| With the rapid development of social network and big data,a variety of new databases are emerging and developing.NoSQL database,in particular,has unique advantages for specific data,and there are many NoSQL databases which are widely used in practical systems,and playing an important role.Usually,large scale systems with high real-time requirements will use both SQL database and NoSQL data to ensure query speed and reduce data storage overhead.However,the cost of maintenance and application development of a system with a variety of heterogeneous database system is huge.When introducing a new database component development and maintenance personnel have to learn new interfaces,test new drivers,and write other extract-transform-load components.Aiming to solve this practical problem,this paper mainly studies the unified access method of distributed heterogeneous databases.The work of this paper can be summarized as follows:1.This paper discusses the method of solving the semantic heterogeneity of data based on ontology,and proposes a general schema method based on this.It can not only deal with the semantic heterogeneity of traditional relational database,but also deal with the semantic heterogeneity of new non-relational databases.2.The basic framework of foreign data wrapper based on PostgreSQL is proposed,and based on this framework,Hbase_FDW is designed and implemented.Hbase_FDW automatically transforms standard SQL statements into Hbase internal API operations,so that users only need to use standard SQL statements to do relevant operations on Hbase.3.The unified access method of distributed multi-heterogeneous databases is proposed.According to the actual requirements,this thesis combines the design and implementation of the Hbase_FDW and the official MySQL_FDW,Redis_FDW,MongoDB_FDW and PostgreSQL_FDW,integrate the distributed databases system of Umeng+,by only use the standard SQL users can do cross-database and cross-node operations. |
PDF Full Text Request