Design And Implementation Of Distributed Monitoring System In Cloud

Rapid development of Cloud Computing brings more convince to individual and enterprise users on elastic resource configuration. At the same time, users and service providers pay more attention to Cloud Monitoring. The reason is that Cloud Monitoring provides key figures for cloud platform, application and user, and serves cloud service provider important proof for infrastructure management to implement elastic resource configuration so as to satisfy users’ real time requirements. Through all of above, we can serve high quality cloud service.This thesis studied and analysized existed monitoring systems and compared common used distributed computing technologies. Finally a distributed monitoring system CheckerMS in cloud is developed. Based on REST(Representational State Transfer) architecture style, CheckerMS implemented a distributed, cross cloud platform and plugin mechnisim of resource monitoring architecture. In current version of CheckerMS, monitored resources include physical machine, VMware virtual machine and hypervisor.Mounts of online tests proved that:1) Timing and real time data collection is implemented. Real time collected data can be viewed through web browser, timing collected data is stored into database ready for disposion and analysization.2) REST based distributed architecture decreases weight of the architecture and difficulty of system development, make system extension easier and avoids singal collect server downtime because of too high load.3) Cross-platform feature enables CheckerMS being planted into different cloud platforms to perform monitoring which greatly improved usablity of CheckerMS.4) Plugin mechanism of resource monitoring decouples monitoring architecture and data collection detail and interfaces provided by CheckerMS brings convince to new resource extension in the future.5) Robustness mechanism of CheckerMS can promise users substained and stable monitoring service.