Scaling Up to 30% While Reducing Costs by 20% Using AWS Graviton3 Processors with Instructure

Instructure provides learning management solutions for higher education and K–12 schools worldwide. The company offers various digital tools for collaborating through videoconferencing and online discussions. Students can manage their calendars, read course content, and submit assignments. Teachers can grade the work on the same platform and submit feedback.

Instructure is a cloud-native company, having chosen Amazon Web Service (AWS) for its reliability, global reach, and sustainability, says Pendleton. “We saw the value of the cloud from the beginning and moved in that direction.” Instructure is running on Amazon Elastic Compute Cloud (Amazon EC2) instances, which provide secure and elastic compute capacity for virtually any workload. When online learning increased during the COVID-19 pandemic, Instructure began to explore using AWS Graviton–based Amazon EC2 instances, powered by custom-built AWS Graviton processors, to deliver high performance at a lower price for cloud workloads.

Instructure first migrated its compute-intensive workloads to AWS Graviton2 processor–based Amazon EC2 C6g Instances, which optimize for both higher performance and lower cost per vCPU. The migration from Amazon EC2 C5 Instances was seamless. The primary programming languages used by Instructure, Ruby and Java, support Arm-based instances. Hence, there were no source code changes required. When AWS launched AWS Graviton3 processors in 2022, Instructure performed load tests on AWS Graviton3 processors that are based on Amazon EC2 C7g Instances. These offer up to 25 percent better performance over the sixth-generation Amazon EC2 C6g Instances based on AWS Graviton2 processors. The load tests assessed the new instances’ cost and performance benefits, and the results compelled the company to migrate to AWS Graviton3–based instances.

Overall, Instructure observed up to 30 percent improved performance by migrating to AWS Graviton–based instances. “We saw better 99th percentile performance during load testing of the Amazon EC2 C7g instances, which led to lower error rates. That kind of consistency and reliability is meaningful to us and our customers,” says Pendleton.

Instructure uses AWS Graviton processors to scale its solution and uses Amazon EC2 Auto Scaling, which makes it possible for users add or remove compute capacity dynamically to meet changing demand.

On serverless AWS solutions, Instructure streamlines its infrastructure management to further optimize the way that it uses compute power. The company uses AWS Fargate, a serverless, pay-as-you-go compute engine for building applications. Instructure also uses AWS Lambda, a serverless, event-driven compute service to run code for nearly any type of application or backend service.

After migrating to AWS Graviton3 processors, Instructure saw a 30 percent boost in throughput performance and improved load performance running on Amazon EC2 C7g Instances over Amazon EC2 instances not based on AWS Graviton3 processors. “Migrating to AWS Graviton3 processors has helped us save costs on scaling while empowering us to offer our users a smoother and faster experience,” says Pendleton. The company achieved up to 20 percent better performance from its application servers while running fewer instances at peak times. The organization also observed that the Amazon EC2 C7g Instances were delivering better results against their cost, which was reduced by 15–20 percent. “These cost savings mean that we can invest in more novel, interesting solutions, like new data services and machine learning. Our engineers can also spend less time doing mundane tasks and more time innovating to benefit customers,” says Pendleton.

Instructure could also manage more requests while reducing its response times from 1.5 seconds to 500 ms using the Amazon EC2 C7g Instance clusters. As a result, millions of concurrent users can complete tasks with less interruption. “We’re able to take that in-person student-teacher experience and either extend it or, where needed, replace it,” says Pendleton.

Instructure plans to migrate its remaining databases running on older instance types to AWS Graviton3 processors. The company is reinvesting its savings from Amazon EC2 into developing data services on AWS that give customers insight into at-risk students so that it can engage them proactively. To do so, Instructure expects to expand its use of Amazon Simple Storage Service (Amazon S3)—which offers industry-leading scalability—and add additional AWS services, such as Amazon Redshift, which offers cloud data warehousing, and Amazon EMR, a cloud big data solution for petabyte-scale data processing, interactive analytics, and machine learning.

“AWS has consistently been a fantastic vendor for us. It is flexible and responsive,” says Pendleton. “Working alongside AWS, we can build solutions that meet our customers’ needs.”