DeLaval Builds IoT-Based Monitoring System for Cow Behavior Analysis with Knowit on AWS

Sweden-based DeLaval is a global company that provides milking equipment and services for dairy farmers to support sustainable milk production. The company’s aim is to help farmers minimize time-intensive, day-to-day tasks to make the milking process as productive and profitable as possible. It also focuses on animal welfare and increasing the sustainability of milk production. DeLaval is known for its novel products in the dairy industry, such as its Voluntary Milking System (VMS) that allows cows to be milked according to their milk production capacity and needs, and a device that measures progesterone in milk, a naturally occurring hormone used to manage the reproductive cycle of dairy cows.

Founded in 1883, DeLaval today has a global footprint with 4,500 employees and 18 factories worldwide. It focuses its innovation efforts on farms with only a few cows, rising to ones with many thousands of animals. Its key markets include, China, Europe, Japan, the Middle East, and North and South America.

Beyond its core operations, the company wanted to enhance its position as a service provider by building a cloud-based platform to offer new digital offerings. For its first new application on the platform, it wanted to provide near-real-time monitoring of cows, using Internet of Things (IoT) sensor data, to see when they are in heat and fertile, supervise their welfare, and to track animal location, for example, to check if a cow is in the barn or out in the field.

Rather than build a team of in-house developers, which would be time consuming, the company turned to AWS Partner Knowit because of its cloud development and IoT expertise. After evaluating several cloud providers, it decided to develop a serverless architecture on AWS, because of its wide array of available services. “Knowit became embedded in our team very quickly and has gone way beyond being just a contractor,” says Yvo Bogers, director of technology architecture at DeLaval Digital Services. “We quickly realized that we could rely on them to fill the knowledge gaps we had.”

Knowit began work on the new platform in 2020 completing the project over a 3-year period—this was necessary for DeLaval so it could test the efficacy of items such as IoT tags. Data is collected from tags on the cow’s ear by antennae on a customer’s farm. This helps detect changes in animal behavior by tracking multiple variables, such as the type and intensity of movement or activity the animal is engaged in. The application was also designed to track cows’ location using positioning data.

Knowit worked to control costs after the platform prototype proved to be too expensive, by iterating processes and AWS service use. This included more efficient DevOps processes for workloads, database interactions, and log storage. The partner also optimized IoT messaging costs by more efficiently packaging and segmenting data before sending to AWS for processing. “Knowit helped us reduce the cost per cow by 75 percent, making the application more affordable for our farmer customers,” says Bogers. “Tracking the activity and positioning of cows helps farmers focus on essential tasks and keep their costs much lower.”

DeLaval’s in-house team developed machine learning (ML) algorithms that run on AWS IoT Greengrass, to build intelligent IoT devices faster. Near-real-time streaming data is collated and sent to the cloud using AWS Lambda, which runs code without thinking about servers or clusters. Using AWS, DeLaval can also process ML data models at the edge should connectivity be lost. “It’s crucial to deliver important notifications at all times,” says Johan Ribbeklint, business director at Knowit. “If it fails to notify that a cow is in heat, farmers pay a lot more in feed costs.”

The company installs edge servers on each farm—all managed from AWS. When a new device comes online, it connects securely to the platform to download the operating system automatically. This is coupled with a device management service built on AWS using a combination of Amazon DynamoDB—a fast, flexible NoSQL database service for single-digit millisecond performance—with AWS Lambda, and Amazon API Gateway to create, maintain, and secure APIs at any scale. The service is used to provision and deploy configuration packages to the edge for ML models, and to manage database retrieval or dispatch messages.

Because not every farm has a solid network backbone, Knowit worked to optimize bandwidth to improve connectivity when sending data to the cloud. To ensure that DeLaval’s globally distributed customers—including very remote farms—gain high levels of performance, it uses Amazon CloudFront to securely deliver content with low latency and high transfer speeds. Knowit and DeLaval developed a user interface that runs on AWS to provide insights based on the behavior analysis and location data with dashboards, graphs, and data visualizations. This can be tailored to the role and tasks of each individual farm worker.

DeLaval is now looking to develop its platform further on AWS to handle more complex analytics jobs—and it is very happy that its new platform helps farmers to do more with less and control costs. “We were lucky to have such a transparent relationship with Knowit and we were impressed they were so cost aware,” says Bogers. “Using AWS helped us keep our development team focused because you don’t have to build everything yourself. For us, that is gold.”