An expert system for diagnosis and maintaining the AT&T 3B4000 computer: an architectural description

ABSTRACT

Major computer vendors have concentrated on enhancing diagnostic and maintainability aspects of their computer systems to permit a prompt repair interval with a minimal amount of technical support interaction. This paper proposes an architectural description for an automated diagnostic and recovery expert system. The authors obtained sufficient domain knowledge of both the AT&T 3B4000 Computer and the AT&T technical support organization and chose the 3B4000 Computer as a vehicle for the research concepts presented. The architectural description presents a concept based on the information provided by AT&T and does not necessarily reflect actual implementation plans within AT&T.The expert system would be capable of intercepting on-line and off-line messages generated on the 3B4000 Computer. Captured messages would be used by the expert system to define recovery actions that would be automatically or manually executed. The recovery mechanism first references on-line general and technical documents. In addition, various established support data bases and an expert knowledge base are referenced. The combination of documented information and expert knowledge guide the recovery mechanism toward a solution set tailored to correct system errors. This multi-level depth reasoning approach expedites expert system deployment by utilizing established information first, while knowledge-base rules and facts are added in parallel to enhance diagnostic and recovery accurary. The same knowledge and data facilities also would be available to system administration and field support personnel for either accessing user documentation and expert knowledge, or updating documented information and/or expanding knowledge base rules.

REFERENCES

• ATT88: AT&T 3134000 Computer hardware description, issue 1, select code 303-303, AT&T, 1988.
• BAL88: Ballaxd, B. W., Hindle, D., Hirschberg, J. Natural language processing. AT&T Technical Journal, vol. 67, AT&T, 1988.
• CON86: Conry, S. Distributed artificial intelligence in communication systems. Expert Systems in Government Symposium, IEEE Computer Society, 1986.
• GUP84: Gupta, N. K., and Seviora, R. E. An expert system approach to real time system debugging. First Conference on Artificial intelligence Applications, IEEE Computer Society, 1984.
• HAN86: Hanson, R. H. Artificial intelligence in network management systems: A total system viewtx)int. Expert Systems in Government Symposium, IEEE Computer Society, 1986.
• HOW87: Howard, H. C. and Rehak, D. R. KADBASE - A prototype expert system-database interface for integrated CAE environments. Sixth National Conference an Artificial Intelligence, vol. 2, American Association for Artificial Intelligence, 1987.
• MAR88: Marques, T. E. A symptom-driven expert system for isolating and correcting network faults. IEEE Communications Magazine, IEEE Communications Society, 1988.
• VES88: Vesonder G. T. Rule-based programming in the UNIXr system. AT&T Technical Journal, vol. 67, AT&T, 1988.
• XIA86: Xiang, Z., and Sfihari, S. N. Diagnosis using multi-level reasoning. Expert Systems in Government Symposium, IEEE Computer Society, 1986.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE

Additional Classification:
  C. Computer Systems Organization
  C.4 PERFORMANCE OF SYSTEMS
      Subjects: Reliability, availability, and serviceability

  J. Computer Applications
  J.2 PHYSICAL SCIENCES AND ENGINEERING
      Subjects: Electronics

General Terms:
Design, Reliability