Research And Implementation Of Resource Monitoring And Elastic Expansion Technology Of Cloud Platform

The rapid development of cloud computing meets the requirements of large-scale computing and massive data storage. For the cloud platform,resources are numerous and the situation is complex. For a wide range of resources, it is necessary to make use of monitoring system to manage the entire platform. In order to make full use of cloud computing resources,elastic flexible service have also become an integral part of the cloud computing platform.A state-of-art monitoring system should not only be convenient for users to reasonable scheduling of cloud resources, but also provide data support find the cause of the system failure. In this paper, the deficiencies of the system of the implemented cloud monitoring system has been studied, the architecture of the WEB interface, the two level database storage, timely warning of a number of high cohesion, low coupling has been designed. In order to improve the efficiency of monitoring system,this paper proposes a method to use /proc file system for data collection and data transmission. Fault processing subsystem composed of fault monitoring, log analysis, and notification to the police are also implemented in this paper.At present, in the cloud market flexible service, the redistribution of resources has to be forced to suspend the user's application tasks, which makes users a poor sense of experience. In view of the above issues, this paper puts forward an elastic expansion model based on load forecasting,which further reduces the waste of resources, improving the utilization ratio of resources and reduces the cost input. This model is divided into two parts, the first part is the load forecast, which based on he similarity and periodic load history data, using linear regression, load string matching and two complementary algorithms to predict the next time load demand; the second part is flexible, to meets the demand of the resource expansion, the algorithm has been proposed by this paper to get minimum level cost of expansion amount and vertical expansion amount, and than expanding and releasing the corresponding resources.In order to verify the efficiency and reliability of the monitoring system, the design scheme is realized and applied in this paper, the push-pull hybrid model is tested and analyzed. In order to prove the superiority of the elastic telescopic model based on load forecasting, this paper makes a comparative experiment, which turns out that the average error of load prediction is 8%, lower than the auto-regressive moving average model and exponential smoothing model. The method of cost saving 5.88% than the horizontal expansion method, reduces the cost of resource use.