A Next-Gen Network Switch Software Automation Upgrade Method Based on a Two-step Inference Expert System

This paper presents an automatic software upgrade method for an next-generation network (NGN) switch instead of the manual software upgrade. An expert system is designed for the automation upgrade, and a two-step inference is adopted in its inference engine. The expert system can handle most of exceptions and faults during the upgrade and make the software upgrade automation, and its knowledge base can be expanded by on-line learning and off-line learning. This paper gives the detailed implementation of the automatic upgrade strategy and the expert system, and describes the two-step inference algorithm. The automation upgrade method can save the NGN switch upgrade time, improve the reliability of the NGN switch software upgrade and decrease the network administrator (user) work force. The experiment result shows the automatic software upgrade method has the great performance improvement compared to the manual software upgrade.