Overview
Results (summary)
Oracle 12c
Execution time of a set of summary queries is divided by 20 for an additional disk space usage of 42% using ACOD-BI compared to a basic datamart.
It is divided by 3000 for an additional disk space usage of 68% using ACOD-BI compared to a direct use of the data warehouse.
Oracle Autonomous Data Warehouse
Execution time is divided by 15 using ACOD-BI compared to a basic datamart.
It is divided by 45 compared to a direct use of the data warehouse.
Source database
The source is a data warehouse with two fact tables, invoices and orders :
• 25.0 M invoice lines
• 2.25 M order lines
The source data warehouse disk space usage is 4.76 Go.
The data warehouse contains 57 attribute columns available as:
• 179 filter or group by criteria for invoices
• 95 filter or group by criteria for orders
Reporting
Reports are done by a request set :
• 25 distinct requests (on invoices)
• 2385 request executions
The reporting statistics correspond to the execution of a request of the set in 99.9999% of cases and to the execution of a random other request in 0.0001% of cases.
With this statistics ACOD-BI generates a datamart with an optimum execution time for the request set and which allows any other request.
Results (details)
Data warehouse alone (without ACOD-BI)
Request set execution time : Several days.
Datamart – Detail tables only (with ACOD-BI without aggregates)
Datamart disk space usage : 2.29 Go
Request set execution time : 01h47
Datamart – Detail and aggregate tables (with ACOD-BI with aggregates)
Aggregates disk space usage : 0.96 Go
Datamart disk space usage : 3.25 Go = 2.29 Go + 0.96 Go
Request set execution time : 00h05
