Migrating over 60,000 SQL Server databases to AWS with Mindbody

Mindbody’s infrastructure previously consisted of two tiers. The first was a central tier that used SQL Server Enterprise edition with Always On availability groups for HA. The second was a customer tier that contained over 60,000 standalone SQL Server 2016 Standard databases and operated without HA.

“With that number of databases, you can imagine the hardware costs,” says Tim Ford, lead database engineer at Mindbody. “We wanted to move to an operationalized model where we could dial up and down compute resources, and therefore costs, as needed. We also wanted to innovate, and to do so, we needed to adopt infrastructure-as-code and other modern methods to get away from repetitive tasks.”

When Mindbody needed to exit its data center because of an expiring lease, it saw an opportunity to transform its infrastructure. The company chose to migrate to the cloud, selecting AWS for its proven track record in supporting large-scale migrations of business-critical workloads.

Mindbody’s migration began with an AWS Optimization and Licensing Assessment (AWS OLA), a program that empowers both new and existing AWS customers to assess and optimize their on-premises and cloud environments. Through the assessment, Mindbody evaluated its resource requirements and mapped on-premises workloads to appropriate instances on Amazon Elastic Compute Cloud (Amazon EC2), secure and resizable compute capacity for virtually any workload.

Using AWS OLA, Mindbody optimized compute and memory usage, achieving a cost savings of more than 20 percent. The assessment also helped Mindbody better understand its License Mobility through Software Assurance rights and choose between bring-your-own-license and license-included models.

During the initial testing phase, Mindbody saw an increase in latency when migrating databases from its Los Angeles data center to the US West (Oregon) AWS Region. There were more than 20 ms delays between Mindbody’s application, which was still on premises, and the database layers that were hosted on AWS. On the recommendation of AWS experts, the company migrated its databases to an AWS Local Zone in Los Angeles to bring them closer to the application. This reduced latency significantly.

To maintain HA for its SQL Server databases without relying on a shared-storage option, Mindbody engaged SIOS Technology Inc. (SIOS), an AWS Partner, for a solution. SIOS provided Mindbody with block-level replication between volumes in Amazon Elastic Block Store (Amazon EBS), which provides easy-to-use, high performance block storage at nearly any scale. Mindbody also adopted a third-party tool to automate consistent deployments of its SQL Server environments in the cloud.

With the infrastructure foundation in place, Mindbody scheduled database migrations during periods of off-peak customer activity. “We moved databases one at a time, using our existing monthly maintenance windows for the cutover phase,” says Ford. “That way, we didn’t need to fully shut down operations, and we achieved the migration with minimal impact on our customers.” Overall, the initial migration took 18 months.

After migrating the databases and the application to the AWS Local Zone, Mindbody conducted another AWS OLA. As a result, the company saw an opportunity for further optimization of performance and costs in moving to the US West (Oregon) Region. Mindbody also used a combination of SQL Server Always On availability groups and SIOS’s block-level replication to migrate the databases. The company was given a 30-minute maintenance window for the AWS to AWS migration cutoff of all databases and applications, but the process took place in 12 minutes.

After the migration, Mindbody improved backup performance by 30 percent and decreased latency in one Region. The company also achieved a cost savings of almost 13 percent by moving its hosting to the parent Region, and it expects an 8-month return on investment.

“By migrating SQL Server to AWS, we not only increased availability but also boosted disaster recovery for both our data and our customers’ data,” says Ford. “We’ve increased the frequency of our log backups by four times using AWS, significantly accelerating our recovery speed.”

Thanks to the successful migration, Mindbody has laid the foundation for ongoing modernization and optimization. To simplify database management and further improve scalability, the company is now exploring managed database services such as Amazon Aurora, which provides high performance and availability at a global scale.

Mindbody is also modernizing its application to support the cross-platform solution .NET, which will empower the company to run on Linux and adopt AWS Graviton processors. These are custom-designed server processors developed by AWS to provide excellent price performance for cloud workloads running on Amazon EC2. By migrating to Graviton processors, Mindbody might save an additional 30 percent on compute costs while maintaining or even improving application performance.

Mindbody’s strategic collaboration with experts and use of the right AWS services helped it overcome complex challenges and achieve its business objectives. You can learn more about Mindbody’s thoughtful, phased approach to migration from its re:Invent 2024 presentation.