Study On Application Of NoSQL To Information Feedback System And Its Implementation

In recent years, with the big data and unstructured data increased, the limitations of traditional relational database becomes apparently in these aspects: big data storage, performance scalability, concurrent request processing capability etc. The development of cloud computing drives the study on the new data management solution, then NoSQL database(Not Only SQL) appeared timely. NoSQL database is not clearly defined, it is usually described as non-model, lightweight and open source, NoSQL has good expansibility and flexibility, as well as good performance in dealing with concurrent requests. Most of web systems were designed based on the relational database, now become difficult to meet the new data management request: storage and management of large amount of unstructured data, high concurrent data response request. In this paper, we proposed a hybrid data storage scheme based on relational database and NoSQL database, NoSQL database bears most of the system's reading and writing pressure. At present, the performance of information feedback system in content management and system utilization is not good, so we designed and implement a new type of information feedback system, based on NoSQL technology and improved interactive design.Firstly, this paper introduced the concept of NoSQL database and traditional relational database, analyzed the differences of them, and compared the common NoSQL database in principles and features. Considering that more and more information service systems have a high demand for the response, more and more content of system are unstructured data, this paper selected the document database-MongoDB as the key technology for its extensive applicability. Secondly, analyzed the principle of MongoDB such as auto-sharding and data balancing algorithm, and pointed out the defects. Through the performance test of MongoDB and MySQL, it was confirmed that MongoDB has a better performance on concurrent request processing than MySQL in the same experiment environment. Thirdly, took the advantages and disadvantages of Mongo DB and MySQL into consideration, to design a hybrid data storage scheme. Finally realized an information feedback system to verify its feasibility. The information feedback system was designed from the view of the user and the system itself. Considering the practical application, the business process, system framework and data management mode of the system were designed. Differed from the redundancy, information fragmentation and low scalability of traditional design patterns, the new information feedback system has the characteristics of platform-based, lightweight, high scalability and so on.The hybrid storage scheme based on MongoDB and MySQL, takes the advantage of expansion flexibility of MongoDB, as well as its strength in storing large text, pictures and other unstructured data, the scheme also combined with the stability and the reliability of MySQL, it improved the content management and user experience of traditional information feedback system. The test results that the system based on MySQL and hybrid storage scheme show that the hybrid scheme proposed in this paper can reduces the read/write load of MySQL database, it can effectively improve the query(read) efficiency by 25%, and set aside the ability to expand.