Cache target tables in ComputeDB are created either as row-store tables (in the case of Single Table Caching) or column-store tables (in the case of Multi-table caching). Row-store tables more commonly used in relational databases. The advantage of Column-store tables is that they do not have primary keys defined and are efficient while performing analysis on large data sets.

ComputeDB uses an in-built engine to move data to the cache tables. The distributed engine that is used to perform this task uses an optimization mechanism that can process data coming from multiple datasets more efficiently. The cached data is stored in memory with an option to spill to disk if necessary, which makes queries that are run against the cached data much faster.

By default, TDV is configured to use ComputeDB’s Native Data Loading option and the Bulk loading mode is set to “Virtual Scans”.

ComputeDB supports both Single table as well as Multi table caching. Refer to the sections Caching to a Single-Table Database Target and Creating a Multiple Table Cache on a Database Target in the User Guide for more details on how to configure.

The ComputeDB console helps the user to monitor the resource usage by the cluster members and jobs. To access the console, start your cluster and open http:<localhost>:5050/dashboard/ in the web browser. Refer to the ComputeDB documentation for more information.