At present, most systems use open source databases such as MYSQL. The database operation will produce the corresponding logs. In order to ensure that business operations and statistical analysis do not affect each other, we usually copy data from the production library to the analysis library through log synchronization (such as POSTGRESQL or HADOOP). A large number of statistical analysis work will be taken in the analysis library, To do this, a data replication tool for heterogeneous databases needs to be developed to capture and parse BINLOG logs generated by the source MSYQL database in a timely manner and write them to an analysis library such as POSTGRESQL or HADOOP for as-real-time replication of heterogeneous databases