Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Building a Hybrid Database-as-a-Service Platform with Azure Stack HCI
Overview Solution Architecture Lab Setup Performance-Oriented Test Scenarios
Test methodology introduction
HammerDB initial configuration
HammerDB CLI Configuration
SQL MI creation and configuration
Test groups defined
Test Group 1 results
Test Group 2 results
Test Group 3 Results
Test Group 4 results
Test Group 5 results
Test Group 6 results
Reporting visualizations
Testing synopsis
Functional Test Scenarios Conclusion Appendix A: On-premises monitoring setup Appendix B: SQL MI and HammerDB Test Harness details Additional Resources

Test Group 6 results

Test Group 6 results

  • As our final test, we ran the same HammerDB TPC-C workload across all ten Azure Arc-enabled SQL Managed Instances. We observed that our overall TPM numbers were lower, but we saw impressive results on the first four quad-vCPU instances when compared to the rest of the instances, as shown in the following figure. 

    With the previous test, the TPM ramp-up and user step-up workload jumped in 40,000 TPM increments. With this denser workload profile of ten SQL MIs, the TPM increased to averages around 20,000 TPM for each stepped user increase. 

    Note: We never observed CPU, memory, or storage performance saturation at any level of the solution stack. This means that we could have pushed the system even harder if wanted. 

    Chart, table Description automatically generated

    Figure 26.    Test-6 user count and TPM metrics

    For SQL MI specifications, see Test groups defined.