This past November at Ignite we announced the public preview of the new Ebdsv5 series of VMs which are ideal for database and other I/O intense workloads because they offer higher remote storage throughput and IOPS per vCPU than any other VM series in Azure. If you haven’t read it yet, be sure to read the blog we published then about the amazing price-performance these VMs offer for your SQL Server workloads running on Azure VMs. If you’ve been following us on Twitter or watching any of our sessions for the past few months, you may have heard us refer to them as the #DoubleSprinklesVMs because of how much we love them for all our SQL Server on Azure VM customers.
Today we are thrilled to announce that the Ebdsv5 VMs are now generally available! You can read more about what these VMs offer in the documentation and in the Azure Compute blog also released today.
Back in November we advertised a 30% improvement to price-performance for these VMs over the Ebdsv4 series. I’m happy to share updated numbers for this series which now show an even better 34% improvement in price-perf. As we did for the previous blog, we used HammerDB to execute a TPC-C1 style workload against an E32bds_v5 VM. The performance metric used for these tests is New Orders Per Minute (NOPM). The following table compares one of our recent E32bds_v5 test runs with the runs we did on the E64-32ds_v4 VM for our original price-perf blog:
As you can see in the table above, we added a few more P30 drives to match the maximum IOPS the VM is capable of. With the lower compute costs and higher performance, this drives the price-perf even lower than the tests we did last November.
If you are currently running SQL Server on older generation VMs in Azure, consider taking a look at the new Ebdsv5 VM series. You may be able to lower your cost AND boost performance with a simple migration to one of these new VMs. And be sure to stay tuned over the coming months as we continue to deliver enhancements to SQL Server on Azure VMs that make Azure the best place to run SQL Server in the cloud!
1The HammerDB TPC-C workload is derived from the TPC-C Benchmark and is not comparable to published TPC-C Benchmark results, as the HammerDB TPC-C workload results do not fully comply with the TPC-C Benchmark.