Expert system supporting system administrators managing in a distributed, heterogeneous environment

The paper discusses an expert system providing a solution to the complex problem of system administration. It operates in a distributed system management environment that monitors hosts across a network, detects complex problem conditions, and takes automated, unattended actions to resolve common problems. The system was specifically developed to support system administrators managing distributed multi-vendor heterogeneous networks in financial services in Russia. The system intelligence is based on client daemon processes on all hosts of the network that independently can gather system data into a local database and analyze the data using a built-in, rule-based expert system. The system can manage problem events and actions using a sophisticated alert handler and action manager; it distinguishes itself by its scalability and ability to be easily customized by the primary users in order to meet their specific site requirements.