It’s been over 11 years since AWS began supporting Microsoft Windows workloads. In that time, AWS has innovated constantly to maintain its title as the #1 cloud provider for these workloads. You can run the full Windows Stack on AWS, including Active Directory, SQL Server, and System Center.

Many third parties have completed studies that show why AWS is superior when it comes to performance, cost, and reliability. In 2018, the next-largest cloud provider had almost 7x more downtime hours than AWS. Additionally, SQL Server on AWS boasts a 2-3x better performance record. When costs are calculated correctly, SQL Server running on AWS’s competitor’s platform would be almost twice as much. This includes the cost of storage, compute, and networking.

Reliability is the quality that puts AWS high above the rest. AWS has 64 availability zones within 21 different regions. AWS customers can deploy their applications across multiple zones in the same region for fault tolerance and latency. Instead of having a single-region instance that scales up, AWS’s services are divided into smaller cells that scale out within a region. This design reduces the effects when a cell-level failure occurs. Notably, AWS has never experienced a network event that spans multiple regions.

When migrating your SQL Server Workloads to AWS, there are a few things you should consider. It’s important to optimize your total cost of ownership, which includes optimizing your workloads to benefit from the scalability and flexibility of the cloud. On-premises servers are not optimized, in fact, 84% of workloads are over-provisioned. Many Windows and SQL Server 2008 workloads are running on older and slower server hardware. To optimize your cloud migration, you need to size your workloads for performance and capability, not by physical servers. To reduce cost, you can also decrease the number of licenses that you use by server and core counts.

Another strategy is to decouple your storage and compute processes. When these are combined, they must be scaled together. On the cloud, compute and storage can be separated. Decoupling makes elasticity easier to achieve and manage. Many people question this because SQL Server instances often contain logic to ingest or process data before it is stored in a schema. Many time, ETL logic is written within SQL Server processing engine and the servers are sized to handle a large volume of ETL processes. These ETL processes often run a couple of times a day and the capacity is only needed during the time it is running. By moving the ETL logic outside of the SQL engine, you can utilize the elasticity of the cloud and expand your compute power whenever needed. This will reduce your SQL Server footprint in the long-run. Of course, this doesn’t apply to every use case in SQL Server, but if you have ETL logic, enrichment logic, or load logic inside your SQL Server, decoupling might be the correct choice. This was the case with one of our customers, Telarix. See here to read their white paper.

As part of your migration, you should consider running your SQL Server on a Linux Instance within the AWS platform. The majority of current Windows functionality is supported on the Linux platform. Additionally, there is a minimum of a 20% cost benefit of running SQL Server on a Linux instance! This decision will give you the best performance and save you the most money.

You can also use Amazon RDS to upgrade your database instance to SQL Server. This is performed in place and is initiated with just a couple of clicks. Before you upgrade, you can create a snapshot backup, use it to create a test DB instance, then upgrade that instance to the desired new version. You can also opt-in to automatic upgrades that take place within your preferred maintenance window. 

If you’re considering migrating your Windows Workloads to the cloud, AWS is the optimal choice because of the price, performance, and reliability. This is the perfect time to migrate and modernize your outdated servers. Contact us to learn more or get started on your project.

Share this post on: