Appendix A. About OLTP and OLAP Workloads

Postgres Pro is traditionally used for transactional (OLTP) workloads. It determines architectural decisions on data layout and processing algorithms. OLTP workloads have the following characteristics:

  • operate on small volumes of data

  • perform read/write operations on most columns

  • require strict ACID guarantees

  • require short response time

With these characteristics, an optimal approach is to store and process the data in a row-based format.

Analytical (OLAP) workloads are significantly different:

  • process large volumes of data

  • perform read-only operations on a small subset of columns

  • do not require strict ACID guarantees

  • prioritize throughput over latency

Considering the difference, Postgres Pro AXE holds a significant advantage due to its architecture:

  • local, network, or S3 storage for better scalability and price-to-performance ratio

  • columnar layout on storage and in memory to work efficiently with individual columns and for higher level of compression

  • vectorized execution to process the data in batches

Postgres Pro AXE enhances the traditional Postgres Pro OLTP functionality by offering possibilities to operate with OLAP and HTAP workloads.