The second test adds a DSS workload to the system by generating TPC-H-like workloads. Note that a “TPC-H–like” workload means that the test results are not certified. The TPC-H benchmark consists of ad-hoc queries and concurrent data modification across large sets of data. Businesses might use a DSS to analyze a large volume of data to generate reports that facilitate evidence-based business decisions. We monitored throughput and storage metrics only as part of running a DSS workload. Therefore, we do not review metrics such as TPC-H Composite Query-per-Hour (QphH@Size) in this document.
For the DSS use case, we ran one Oracle database and two SQL Server databases in parallel to generate throughput on the PowerMax 2000 storage array. The DSS workload ran in parallel with the OLTP workload, creating a combined workload on the system.
The following table details the configuration of the TPC-H–like tests:
Table 4. TPC-H-like benchmark configuration for DSS workload use case
Database scale factor |
1,000 |
3,000 |
4,000 |
Database size (TB) |
2 (VM1 + VM2) |
3 |
5 |
Number of virtual users |
2 |
1 |
3 |
Test duration (min) |
30 |
30 |
