16. Kudu Data Store

The GeoMesa Kudu Data Store is an implementation of the GeoTools DataStore interface that is backed by Apache Kudu. It is found in the geomesa-kudu directory of the GeoMesa source distribution.

Apache Kudu completes Hadoop’s storage layer to enable fast analytics on fast data.

GeoMesa leverages Kudu predicate push-down, column selection, and more, all through the GeoTools API. The Kudu Data Store is a good choice for running Spark analytics on a few attributes at a time, as the columnar storage format minimizes the underlying data reads. In addition, due to Kudu’s compression options and ability to rapidly query non-indexed columns, space on disk is minimized compared to more complex systems like HBase.

Warning

The GeoMesa Kudu data store is an alpha-level feature, and hasn’t been robustly tested at scale.

To get started with the Kudu Data Store, try the GeoMesa Kudu Quick Start tutorial.