Bayer Crop Science Case Study

Bayer Crop Science, a division of Bayer, provides a range of products and services to enable sustainable agriculture for farmers worldwide. The company offers seeds, seed treatments, herbicides, insecticides, and other products to help farmers control pests and weeds and maximize crop production.

As a leading sustainable-agriculture company, Bayer Crop Science continuously strives to give its customers expert advice on where and when to plant its products. To do that, the organization supports a large geospatial-data platform containing sets of vector and raster data, soil and elevation layers, maps, and digital aerial imagery of farms. The company performs analytics on these data sets for internal data scientists and external business partners. After analyzing the data sets, Bayer Crop Science recommends specific types of hybrid seeds for use in planting and harvesting in various farming environments.

As demand for the geospatial platform grew, Bayer Crop Science was concerned that its on-premises data center was not able to scale to support the expanding system. “The platform is growing very quickly,” says Vishnu Alavur Kannan, analytical technical platform lead at Bayer Crop Science. The organization also needed to better manage the “bursty” nature of the platform. “Our workloads are seasonal, based around planting cycles, and we’ve always had to plan for that by using compute to match traffic spikes,” says Kannan. “We wanted more elasticity, so we could do that in a faster, more cost-effective way.” Additionally, the company sought to give its developers a way to create new software features more quickly. “We wanted to accelerate development by enabling developers to rapidly test and launch new features,” Kannan says.

To support these requirements and scale the geospatial platform more easily, Bayer Crop Science wanted to implement a cloud-based environment for the platform.

As part of a cloud-first strategy, Bayer Crop Science had already turned to Amazon Web Services (AWS) and moved several internal applications to the AWS Cloud, running on Amazon Elastic Compute Cloud (Amazon EC2) instances. “We had already made the decision to migrate key applications to the AWS Cloud,” says Kannan. “AWS is the leader in the cloud space, and we wanted to continue adding more AWS services.”

Bayer Crop Science then decided to deploy a shared file system to store all the data its applications require. After exploring and testing several file-sharing systems, the company implemented Amazon Elastic File System (Amazon EFS), a service that provides scalable file storage for use with Amazon EC2 instances. Using Amazon EFS, Bayer Crop Science can automatically expand and shrink its geospatial data platform by taking advantage of the service’s elastic storage capacity. Amazon EFS also provides scalability, so Bayer Crop Science can scale performance for its geospatial database platform automatically as it grows.

Using Amazon EFS allows Bayer Crop Science data scientists and developers to share large geospatial data sets internally. Using the data stored in Amazon EFS, Bayer Crop Science data scientists and developers perform highly parallelized analytics, enabling more detailed prescriptive recommendations for the use of the company’s seeds. For example, Bayer Crop Science can better manage interactions between its different growing zones and can also better predict the ranking of hybrid and inbred seeds within and across different customer growing and harvesting environments.

Using Amazon EFS to process and store data, Bayer Crop Science has a better way to grow its large-scale geospatial system. “We can now run analytics at scale for our geospatial data platform using Amazon EFS,” says Carrington Wong, platform engineer for Bayer Crop Science. “We can provision compute and storage in seconds, which means we can quickly expand the platform to meet the growing demand for analytics. We can also use the elasticity of the cloud to automatically scale up and down in support of our seasonal workloads.”

By relying on Amazon EFS, Bayer Crop Science can also quickly scale storage and throughput, enabling better performance for its platform, and it has been able to achieve consistent latencies and higher availability. In addition, Bayer Crop Science’s system is now easier to manage. “Amazon EFS gives us a shared file system in the cloud, with minimal setup and no management required,” says Kannan. “As a result, we have a much more efficient way to manage our petabyte-scale geospatial data platform, enabling analytics at scale and integrating analytics with product platforms. These capabilities are especially important for us because we have internal resource constraints.”

The company has also been able to speed the development and test cycle for its developers. “We can experiment with new feature development much faster now using AWS,” says Wong. “It takes us just a few minutes to stand up a complete end-to-end dev and test environment, where before it would take us an average of four months to do it because we had to procure, provision, and configure hardware.”

By avoiding the need to run its geospatial platform out of a data center, Bayer Crop Science has saved time and money. “Before we went to AWS, we had to have an entire team of people working on setting up hardware and deploying and configuring it,” says Kannan. “Now, because everything is in the cloud, they can spend their time focusing on more important corporate initiatives. And we also save significant amounts of money that used to be tied up in hardware and maintenance.”

Bayer Crop Science will also be able to more effectively grow its business by using AWS. “We are taking advantage of the global presence of AWS to more easily expand throughout the world,” Kannan says. “We can take our applications and services closer to our business users in more regions. Our business growth overall will be simpler and faster because of the AWS Cloud.”