In a recent benchmark, Microsoft claims that SQL Server on Azure Virtual Machines can be up to 57% faster and cost up to 54% less than running a similar workload on AWS EC2.
According to the GigaOm benchmark, SQL Server 2019 on Azure achieves 57% higher transactions per second (TPS) than running on an AWS EC2 instance with gp3 volumes. Without including AWS license mobility and Azure Hybrid Benefits, workloads on Azure could be 34% less expensive.
Source: https://research.gigaom.com/reprint/sql-transaction-processing-and-analytic-performance-price-performance-testing-22-585-microsoft/
The study compares prices and performances of SQL Server 2019 Enterprise Edition on Windows Server 2022 running on Azure E32bds_v5 instances with Premium SSD v2 disks and AWS EC2 r5b.8xlarge ones with General Purpose (gp3) volumes. Both Azure and AWS virtual machines provide 32 vCPU and 256 GB of memory.
Compared to previous benchmarks on Azure, the results highlight the benefits of the Ebdsv5 virtual machines. Peter Carlin, VP for Azure Database Services, writes:
The Ebdsv5 series of VMs, part of the Ev5 family, are ideal for database and other I/O-intensive workloads because they offer higher remote storage throughput and IOPS per vCPU than any other VM series in Azure.
As the latest Ebdsv5 VMs support higher IOPS when using Premium SSD v2, SQL Server can now perform better in high-frequency transactional workloads. Pam Lahoud, principal PM manager at Microsoft, explains:
The great thing about Premium SSD v2 storage is not only does it have lower latency than Premium SSD, but it is also fully customizable so you can configure the capacity, IOPS, and throughput according to your needs. And you can scale it up or down on the fly! A single Premium SSD v2 disk can be configured for up to 80,000 IOPS and 1,200 MBPS and can have a capacity of up to 64 TiBs.
The report was sponsored by Microsoft and the benchmark setup was derived but not following the TPC-E and TPC-H specifications. GigaOm explains:
The workloads are not official TPC benchmarks nor may the results be compared to official TPC-E or TPC-H publications. (...) Microsoft chose the competitors, the test, and the Microsoft configuration. GigaOm chose the most compatible configurations for the other tested platform and ran the testing workloads.
This is not the first time that Microsoft releases a GigOm report comparing the performances of SQL Server deployed on Azure versus AWS. In the past, AWS replied with the articles "AWS continues to beat Azure on SQL Server price/performance" and "Fact-checking GigaOm’s Microsoft-sponsored benchmark claims".
While some experts questioned the licensing impacts, Fred Wurden, VP of AWS Commercial Software Services, highlighted mismatches in the configurations and the price calculations, questioning the methodology:
This claim by GigaOm and Microsoft is not reproducible and was created by utilizing a modified TPC-E benchmark which uses Microsoft’s proprietary benchmark tool.
AWS has not made any public statement regarding the latest GigaOm report and results.
