Knewton's goal is to personalize learning for students around the world. Education companies use Knewton technology to power courses that dynamically adapt to each student’s unique needs. By analyzing data to figure out what a student knows, Knewton recommends what to study next, helping more students master material and get ahead. Knewton-powered analytics identify knowledge gaps and predict performance to help educators, parents, and administrators better support every student.
Knewton has offices in New York City and London.
Founded in 2008, Knewton has developed a heavy-duty infrastructure that powers adaptive learning for publishers and education companies. The company has been on Amazon Web Services (AWS) since its inception. As Chief Technology Officer (CTO) Ryan Prichard explains, “Developing on AWS kept our costs low for provisioning and maintaining the environment. We wanted to be on the cloud so we could expand and support a large volume of students and teachers during peak times like back-to-school season. AWS was the clear leader in the cloud space.”
Knewton launched an enterprise-ready platform, which analyzes and normalizes anonymous learning data, enabling publishers to create more effective digital course materials that continuously adapt to each student’s unique needs. Over six million students will use Knewton-powered products in 2014. Knewton is quickly expanding to support new subject areas, grade levels, and classrooms; the company currently has partnerships across the U.S., Europe, East Asia, and beyond. As Knewton continues to grow, it’s critical that the organization can quickly scale its platform to provide a reliable experience for students and teachers around the world.
Why Amazon Web Services
Prichard comments: “If we weren’t running on AWS, it would have slowed down our growth. We would have missed business opportunities and key features wouldn’t have been developed early on. We wouldn’t have been able to get to the place we are now — with partnerships with 19 publishers, education companies, and institutions around the world — if we had to hire a huge in-house operations team to operate multiple data centers.”
Analyzing large data sets
The adaptive learning platform analyzes massive amounts of anonymized data, as students use Knewton-powered products for extended periods of time while studying at home and in the classroom. Knewton uses Amazon Elastic Map Reduce (Amazon EMR) to analyze large data sets across the platform in parallel in order to calibrate its recommendation and analytical models. This type of analysis is done periodically; by leveraging Amazon EMR, Knewton’s teams don’t have to deal with the additional cost and operational overhead of managing their own Hadoop cluster.
The company uses AWS CloudFormation to provision more than 1,000 Amazon Elastic Compute Cloud (Amazon EC2) instances. “We moved about 80 percent of our instances to Reserved Instances, which has cut our monthly bill almost in half,” says Prichard.
Amazon Simple Storage Service (Amazon S3) provides storage for configuration files and backups. Elastic Load Balancing distributes traffic among instances. Figure 1 demonstrates Knewton’s adaptive learning platform on AWS.
Figure 1. Knewton Adaptive Learning Platform on AWS
By running on AWS, Knewton can scale to manage large traffic spikes. “At the beginning of the school year, we’ll see students using Knewton-powered course materials all at once,” says Prichard. “One of the best things about AWS is the ability to scale up quickly during busy times with very little advance notice then scale back down once traffic resumes to normal.”
The company also takes advantage of the AWS Cloud to make informed infrastructure decisions at scale. When deciding on the technology behind their next-generation graph database, Knewton was able to first prototype a small cluster to validate the functional feasibility of the solution. Their team then scaled that cluster up to the size of a fully provisioned production cluster — ensuring the benefits of the new solution would be realized at the scale forecasted over the next 12 months. At the completion of the functional and performance tests, all database instances were decommissioned. This relatively small investment helped make Knewton much more confident in a critical infrastructure decision.
From a business perspective, AWS has allowed Knewton to meet the demands of students and partners, and expand into new markets to become the leading adaptive learning platform. “We’re able to work with partners all over the world while running our platform with AWS,” says Prichard. “AWS offers an infrastructure that is compliant with data regulations in various countries, which is essential as we already have partners spread across four continents. Being able to deploy in an Amazon Virtual Private Cloud (Amazon VPC) and operate in multiple regions has been a huge strategic advantage for us.”
“The AWS team has been a Knewton partner since the early days of our organization,” says Prichard. “AWS has provided a reliable infrastructure and helped us power materials for education companies internationally. I look at the AWS team as a valuable partner in Knewton’s growth.”