Research On The Model Of Select The Secure Cloud Providers And Resource Monitoring System

Cloud computing can run multiple services on physical servers and use network to distribute services to users through virtualization technology. As a new service model, users don’t need to spend a lot of costs in building and maintenance of their own data centers or computing centers. They just ask the services from computing service providers. On the other hand cloud computing uses unified management model, so suppliers in cloud computing not only can improve the efficiency of management and service, but also can reduce operating costs. With the development of cloud computing, users who use cloud computing services have become more common and there are more cloud computing services to be chosen. So how to choose a suitable service provider and how to monitor the cloud computing platform and raise the utilization ratio of cloud computing resources have become the hot spot of researching. This paper aim at how to choose a more secure cloud computing service providers for user. Then discuss on how to monitor cloud computing resources and improve the utilization rate of resources for suppliers of cloud computing.The main tasks as follows:First, the cloud computing service level agreement often only focuses on performance while seldom emphasizes security parameters. Besides, customers are confused by several suppliers in cloud computing market. To solve problems above, cloud computing security service level agreement(SecSLA) with detailed indexes is put forward for the first time and an improved ELECTRE method is introduced to the cloud provider selection process. Goal-Question-Metric method was employed to form 50 bottom evaluation indexes for SecSLA;The modified ELECTRE Ⅱ method combined the value of net advantage and net disadvantage to simplify the decision process. The calculation and analysis of instance prove the following aspects:the proposed SecSLA has a relatively comprehensive coverage of the cloud computing assessment factors, such as incident treatment, vulnerability and patch management, configuration management, compliance and availability; the suggested method eliminates the redundancy and dependence on experience of ELECTRE Ⅱ, which makes the realization of automatic process easier. The final evaluation order conforms to the result of ELECTRE Ⅱ, gray correlation method and TOPSIS method, which indicates that the designed method can play a guiding role in the selection of cloud service provider.In view of cloud computing monitoring, firstly, proposing features and program requirements of cloud computing, summarizes cited data that the monitoring platform need to be collected, for example CPU, memory, disk, network usage. Then according to the characteristics of cloud computing across the region, To design a cluster virtual machine resource monitoring system with a hierarchical structure. Not only monitoring its own virtual machine on each host server resources, but also monitoring the host server resource usage in the data center layer of each region, and summary the whole situation of the cloud computing clusters in a client-sever, and estimate the utilization factor of the whole clusters to provide data support for migration management to assess the state of the virtual machine and the virtual machine load. Finally monitoring the cluster resource by libvirt and test on cloud environment through the deployment of open source OpenStack.