TDV provides a parallel caching option that is used to maximize performance of data caching. If cache loading performance gains cannot be realized using the TDV native loading option, you can try the TDV parallel option. This option uses statistics derived from a unique numeric key to create multiple partitions that can be used to load the data using parallel processes.

During parallel loading, the reader and writer work in parallel and are not blocked or waiting for one another. The reader reads the table rows into TDV memory and the writer pushes them into target database. You can have multiple readers and corresponding writers that work on different partitions simultaneously. These parallel threads use JDBC-based read write calls.

This performance option is available as listed for the Native Cache Target Support column of the table in Supported Cache Target Storage Types.

It requires some setup beyond TDV configuration parameters. The parallel cache loading option requires that the view that you are caching has:

Because partitioning is being determined on the key column, the cardinality statistics help determine how many unique partitions can be loaded in parallel. The number of unique partitions that can be found by the statistics collection determines how many threads can be used to load the data into the cache.

The next time your cache is refreshed, this cache loading option is attempted.