Matterport Uses AWS to Offer 3D Experiences and Reduce Compute Costs by 70%

Matterport, a software-as-a-service (SaaS) provider, helps professionals in a range of industries create better visual experiences by creating 3D models of built spaces. Customers use Matterport’s Pro2 camera or compatible third-party 360 cameras to take digital images of spaces, capture visual and depth data, and upload the images to the Matterport Cloud to create detailed 3D virtual models. Using these virtual models, everyone from architects to real estate professionals to insurance appraisers can take advantage of the most accurate depiction of a physical space.

Matterport’s application requires a large amount of compute capacity. “We are typically processing thousands of customer models per day, and the volume is continuously growing, so we needed the ability to scale to support that,” says Simon Wynn, senior vice president of software engineering for Matterport. Since its founding in 2011, Matterport has used Amazon Web Services (AWS) to enable scalability, relying on Amazon Elastic Compute Cloud (Amazon EC2) for on-demand compute capacity and a proprietary batch-processing solution to process customers’ 3D models.

As the company’s business has grown, however, Matterport needed even more scalability. “Our business is growing quickly,” says Leo Sutic, principal engineer for Matterport. “Our systems were not scaling well and were not durable enough to satisfy our customers’ requirements for fast, reliable processing of their models.” Matterport also sought to reduce costs as data volume expanded. “We spent a lot of time maintaining our proprietary environment, which took too much time and added to our costs,” Sutic says.

To solve its scalability and cost challenges, Matterport began using a microservices architecture based on the Kubernetes container-orchestration system running on Amazon Elastic Kubernetes Service (Amazon EKS).

More recently, the company started relying on AWS Batch job queues to process 3D models uploaded to AWS. “Using AWS Batch, we have automated the uploading and assigning of job queues for 3D models,” says Ameet Patel, director of cloud infrastructure engineering for Matterport. The company also started running its Matterport Cloud 3.0 application environment on Amazon EC2 Spot Instances to reduce costs. Amazon EC2 Spot Instances are spare compute capacity that is available at large discounts compared to the cost of On-Demand Instances.

As a result of using AWS Batch and Amazon EC2 Spot Instances, Matterport now has the scalability it requires to meet customer demand. “Without the automation of AWS Batch, it would have taken us more than a year to create our containerized application environment,” says Patel. “Instead, it only took a few months. Now, relying on AWS Batch, we can automatically process tens of thousands of 3D models a day without our customers having to wait a long time for their completed models.”

Using Spot Instances, the company has achieved its goal of lowering its compute costs. “We have reduced our costs by 70 percent using Amazon EC2 Spot Instances, which saves us more than $1 million annually,” says Wynn.

Because it is spending less money on compute, Matterport can offer more flexible pricing to its customers. “The reduced costs we achieve using Amazon EC2 Spot Instances help us directly improve our business,” Wynn says. “For example, we recently introduced a freemium pricing model, including a free tier where customers don’t pay fees when creating 3D models. We have also begun offering support for third-party 360 cameras including the Ricoh Theta V and Insta360 One X. We could not have offered these new capabilities without the cost savings we realized through our use of AWS.”

By taking advantage of AWS, Matterport can continue to improve service delivery for its application. “We have more and more international customers, so our platform needs to be up and running 24/7,” says Wynn. “As we keep moving into a containerized approach running on AWS Batch and Amazon EC2 Spot Instances, we will be able to optimize the performance of our platform while ensuring the highest availability and also reducing costs. We can better serve our customers and grow our business by using AWS.”

To learn more, visit thinkwithwp.com/websites.