Design And Optimization Of Cloud Resource Management System Based On Mongo DB

With the improvement of computer technology and Internet technology, databases are facing the requirements of high concurrent read and write, high scalability and high availability. However, the traditional relational databases become inadequate. In recently years, new unrelated databases have shown. By its advantages of open source design, high performance, easy to deploy, easy to use and convenient storage, Mongo DB is accepted by a growing number of companies and individuals. Auto-Sharding of Mongo DB has brought great convenience to users, but there are still some defects in terms of Data Balancing, which need further research and optimization. By deep research of database management, the design and development of cloud resource management system which based on No SQL are of great theoretical and practical significance.Based on Mongo DB data management solutions, a feasibility analysis is conducted, the write performance tests between Mongo DB and a traditional relational database My SQL carries out. By analyzing and comparing experiment data, a specific data management solution is established and high security demand user information is managed by My SQL with its strong stability, while large amount of data and flexible data structure are handled by Mongo DB with its superior read and write performance.Then, a Mongo DB cluster architecture and its performance are analyzed, focusing on its data fragmentation strategy and balanced strategy. The main factors affecting the performance of the cluster is researched, and the characteristics of each shard key and the principles of data equalization algorithm are studied. On this basis, an optimization scheme which based on key-based chip design, query frequency calculation and primary key generation is proposed. Its insert and query performance is tested under a distributed cluster environment, to verify its effectiveness and feasibility.Finally, a cloud resource management system database is designed. The setting of its parameters, the creation of its index and the improvement of its primary key generation are implemented according to the optimization scheme. Under the experimental environment, Replica Set clusters and Sharding clusters are built with Mongo DB, and main functions of the system are developed.