面向时序大数据的数据库性能研究

Research on database performance for time series big data

在大数据背景下,物联网技术的广泛应用产生了大量的时序数据。如何合理选择最适合的数据库来存储时序大数据是一个重要的研究内容。然而现有的数据库性能对比研究没有考虑数据的具体应用场景,缺乏特定场景下的性能对比实验。为了在存储时序大数据时能够从不同存储结构的数据库中选择最适合当前场景的数据库,文章对关系型数据库、本地NoSQL数据库以及公有云NoSQL数据库在燃气大数据的应用场景下进行了定量定性的实验与分析。实验结果表明,相比于关系型数据库,NoSQL数据库更加适合存储时序大数据。然后,进一步提出了针对不同应用场景下时序大数据的数据库选型建议。

In the age of big data,the widespread application of the Internet of Things(IoT)technology has produced a large amount of time series data.How to choose the most suitable database to store time series data is an important research content.However,the existing database performance comparison studies do not consider the specific application scenes of data,and lack performance comparison experiments in specific scenes.In order to choose the most suitable database for the current scene from the databases with different storage structures when storing time series big data,we conduct quantitative and qualitative experiments and analyses on relational databases,local NoSQL databases and public cloud NoSQL databases in the application scenes of natural gas big data.Experimental results show that NoSQL databases are more suitable for storing time series big data than relational databases.Furthermore,the database selection suggestions for different application scenarios are proposed.