Research And Design Of Security Monitoring System Based On Cloud Computing Platform

In recent years,cloud computing has gradually become a viable and effective service and computing model.The use of related technologies has greatly improved resource utilization,and more and more deployed in university laboratories and enterprises.With the increasing scale of cloud computing and the high concentration of resources in cloud data centers,how to ensure the stability,reliability and security of cloud services cannot be ignored.Clouds are inherently complex due to the large amount of resources involved and the need to meet SLAs(Service Level Agreements)for different users.Therefore,in order to ensure that tasks are allocated resources must be flexible and dynamic,it is necessary to monitor the resources of the cloud platform in real time.Through the research on the learning and monitoring technology of cloud platform architecture,the current monitoring system has different shortcomings such as:no load prediction function and alarm mechanism,the stability of the system is poor,when the virtual machine reaches a certain number,the performance is sharp.Falling and so on.According to the actual monitoring requirements of the existing cloud computing platform,the related technologies in the cloud monitoring system are studied,and the cloud monitoring system is designed to realize real-time monitoring,prediction and alarm of the virtual node resources and service status of the cloud platform.The platform provides a foundation for efficient,stable,and secure service establishment.The main work of the thesis is as follows:1.Through the research and analysis of the current monitoring system of cloud computing platform,the organization structure of the paper is given,the research background and significance of the cloud monitoring system are expounded,and the key technologies are introduced.The motivation and comparison of cloud monitoring are analyzed.Currently commonly used monitoring systems.2.In the research of cloud monitoring load forecasting,based on the traditional neural network prediction method,a load forecasting method based on neural network and FLS is proposed.This method starts from four aspects by formulating input and output variables and prediction rules.Fully considering the real situation of the virtual machine load,the prediction is realized according to the monitoring history and the current CPU utilization and workload of the virtual machine.3.In the research of cloud monitoring alarm mechanism,the 1:N:N monitoring alarm mechanism is proposed,and the alarm process is improved,which can be applied to multi-scene monitoring,realize automatic alarm function,and adopt time in data processing.Sliding window mechanism to ensure better monitoring services.4.Perform an architecture analysis on the cloud monitoring system and divide the monitoring system into three technical modules.In the data acquisition module,the Pull&Push data acquisition mode is adopted,so that it can intelligently switch between the push mode and the pull mode according to different conditions,and the collected data is uniformly encapsulated and transferred into the data storage module.In the data storage module,a data double write strategy is adopted to avoid data loss.In the monitoring alarm module,the improved monitoring alarm model and the load prediction mechanism are adopted,and the function of notifying the relevant personnel when the fault occurs is realized.5.The cloud monitoring system adopts the MVC design mode to design the relevant functional modules of the cloud monitoring and carry out related design of the database.Finally,the platform is functionally tested and performance tested and analyzed.