Research On Technology Of Multi-domain Services Fault Management In Cloud Environment

With the development of cloud computing platform, more network resources are published on the Internet by means of services. So how to manage these services from different administrative domains has become a focus of research in service management. Facing the challenge of multi-domain service fault management in cloud environment, what we have completed are as follows:A framework of multi-domain service fault management. The system of intra-domain fault management defines the main content of the services fault management in one domain. Based on the system of intra-domain fault management, we designs the framenwork of multi-domain service fault management, including collaborative mechanism and interaction model.The intra-domain fault management techniques. We can collecting the data of failure through monitoring managed objects by deploying active probe. To the reliability problem of prior probability, we propose a fault location algorithm based on probability updating, which can updating prior probability with historical fault probability. A regulation strategy of services is proposed to improve service availability, which can replace the failed service.The multi-domain fault management techniques. The paper proposes an inter-domain dependency model based on the relationship between symptoms and domain, and an algorithm to divide the cluster of symptoms for the requirements of collaborative diagnosis. We propose a multi-domain fault diagnosis algorithm to locate the fault by estimating the relationship between symptoms and domain. To the problem of algorithm performance, we also design a screening algorithm. According to the requirements of multi-domain fault management, a multi-domain fault interaction protocol is proposed to solve the interaction problem of inter-domain diagnostic message.Based on the above study, we establish a multi-domain fault management system, which can implement the intra-domain fault management, monitor the across-domain services, diagnose the symptoms efficiently, and dispose the fault.We design a solution of multi-domain service fault management. The solution based on the framework of multi-domain service fault management, support by intra-domain fault management techniques and multi-domain fault management techniques, implement in multi-domain service fault management system.