Researchondistributeddata Access Layer Based On Shared-Nothing Structure

With the development of network technology, the amount of information generated by the various services is increasingly large. The data warehouse applications and data analysis is also growing,which require a higher level for database performance. In order to improve the performance of database services, the concept of distributed databases was proposed, utilizing multiple terminals to provide parallel data storage and processing. Distributed database can meet the demands of massive data analysis and processing performance with relatively low cost, with good scalability. This paper s study is mainly focused on the Shared-Nothing architecture-based distributed database access layer design.This article describes the related distributed database technologies, comparing the pros and cons of the three distributed architecture, and determined using Shared-Nothing architecture as the overall system architecture and the MySQL database architecture to achieve module design of distributed database access layer functions, including: system architecture design and implementation, design and implementation of the OAM module, system control module design. Put forward the idea of using the column storage mechanism to enhance the query performance of distributed database systems, including data storage methods and strategies to achieve them. This paper conducted the testing for the designed system by the benchmarks and scalability, making a summary evaluation for the system performance according to the performance test results.The performance results show that this system is excellent in query performance and scalability, compared to the traditional column-based database. This system would meet the demand of the additional database performance upgrading requirements brought by massive data analysis with lower cost, which would has a good application prospect.