Research On Design And Implementation Of Cloud Security Monitoring System Architecture

With the strong communication ability,storage capacity and computing power,cloud computing became the hot spot of research in various countries' government,multinational enterprises and research institutions.However,cloud computing as a strategic new industries flourish,there are still many users refused to use cloud service,because of its security issues.Users are lack of cloud security monitoring expertise and tools,and the safety assessment information provided by cloud providers are lack of credibility,under this condition,the effective way to improve the confidence of cloud computing services and user's trust to the cloud computing is to build a third-party assessment of cloud security monitoring.However,most of the cloud monitoring architecture or solutions is not to deploy from the perspective of an objective third party,and it's more for a single type of cloud platform,they can't meet the demand of user personalized customization.Thus,the paper focuses on the design of the cloud monitoring system architecture.Firstly,in response to the object-oriented single issues of the current cloud monitoring architecture,and customization needs,the paper put forward a cloud monitoring system based on domain and P/S mode(DPS-CMS).By adopting the idea of domain,DPS-CMS can flexibly and quickly locate cloud nodes and the communication link failure,it's easy to troubleshooting and the specific implementation of cloud monitoring,and have a high degree of regional autonomy.In addition,the DPS-CMS used P/S communication model,users can personalized customize their own monitoring contents based on topic.Secondly,in response to the condition of DPS-CMS node's resource capacity reaches a critical value,enforce monitoring command causes the cloud monitoring service unavailable problem,the paper proposed a method of load prediction based on Bayes.When the load of cloud nodes is too high,it can reduce the frequency of monitoring to ensure the availability of cloud monitoring and cloud service.Finally,in view of the problem of redundant and useless data transmission on DPS-CMS TopicAgent layer,the paper proposed a data preprocessing method based on comentropy,based on the calculation results of copies keyword information entropy,chose the forwarding information of Topic_Agent.Experiments show that the method can reducethe redundancy in the network and useless data transmission effectively,which reduced the load and increased the utilization rate of resources in network.