Research On Fault Diagnosis And Self-Healing Based On NVP Technology

Cloud computing provides users with highly scalable elastic computing services.With the continuous increase of cloud computing scale and resources,a variety of failures occur during the operation of cloud computing system,and serious failures such as downtime may cause huge losses.Therefore,it is necessary to study fault diagnosis and self-repair technologies to ensure the stability and improve the availability of cloud computing system services.This thesis studies a mechanisms based on NVP technology to find suspicious fault systems,establishes fault diagnosis and self-repair mechanisms,and optimizes the antifailure model to eliminate the effects of aging.The main work is as follows:(1)The screening mechanism of suspicious system based on NVP technology is studied.NVP monitors task completion and collects monitoring indicators such as cpus,memory,disks,and networks of VMS in real time to detect faulty systems in the cloud system.(2)The hybrid fault diagnosis and self-healing mechanism are established.For the system which may be affected by the fault,a hybrid fault diagnosis and self-healing mechanism is established.Firstly,a multi-valued decision graph is used to determine the severity level of the system according to the symptoms.Then for the system at the severity level,the fault result corresponding to the symptom is diagnosed by using the fuzzy logic integrated neural network,and the corresponding prescription is applied to self-repair the diagnosis result.Finally,an example is given to verify the effectiveness of the proposed fault diagnosis and self-healing mechanism.(3)The rejuvenation strategy of aging system is optimized as a prescription for diagnosis and repair of intermediate fault state system.Firstly,the architecture of virtualization technology in cloud system is analyzed,and the combined software rejuvenation strategy is studied.In order to determine the aging state and rejuvenation time,weibull distribution was introduced to simulate and model the aging state transition process of the system,and the task success probability was evaluated based on this model,the probability of mission success is evaluated.Then,on the basis of AHP supplemented by objective parameter entropy analysis,the AHP is optimized,and the rejuvenation decision function is constructed in real time as the basis for the rejuvenation decision and the software aging.Finally,the rationality and effectiveness of the hierarchical structure and judgment matrix construction,and the optimization effect of the rejuvenation strategy on system stability and average task completion time are verified through experiments.