Mission critical workloads require high performance to maximize throughput and minimize response time. SQL Server on Windows has proven over the years to be able to deliver both. As organizations begin to run their workloads on container platforms such as Kubernetes, they need to be confident in performance and scale. Therefore, Dell Technologies and Microsoft partnered to test the performance and scale Arc-enabled SQL MI to demonstrate power and scale.
Performance test
The first test demonstrated how Arc-enabled SQL MI would perform relative to how customers run SQL Server today. Therefore, SQL Server running on Windows virtual machine on VMware was used as the performance baseline to compare SQL Managed Instance. Architecturally, these environments are similar in that there is an operating system and hardware abstraction layer on both. In this case, a single instance was running on both environments to keep the test as simple as possible. The MSOLTPE1 workload was used to compare the performance between these two different environments.
1 The MSOLTPE workload is derived from the TPC-E Benchmark and is not comparable to published TPC-E Benchmark results, because this implementation does not comply with all requirements of the TPC-E Benchmark.
Scalability test
The second type of test demonstrated Arc-enabled SQL MI scale, specifically the linear scale of multiple Arc-enabled SQL MIs running under two scenarios:
- Run an increasing workload across 12 Arc-enabled SQL MIs to demonstrate an application load increasing.
- Run a fixed user workload and ramp up the number of running Arc-enabled SQL MIs from one to 12 and then back down to one. This demonstrates what happens when additional workloads or applications are added or running a larger distributed or sharded data application.
Both test two scenarios were run using HammerDB to run the TPROC-C workload.
