Microsoft Workloads on AWS

How Autodesk transformed a “good enough” app into “great”

Too often, “good enough” simply isn’t “good enough.” This is especially true for a large multinational company like Autodesk with ambitious growth plans. Autodesk is a leader in 3D design, engineering, and entertainment software. The company started its cloud modernization journey years ago, moving workloads from private datacenters to Amazon Elastic Compute Cloud (Amazon EC2), among other AWS services. One of the applications it migrated was Autodesk single sign-on (SSO). This is a mission-critical service integrated with all Autodesk products. It serves over 142 million users and responds to over 145,000 API requests per minute.

Autodesk needed to modernize to gain flexibility and scalability to support anticipated growth so moved to fully managed Amazon Aurora MySQL. This shift helped streamline the management and resiliency of the Autodesk SSO service, optimize costs, and reduce overhead for infrastructure maintenance.

Taking modernization to the next level

Data is the primary fuel of digital transformation, which is why Autodesk began collaborating with AWS years ago to move database workloads to AWS. For example, Autodesk recently moved their Autodesk Access Control Management (ACM) application from managed MySQL databases running on Amazon RDS and self-managed MySQL databases hosted on Amazon EC2 to Amazon Aurora MySQL. This change resulted in a 10x reduction in CPU utilization, from peaks reaching 100% with MySQL to less than 10% with Amazon Aurora, and a 2x improvement in response time.

With that modernization success behind them, Autodesk was ready to tackle SSO.

Although Autodesk achieved significant growth with AWS, there was room for improvement. After a few years, the complexity of their self-managed database infrastructure started to present challenges such as difficulties conducting root-cause analysis of incidents they might be facing.

Although they automated many of the necessary steps, Autodesk’s backup management is self-managed infrastructure and therefore imposed significant overhead, especially because of their cross-region setup. Patching was also a significant time sink for the database engineers: They had to support multiple environments, including testing, development, and production. Finally, provisioning new nodes in Autodesk’s self-managed infrastructure was a time-consuming process, making elasticity an elusive goal.

Something had to change, but the risks of change seemed daunting. Autodesk saw the value of a fully managed database service, but such a migration would involve moving a large amount of data from the source database to a target database. SSO was used by millions of users 24/7/365 and anything that threatened to disrupt the customer experience had to be approached with care. This application was mission-critical; there was no room for error. This potential risk, however, was outweighed by the upside associated with modernizing with AWS and improved experience for Autodesk customers.

The benefits of modernizing with AWS

Autodesk knew they needed to modernize in order to remain competitive and best serve their customers. It helped that they already had good experience with AWS, with extensive partner support available. As Tulika Shrivastava, a software architect at Autodesk, notes, “Our entire ecosystem was already present on AWS.” The fact that an ecosystem of unparalleled depth and breadth was already running on AWS gave Autodesk confidence to move forward, and the company derived near-immediate benefits. “Moving to a fully managed database service like Amazon Aurora MySQL helped us remove all this infrastructure maintenance overhead,” notes Shrivastava. “It also gave us the ease with which we could scale up and down our infrastructure based on the load.”

Let’s dig into scalability for a minute.

Generally in Always On setups, the primary node processes the read-only routing request. This feature ensures that the requests always goes to a healthy secondary node but in turn it makes the primary node a bottleneck for scalability.

Amazon Aurora MySQL uses read and write cluster endpoints for read and write instances. This makes it easy to scale without any single instance becoming a bottleneck. For these reasons, Shrivastava continues, “Moving to [Amazon] Aurora [made it easy to] scale up and down the database infrastructure, which gave us the flexibility to use lower[-cost and capability] instance types in our pre-production environment, and scale them up with a click when needed.”

Another reason for migrating to Amazon Aurora MySQL was cost. According to an initial cost analysis, the company would save approximately 40% to 50% of overall database costs every month using Amazon Aurora MySQL.

And what about the migration risks that previously concerned Autodesk?

According to Shrivastava, Autodesk was able to migrate to Amazon Aurora MySQL with ease. Autodesk’s migration plan involved schema migration, application migration, data migration, and a rollback plan in case things went wrong. To ensure this went smoothly, Autodesk relied on the AWS Database Migration Service, partnering closely with AWS’ migration experts.  There were further advantages to be gained in the process, as Shrivastava explains: “Due to the choice of the character set for our schema migration, we were able to reduce the database size considerably. Because of that, we were able to optimize the instance type further and save costs.”

By going all-in on their modernization efforts with AWS, Autodesk dramatically lowered the overhead associated with managing their database infrastructure, improved scalability, lowered costs, and achieved significant performance gains.

What are you waiting for?

Given the benefits Autodesk and others have experienced, what’s your plan for embracing transformation and modernizing your infrastructure?

Read about how other companies are modernizing on AWS as we regularly showcase the different approaches and strategies companies are using to transform their infrastructure.

Let AWS help you assess how your company can get the most out of cloud. To have us create an assessment, email us at WhatsYourModernizationPlan@amazon.com and please consider joining the conversation using the #WhatsYourModernizationPlan hashtag. Visit Windows on AWS to learn more.

Matt Asay

Matt Asay

Matt Asay (pronounced "Ay-see") has been involved in open source and all that it enables (cloud, machine learning, data infrastructure, mobile, etc.) for nearly two decades, working for a variety of open source companies and writing regularly for InfoWorld and TechRepublic. You can follow him on Twitter (@mjasay).