Research And Implementation Of The Virtual Machine Monitor Alarm Mechanism Base On CloudStack Cloud Platform

In recent years, the rapid development of cloud computing has become one of the hottest technologies in the field of information technology, it integrates network resources through distributed and virtualization technology to form a pool of shared resources, and as a service to the user to provide the necessary resources. CloudStack cloud computing platform is emerged in such an environment.Actually, it’s necessary for CloudStack IaaS cloud platform that build an effective resource monitoring, and CloudStack still lacks an effective means to monitor the virtual machine. In addition to the approach that remote login each virtual machine to check its health, user can not collect information of all the virtual machine.In response to these problems, this thesis studies the existing cloud monitoring techniques in-depth, and ultimately presents a virtual machine monitor alarm mechanism based on CloudStack cloud platform. The mechanism is divided into SNMP Agent and SNMP server. In the SNMP Agent, the interface provided by virtual library Libvirt timing collect performance data of virtual machines, these monitoring data will be converted to JSON data format, and send to SNMP Server in SNMP Trap Message. In the SNMP server, cloud platform administrator sets up corresponding alarm rules for the virtual machines. When the SNMP Trap server listens to the message, the message will be resolved to obtain performance data first, and then filtration module process performance data to obtain the final alarm events, and alarm events are stored in the database. In order to optimize queries, this mechanism has also increased a alarm migration that contains setting migration policy, the migration policy can be set according to an alarm event "Update Time" field. The administrator determines the alarm event table how long time the alarm event is set by the configuration file, and the alarm events of this time are stored in the table.Finally, this paper validates the mechanism through experimental, and the results show the feasibility and effectiveness of the mechanism. The Virtual Machines in the cloud platform are effectively monitored by the alarm events information of virtual machine.