Ginkgo Makes Biology Easier to Engineer and Scales Cost Efficiently Using AWS Batch

Ginkgo was founded by five scientists from the Massachusetts Institute of Technology (MIT) in 2008, and it uses synthetic biology to provide biological solutions for companies across a wide variety of sectors. For example, Ginkgo’s customers apply its solutions to further drug development or reduce the need for fertilizers in agriculture. Initially, Ginkgo relied on on-premises data centers to power its solutions. However, as the company grew, it was clear that a new approach to high performance computing (HPC) was needed. Because so much of Ginkgo’s work involves biological research and development, the company’s need for computing power to deliver high-quality biology services is highly variable. “Cells just divide. That’s the original exponential growth,” says Chris Mitchell, principal software engineer at Ginkgo. “If you’re dealing with biology, you need to accommodate scale.”

To unlock scalability while freeing up time, Ginkgo migrated its systems to AWS Batch, a service to efficiently run hundreds of thousands of batch computing jobs. “One of the main drivers behind our cloud migration was that we could not handle the data load locally and do all the analysis that our customers need without spending significant resources and time upfront,” says Duran. To get the most out of its services, Ginkgo communicated with AWS to discuss architecture, refine solutions, and optimize spending. Now, Ginkgo uses AWS services such as AWS Batch to take care of the undifferentiated heavy lifting so that Ginkgo teams can focus on delivering high-quality scientific research and engineering at scale.

Along with AWS Batch, Ginkgo uses Amazon Elastic Block Store (Amazon EBS), a scalable block storage service, to optimize its storage volumes. The company runs immense HPC workloads to discover the best way to program organisms, and it uses Amazon EBS to store the petabytes of data it needs for analytics. “This year, we’ve saved up to 90 percent of our Amazon EBS costs by reducing the size of the volumes we use,” says Dima Kovalenko, manager of cloud infrastructure at Ginkgo.

In addition to AWS Batch and Amazon EBS, the company uses several other services to achieve flexibility. For example, it uses Amazon Simple Storage Service (Amazon S3), a highly secure and available object storage service, to store data cost efficiently after experiments have run their course. More specifically, Ginkgo uses Amazon S3 Intelligent-Tiering (S3 Intelligent-Tiering), a service that automates storage cost savings by moving data when access patterns change. “Using S3 Intelligent-Tiering, we’ve saved up to 60 percent of our storage costs,” says Kovalenko.

With its solutions running on AWS, Ginkgo’s teams can now focus more closely on science. Rather than dealing with infrastructure management, the company can turn around projects and provide its customers with answers as quickly as possible. Ginkgo has developed an efficient platform that automates much of the lab work that biotechnology companies would otherwise have to do for synthetic biology. “When people come to us with an exciting idea, we empower them to do rapid experimentation to see if their idea is worth pursuing further,” says Kovalenko. “We don’t have to wait 6 months to set up a data center, and that makes a huge impact for our customers.” Using AWS, the company has met its goal of scaling to meet growing customer demand.

Now that its systems are on AWS, Ginkgo can scale to onboard many more customers on its biology platform. “Since going public, we have to comply with Sarbanes-Oxley regulations, and many of our customers have specific security requirements for their programs,” says Duran. “Without migrating to AWS, we couldn’t have accommodated the needs of our internal teams nor those of our customers cost efficiently.”