Service Reliability Oriented Modeling for the Failure of Cloud Data Center

As the failures in cloud data center is inevitable, some severe failures may prevent the cloud data center from continuing providing service, resulting a great loss of economy. Thus, analyzing and seeking for the rules of the failures of cloud data center are the preliminary procedure of providing better services from the cloud data center. In this paper, from the perspective of service reliability, which is defined as the ability of successfully processing service requests, the data from the cloud data center of an insurance company are collected and will be analyzed. Some relationships between the average response time and the access amount are supposed to be realized. And the regularity of the access amount and access failure amount has also been discovered. Our finding may have a help for improving the service reliability and providing a theoretical support for reliability modeling and resilience analysis of cloud data center.