Performance Benchmark – SQL Server Workload on AWS and Azure
By Fred Wurden, General Manager, AWS Commercial Software Service, AWS Benchmarking
AWS Drives Consistent Superior Performance at a Lower Cost for Online Transaction Processing
In our continued quest to be customer obsessed and offer our customers the best cloud infrastructure for running Microsoft SQL server workloads, we worked with an independent third party to run SQL Server workloads on Microsoft Azure Configurations that get highlighted in public benchmarking studies against one of the comparable options on AWS while running the same workload. For one of the Microsoft Azure configurations, we even picked more premium options (on the basis of performance specs and associated cost) to make the comparison more even. In this blog, we will share the results of this performance benchmark published by Principled Technologies on 12/03/2021.
The study concluded that Amazon Elastic Block Store (EBS) GP3-backed EC2 r5b.16xlarge instance delivers:
- 1.79 times higher transactional throughput
- 1.9 times lower average transactional latency
when compared with Azure Premium SSDs backed E64ds_v4 Virtual Machine (VM) to deliver more consistent transactional database performance.
For this benchmark we selected Amazon Elastic Compute Cloud (Amazon EC2) r5b.16xlarge instance with GP3 volumes and Microsoft Azure E64ds_v4 VM with two storage configurations, one with striped Premium SSD volumes and one with a single ultra disk volume. This made the AWS and Microsoft Azure configurations as comparable as practical in compute and memory configurations. On Amazon EC2, r5b.16xlarge, Intel® Xeon® Platinum 8259CL processor with a base core frequency of 2.50 GHz was selected for this testing. On Azure, two E64ds_v4 VMs, Intel Xeon Platinum 8272CL processor with a base core frequency of 2.60 GHz were selected. Both Azure VMs had the same number of vCPUs (64), and comparable on-demand hourly pricing. The Amazon EC2 r5b instances had 512 GB of memory, and the Azure E64dsVMs had 504 GB.
Principled Technologies ran TPROC-C OLTP workload from HammerDB v4.2 against the selected AWS and Azure instances.
Test Configurations for Amazon EC2 and Microsoft Azure compute instances
Source: Principled Technologies
Based upon Principled Technologies benchmarks, GP3 backed EC2 r5b instance handled 1.79 times the new orders per minute (NOPM) than the Premium SSDs backed Microsoft Azure E62ds_v4 VM (Figure 2). Also, the GP3 backed EC2 r5b instance delivered 1.9 times smaller latency than the Premium SSDs backed Microsoft Azure E62ds_v4 VM (Figure 3).
From a cost perspective, AWS r5b.16xlarge with GP3 storage costs $46.50 per 1000 NOPM whereas Azure E64ds_v4 with Premium SSDs and Azure E64ds_v4 with ultra disk costs $91.93 and $148.60 per 1000 NOPM respectively. The realized performance benefits that AWS configuration offers translate into a more optimized customer experience for your organization via more NOPM and lower latencies.
Behind the study: Why TPC-C workloads?
Online Transaction Processing (OLTP) has been a major optimization target for all IT infrastructure providers to attract enterprise workloads belonging to a variety of market verticals including banking, E-Commerce online platforms and Financial Markets. To perform this study Principled Technologies chose HammerDB’s TPROC-C OLTP benchmark, which runs a TPC-C like workload. HammerDB OLTP benchmark it is more widely adopted in our target customer segment that is looking to run Microsoft SQL Workloads on cloud-based infrastructure. We recognize the added sophistication of TPC-E in the form of more tables, columns, transaction types, check constraints and referential integrity. However, the capability of our target customers to be able to duplicate our experiment and validate our claims and findings drove us to using HammerDB’s TPROC-C workload since it is a freely available tool. If Microsoft decided to open source their TPC-E implementation, we will plan on publishing the TPC-E based benchmarking results as a follow-up.
Why you should run your Windows and SQL Server workloads on AWS
To us, these results aren’t surprising. In fact, they’re in line with the success that customers have been finding running Windows on AWS for over 12 years. Customers like Pearson and Expedia have found better performance and enhanced cost savings by moving their Windows, SQL Server, and .NET workloads to AWS. In fact, RepricerExpress migrated its Windows and SQL Server environments from Azure to AWS to slash outbound bandwidth costs while gaining agility and performance. In this test with Principled Technologies, the results demonstrate that AWS offers better price-performance for all Windows workloads, as well as better ways to run Windows in the cloud. Whether you want to rehost your databases to EC2, move to managed with Amazon Relational Database for SQL Server (RDS), or even modernize to cloud-native databases, AWS stands ready to help you get the most out of the cloud.
AWS can help you assess how your company can get the most out of cloud. Join the millions of AWS customers that trust us to migrate and modernize their most important applications in the cloud. To learn more on modernizing Windows Server or SQL Server, visit Windows on AWS. Contact us to start your migration journey today.
Fred Wurden is the GM of Enterprise Engineering (Windows, VMware, RedHat, SAP, benchmarking) working to make AWS the most customer centric cloud platform on Earth. Prior to AWS, Fred worked at Microsoft for 17 years and held positions, including: EU/DoJ engineering compliance for Windows and Azure, interoperability principles and partner engagements, and open source engineering. He lives with his wife and a few four-legged friends since his kids are all in college now.