TPC-DS was designed to be representative of a traditional report-based workload.
TPC-DS models the decision support functions of a retail product supplier. TPC-DS does not benchmark the operational systems.
The benchmark models the two most important components of any mature decision support system:
The workload imitates the activity of a multi-channel retailer; thus tracking store, web and catalog sales channels.
The TPC-DS benchmark, models the challenges of business intelligence systems where operational data is used both:
The Benchmark is defined as the execution of the Load Test followed by the Performance Test.
TPC-DS imitates the activity of a business model that:
All three channel sub-systems are autonomous and retain possibly redundant information regarding customers, addresses, etc.
System:
The following are examples of business processes of this retail company:
TPC-DS’ modeling of the business environment falls into three broad categories:
Content File Name | Location | Usage |
---|---|---|
Data generator | dsdgen | Used to generate the data sets for the benchmark |
Query generator | dsqgen | Used to generate the query sets for the benchmark |
Query Templates | query_templates/ | Used by dsqgen to generate executable query text |
Query Template Variants | query_variants/ | Used by dsqgen to generate alternative executable query text |
Table definitions in ANSI SQL | tpcds.sql, tpcds_source.sql | Sample implementation of the logical schema for the data warehouse. |
Data Maintenance Functions in ANSI SQL | data_maintenance/ | Sample implementation of the SQL needed for the Data Maintenance phase of the benchmark |
Answer Sets | answer_sets/ | Used to verify the initial population of the data warehouse. |
Reference Data Set | run dsdgen with – validate flag | Set of files for each scale factor to compare the correct data generation of base data. |