Research And Implementation Of A Cloud Resource Monitoring System For CloudStack

Cloud computing is a new large-scale distributed computing model, whichprovides software, platform and infrastructure services through network. Cloudcomputing has elastic scalability, pay-as-demand, broad network access andmany other features, which can improve resource utilization and reduce costs.Because of the technical and commercial advantages, Cloud computing has beenwidely adopted. Cloud computing resource monitoring system provides data tosupport load balancing, scheduling policy, disaster recovery and many othercomponents of Cloud platform. It plays an important role in maintaining thestability and availability of Cloud platform and has become one of the importantresearch focuses in Cloud.This paper begins with an overview of cloud computing technologyfollowed by introducing architecture, principle and network of CloudStack.Cloud computing platform consists of a large number of resources. To monitorthese resources will generate a lot of metrics. When the metrics transfer on thenetwork, it will increase the burden of the network. Network resource is one ofthe important resources in cloud computing environments. In order to minimizethe resource occupation of Cloud platform, this paper proposed a self-adaptivedata transfer mechanism called SAPM which is based on the analysis andresearches of Push/Pull data transfer modes, ACTC algorithm and otherexisting data transport mechanisms. This mechanism can effectively reduce thenumber of metrics transfer on network, but also has better data consistency.Then this paper analyzes the demand of the monitoring system to determinethe system architecture and functionality and designs and implements a resourcemonitoring system for CloudStack. The monitoring system monitors physicalresources, virtual resources and components of the platform. Resourcemonitoring system has three layers, collection, processing and presentation. Itprovides a user-friendly interface with an intuitive monitoring view which helpsusers to understand the resource usages of Cloud platform.Finally, the CloudStack resource monitoring system is setup to monitor theexperimental platform. The result shows that it can monitor physical and virtualresources, platform components of CloudStack. The monitoring system uses auser-friendly dashboard to display the metrics in charts or graphs to users.