Very large distributed system management: a dependable approach

This paper discusses the dependability considerations applied to the development of a prototype system management workbench composed of a set of managed objects whose state corresponds to external managed resources. A method of incorporating manual control operations and coping with unconnected resources is presented.<
>