For as long as databases have been in existence, performance tests have been used to assess performance of a single instance under ideal circumstances. While these tests are valuable to demonstrate what is possible, they are not a typical use case, especially on a container, cloud-enabled platform such as Kubernetes. By default, Kubernetes is a multi-node deployment built and designed for mixed workload and massive scale. Therefore, we wanted to assess PowerStore and PowerEdge, running multiple instances of Arc-enabled SQL Managed Instance GP.
To demonstrate this, a scale up and scale out test was performed on our test environment, consisting of Arc-enabled SQL MIs running on a single PowerStore 9000T and a single PowerEdge R650xs worker node.