Large-scale Cloud Monitoring System With Elastic Extension

The monitoring is an important way to get accurate system information, it is a key part of the cloud management system. It is a challenge to control the large-scale cloud computing system. This paper studies the elastic scaling technology and efficient problems in the monitoring of cloud, and makes a prototype system based on the research.Large-scale cloud monitoring have difficulties on the flexibility and dynamic extension of cloud computing. The existing software is shot of getting information efficiently. They can’t change configuration at runtime to meet the elasticity.Automatic discovery can monitor resources dynamically by changing monitor configuration. The existing technical problems are timeliness and reliability. The paper propose a discovery model based on the advanced message queuing technology. The client sends a message to the queue and the server can get it. It can get the elasticity changes rapidly, then change the monitoring configuration. It ensures the reliability with the support of queue message persistence.There is the efficient problem with the traditional fixed monitor interval model. The paper makes the function of changing configuration dynamically. It propose a prediction model based on Naive Bayesian Classification. It makes trend prediction on the current status based on historical monitoring information. It adjusts the monitoring interval dynamically which can ensure the efficiency and accuracy of the cloud monitoring. Experiments show the model can optimize the information transmission frequency, so as to achieve efficiency of the monitor system.The paper designs and realizes JTangICM intelligent cloud monitoring system. It contains dynamic discovery function and efficient monitoring interval module. It provides extension of monitoring resources, elastic scaling, and unified programming interface. These help deal with large-scale cloud computing environment, and be easy to get secondary developed by administration.