Database Switcher

Home

Research

Data warehouses contain a wide variety of data that present a coherent picture of business conditions at a single point in time. Informational data is extracted from operational data in the source database, transformed for end-user decision-making and stored in the data warehouse.  Since the data in the source database is continuously changing, it is almost impossible to have real-time replication while maintaining a data warehouse. The process of transforming data from the source database to the data warehouse is time consuming. Users of SagaMap may add new attributes to tables in the warehouse schema. These new attributes are the aggregates of the attributes of the master database. As a result, when data is copied from the master database to warehouse database, data for these aggregate functions need to be computed at run-time during update. This causes more delay in the update process. When this update is in progress, applications accessing the warehouse will not get access to accurate data and this results in lack of synchronization.

How does the Image Switcher solve this problem?

The application has been designed in a way that the there can be multiple copies of the warehouse database, where each copy is an image of the warehouse database. When one copy needs to be updated, it is taken offline. Applications that need to access the warehouse database can now access any of the other image warehouses. In this scenario, applications may need to change their database connection to new databases every time the original database (i.e., the database that the application is currently connected to) goes down. It is a bad design for applications to switch databases manually.

To resolve this issue, there needs to be a mechanism by which switching databases is totally transparent to applications so that they do not realize existence of multiple warehouse databases.

Image Switcher has been designed to address this.  A setup has been built that manages database switching in such a way that even if one of the databases go down, applications will be able to access data from one of the other image databases. The setup updates each of the mirror images at regular intervals.

The tasks performed to update a database are as follows:

·        The database to be updated is taken offline

·        The complete schema is dropped

·        Schema of the warehouse is recreated by executing scripts (Note: These scripts were generated by SagaMap during the creation of warehouse scheme.) 

·        Data is migrated from the source database to this newly created schema

This procedure is completely automated and is repeated at regular intervals for each warehouse image. Image Switcher has also been designed not to disrupt the ongoing transactions. When the database is about to be shutdown, it is shutdown in a way that all the ongoing transactions are completed. During this interval no new connections or transactions are accepted. Click here to see the working model of Image Switcher setup.

Image Switcher
Describes the working model of the Image Switcher setup