Abstract: This paper presents an interactive data replication model based on log parsing for cross-database platform, cross-operating system platform and cross-hardware platform. The model extracts the incremental change log data from the Redo log of the database, and then transfers the extracted log data to the preprocessing end using the parallel synchronous transmission method. The preprocessing end performs the transaction merging processing of the original data, and then obtains the queue data of the preprocessing end through the data subscription form on the loading publishing end, puts the data content into the local cache queue file, finally carries on the data transformation to the local cache queue file, and loads the data in storage. The comparison and analysis show that the model not only can guarantee the consistency of data, but also has a low data replication delay.


Key words: Redo log extraction, data preprocessing, data subscription, data loading, data consistency, replication delay