System Requirements for a New Management Control Expert System Generalized Inference Engine Interfacing With a Client/Server Notification System

• Main Author: Brown, Abby H.
• Published: NSUWorks 1998
• Subjects: Computer Sciences
• Online Access: http://nsuworks.nova.edu/gscis_etd/427

Modem organizations struggle with controlling computer platform maintenance expenses. This maintenance effort is further complicated by the existence of multiple, heterogeneous platform environments. A diverse composition of Computer maintenance engineers is responsible for diagnosing and carrying out tasks required to assure quality platform performance. Client/server notification systems are an effective tool designed to manage these computer resources, and yet one of the least mature within the computer platform maintenance domains. The system requirements for a new expert system generalized inference engine were developed based upon research and analysis of expert system generalized inference methods to date, in addition to related efforts accomplished in the Client/server notification system arena. These computer notification systems have done little to address the issues related to the increased complexities in monitoring and managing the maintenance of multiple computer platforms, each with their own unique environments and applications. The systems requirements for a generalized inference engine that utilize binding, matching and unification to categorize alarms written in Prolog have been developed and validated. When the inferencing method was applied to the notification interface, it generalized the heterogeneous data. The rules were applied to an instance of a frame through unification. The inference engine then identified the correct node reporting the error, and applied the appropriate resolution. The requirements also detailed an explanation. The system requirements for the new expert system generalized inference engine provide for effective centralized processing of multiple computer alarms generated by diverse systems. The systems requirements have been tested via test case execution to verify the logical processing of the new generalized inference engine. The results indicate that the growing discipline of Computer maintenance engineering will benefit from the system requirements for the generalized inference engine that provides for effective centralized processing of computer alarms generated by diverse computer platforms. The implication is that organizations will be able to achieve cost-effective centralized heterogeneous platform maintenance with fewer Computer maintenance engineers, each with a lesser degree of expertise required on multiple platforms.