Integration of spatial data in database acceleration for analytics

Abstract

The IBM DB2 Analytics Accelerator (IDAA) integrates the strong OLTP capabilities of DB2 for z/OS with very fast processing of OLAP and analytical SQL workload in Netezza. Today, not all data types supported by DB2 are available for acceleration. One step to narrow this gap is to enable spatial data - in form of polygons, linestrings, and points - for OLAP processing. Since DB2 and Netezza have very different internal architectures and different limitations, a straight-forward mapping from the DB2 data type to Netezza data types is not possible. We developed a prototype geared toward performance and simplicity for spatial data support in IDAA. In particular, all SQL statement that work against DB2 shall work against the accelerator without any changes. In this paper, we describe how the integration with the accelerator is accomplished so that ingestion and access to the spatial data becomes as seamless and transparent as possible for existing applications. The architecture and implementation aspects for spatial data representation in Netezza as well as the access during query processing are discussed in detail.