Design And Implementation Of Cloud Monitoring With MongoDB

People have more needs with the development of Internet. Then people have more individuation by using the Internet so we need a more flexibler and adaptability computer. Now Cloud Computing is coming. The cloud computing can make all existing infrastructure, application platform and services together to improve the utilization of hardware by virtualization tools. Cloud is composed of a large number of physical servers, so simply monitoring by system administrators with their experience is not wise when the number of servers is larger. Cloud monitoring is designed to resolve the monitoring of physical machines and virtual machines. When the size of cloud is increasing to the scale of thousands of physical servers and million of virtual machines, the real-time monitoring information will lead to a massive growth in database storage. In this paper, I did the design and development of cloud monitoring which is based on NeoKylin security cloud as automatic monitoring of all physical nodes in the cloud and all virtual machines, then provide monitor data to system administrators for dispatching virtual machines. In this module, I selected the combination of ganglia and libvirt to achieve the function by contrasting ganglia and nagios. For the problem of massive monitoring data I conducted a variety of databases to study. Ultimately, I choose one kind of NoSQL-MongoDB to storge the monitoring data insead of Relative DataBase. MongoDB is a kind of document-oriented database, its storage type is key-value and auto-sharding are good at the level of expansion. This database is better to such application with largemount of data but low value.This paper finally realized the all physical nodes and virtual machines in the cloud monitoring, and stored these data in MongoDB then displayed them in web. The paper also provides the function of the forward and reverse lookup to the system administrators to monitor use.