Prediction of System Critical Event in Virtualized Medical Applications

Medical applications need reliable and secure computing environment. Traditional medical applications only serve local requests and they are deployed locally. Along with the development of remote medical assistants, the number of requests also increases. Medical applications have be moved to the Internet and become medical services that receive many types of requests. The cloud computing platform provides reliable computing environment for medical applications. Virtualization techniques also increase the availability of medical applications. However, the nature of cloud computing allows co-resident of multiple applications, the influence of other virtual machine will affects the reliability of medical applications. System critical event patterns reflect behaviors of systems. In order to avoid system critical events that affect reliability, resources, applications, and services can be scheduled around predicted failure and limit the impact. Virtual machines, which medical services are running on, produce hundreds of system events. These system events include normal level events and critical events. In order to increase the reliability of medical services, we propose a non-intrusive system critical event prediction framework that increases the reliability of medical service in cloud computing.