Benchmarking PostgreSQL for data warehousing

Often in a competitive market, corporations need tools to increase the accuracy and efficiency of the decisions. Facing this, several companies that have no resources to buy the commercial systems available because of the high costs, have no chance to buy a Decision Support Systems (DSS) platform. This paper is dedicated to a benchmark of PostgreSQL DBMS for data warehouse as part of a low cost platform. To accomplish this task we followed the TPC-H and DBT3 benchmarks to simulate a data warehouse workload. These benchmarks simulate multi-user environment and run complex queries, which execute: aggregations, nested sub queries, multi joins and others. Considering the results, we were able to point some PostgreSQL’s problems in the TPC-H execution, which were the main reason to execute the DBT3. Finally we demonstrate that this DBMS will become a real alternative to data warehousing with improvements on the optimizer and some structures.