System's Availability Maximization Through Preventive Rejuvenation

Virtualization-based autonomic resource management for multi-tier Web applications in shared data center

“…Such parameter, for example, can be a rate of soft errors due to memory leak or the range of drastic fluctuations of CPU utilization. On the basis of performance monitoring, we can obtain the transition probability from state to state of the analyzed system and determine the rule of system preventive rejuvenation accordingly (Langer and Urmanov, 2006). When the Fault Preventor determines that a failure is going to happen related to a certain virtual server, a command will be triggered to request the Configuration Manager to replace the potentially abnormal server with a fresh copy from the backup pool which stores the latest available state.…”