Liberator uses a thread for each DataSource. It can also be configured for multiple connections to the same DataSource, which means each connection is treated as a separate DataSource and therefore has its own thread. This topic shows how the number of threads used for DataSources has an effect on the performance of Liberator.