAGEMENT, PERFORMANCE, AND USER INTERFACE DESING HAVE ALSO BEEN ADDRESSED IN THE PROPOSAL. EXPERT SYSTEM TECHNIQUES MAY BE UTILIZED, BUT THE TECHNIQUES WILL HAVE TO BE EXTENDED WITH FEATURES NECESSSARY FOR A REAL-TIME ENVIRONMENT. METHODS FOR COLLECTING DATA, TECHNIQUES FOR DETECTING ANOMALIES AND STRATEGIES FOR RECOVERING FROM ANOMALIE EXIST, BUT THE RELATIONSHIP BETWEEN ACQUIRED DATA AND SYTEM ANOMALIES DO NOT EXIST. TECHNIQUES THAT DETERMINE THE RELATIONSHIP BETWEEN THE DATA COLLECTED, ANOMALIES THAT OCCUR AND CONTROL PROCEDURES NECESSARY TO RECOVER FORM THE ANOMALIES WILL ALSO BE DESIGNED. THE PROTOTYPE DESIGN OF THE CFD&C WILL BE BASED ON THE OPEN SYSTEM ARCHITECURE STANDARDS AND WILL BE DESIGNED ON TOP OF THE UNIX OPERATING SYSTEM. THE USE OF HIGH-POWERED WORKSTATIONS, SUCH AS THE 700 SERIES WORKSTATIONS FROM HEWLETT-PACKARD AND THE SUN SPARC 2 WORKSTATIONS FROM SUN MICROSYSTEM WITH ENOUGH MIPS AVAILABLE DURING THE PHASE II IMPLEMENTATION WILL ALSO BE RESEARCHED.