| Large-scale collaborative computing platform through the distribution of the host for a unified management and scheduling,so that different hosts to complete the complex computing tasks to solve a single host of the problem of insufficient performance.The collaborative computing platform based on cloud computing has improved the utilization of physical resources and reduced the cost of platform construction by abstracting and virtualization of physical resources with cloud computing.Relying on the cloud computing technology collaborative computing platform with large-scale,scalable,distributed deployment,central management and other characteristics.Therefore,how to manage and schedule the platform resources to ensure that the computing task can be efficient and correct implementation of the distributed collaborative platform design has become the primary problem.However,to achieve efficient management of resources,you need to master the use of platform resources in real time to obtain different sizes of monitoring data for resource scheduling and management to provide a strong data support.At present,the platform resource monitoring field has a lot of related research work,and formed the corresponding software system.However,for the collaborative computing platform,the current resource monitoring system has the following problems.Firstly,fine-grained monitoring of virtual machines can not be achieved in cloud-based collaboration platforms.Secondly,in order to ensure the normal operation of the collaborative computing platform,it is necessary to monitor not only the use of resources(physical / virtual)resources in the platform,but also to analyze the changes of the resources,and to judge the causes of resource changes.Malicious users or software on the illegal occupation of resources and consumption.Therefore,the security situation of platform resources monitoring has become one of the important functions of platform resource monitor.At present,the existing resource monitoring system can only provide the overall situation of the platform or the resource occupancy rate of the physical machine,but can not evaluate the security situation of the platform,the physical machine and the virtual machine.In view of the above problems,this paper designs and realizes the resource monitoring system of collaborative computing platform based on cloud computing with the existing monitoring system.In addition to the realization of the cluster as a whole,the physical machine resource monitoring,the system can run on the physical machine virtual machine and the task of real-time monitoring of the use of resources,and monitoring indicators can be customized according to user needs,greatly extending the original system Monitor the scope,improve the monitoring granularity.In addition,the introduction of security monitoring module in system design,the machine learning method of decision tree technology,combined with the known malicious program resources,respectively,security situation assessment of physical machine and virtual machine.The security situation assessment of physical machine is based on system resource theory,and the security situation assessment of virtual machine is evaluated according to the occupation rate of the normal task when the known task type is executed.Security situation monitoring to achieve early warning of security threats,ensure the availability and reliability of the system.The experimental results show that the monitoring system proposed in this paper can effectively monitor the virtual machine and its task resource occupancy and provide flexible monitoring strategy configuration.In addition,the security situation assessment method based on decision tree proposed in this paper can effectively monitor the abnormal resource occupancy.Experiments show that the prediction accuracy can reach 82% in the absence of other program interference and the characteristics of malicious application of resource consumption,which provides a strong guarantee for platform security management. |
PDF Full Text Request