Reference architecture for distributed systems management

摘要: Management of computing systems is needed to ensure efficient use resources and provide reliable timely service users. Distributed are much more difficult manage because their size complexity, they require a new approach. A reference architecture for distributed management proposed that integrates system monitoring, information management, modeling techniques. Three classes management—network services devices, operating resources, user applications—are defined within this framework, detailed hospital application presented clarify the requirements managing applications. It argued performance applications must be considered from all three perspectives. Several prototypes under study COnsortium Research on Systems (CORDS) described illustrate how such an could realized.