Monsanto finds room to grow with open source software

Farming may be the oldest industry in the world, but the rise of the internet of things (IoT), mobility and other technologies present new challenges for agricultural companies. Scrambling to keep up with the explosion in crop data, Monsanto has turned to an unlikely source for help: open source technology.

Like most agricultural companies, Monsanto has traditionally licensed geospatial information software (GIS) to visualize crop performance and yield information on a map, and to process satellite imagery to understand where disease or blight might be occurring. But the proliferation in data volumes from yield monitors, precipitation sensors, crop health imagery and other sources meant Monsanto would have to buy more enterprise software licenses, says Martin Mendez-Costabel, who leads Monsanto’s geospatial big data engineering and strategy efforts.

Such an investment would prove costly and unsustainable at a time when thousands of users require access to geospatial software from desktop and mobile devices, Mendez-Costabel says. Moreover, with the software scattered all over the enterprise, it proved difficult to manage. Craving more flexibility at a lower cost, Monsanto two years ago began using open-source GIS software from Boundless, a startup whose software helps manage data generated by satellites, GPS systems and a variety of sensors.

Open source helps Monsanto scale

“You can scale up without paying the penalty of spending more money and negotiating additional licenses and installing and configuring them,” Mendez-Costabel tells CIO.com. “Cost savings is a driver, but there is also more flexibility in terms of how much you can customize different tools and analytics models according to your needs.”

Monsanto’s migration to Boundless is part of an open-source renaissance the company has enjoyed since CIO Jim Swanson joined the company from Merck in 2013. Swanson has encouraged engineers to explore alternatives to enterprise applications whose scale-up costs can tax finite IT budgets. Open source technologies, including Hadoop, Kafka and TensorFlow, underpin Monsanto’s science@scale platform, which runs simulations against millions of data points for characteristics such as seed genetics, climate, water, soil and nutrients.

Boundless adds another platform to the Monsanto big data arsenal. Previously, GIS data had been scattered across Monsanto’s R&D and its commercial business units, making it difficult to access and manage crucial information about the best planting plots, soil, elevation, weather and a host of other environmental factors.

Today Boundless allows Monsanto business analysts, agronomists and GIS engineers to pull information about soy, corn and other crops from sensor-fitted farming machines, as well images taken by satellites and drones. The information enables Monsanto to predict the best place to plant crops, which is crucial information in the run-up to planting season. Boundless, which runs on Amazon Web Services cloud infrastructure, provides a “one-stop shop” for Monsanto’s GIS needs, Mendez-Costabel says.

Monsanto’s adoption of Boundless endured its challenges. Mendez-Costabel says some engineers were initially skeptical about using an unfamiliar technology. “Almost 100 percent of the people said, ‘I used to do it this way, how do I do this now,'” Mendez-Costabel says. “There’s a learning curve in an enterprise environment; you can’t just adopt open source and expect everything will be smooth from day one.”

Navigating learning curves for value creation

Training and feedback sessions helped Monsanto even out the learning curve, Mendez-Costabel says, adding that he and his staff spent weeks huddling with users on the platform. Mendez-Costabel says he traveled to South America, Europe and other regions to help teach employees how to use the Boundless technology. In turn, some of the trainees became trainers. To assist with the initiative, Boundless allocated three GIS engineers to work with Monsanto’s IT and commercial business staff on the deployment. The extensive training paid off; Monsanto launched the new platform in about eight months.

The partnership has paid off. In December, Monsanto and Boundless contributed open source geospatial code to the Free and Open Source Software for Geospatial (FOSS4G) community. Among the contributions is an OAuth plugin that determines what data users within the Quantum GIS open source desktop project can access. Monsanto has used the code to federate user access within its own GIS system, enabling employees in the field to access data more easily.

Mendez-Costabel says he expects to extend Monsanto’s work with Boundless to incorporate more mobile development capabilities, such as a software development kit (SDK), which ideally would help developers extend GIS data to far-flung places where internet and cellular access are hard to come by.

On a recent webinar discussing the Monsanto deal, Boundless CEO Andy Dearing says the market for GIS systems is expanding from professionals trained to work with the technology, to business analysts and DevOps engineers who require location data to do their jobs. “They’re not trained in GIS, but they realize that location or a derivative of GIS is important to their job function,” Dearing says.

Monsanto is among several enterprises, including Walmart, MetLife and the Bank of New York Mellon, diving deeper into open source, as the trend continues to climb. More than 90 percent of IT organizations worldwide use open source software within mission-critical IT workloads, Gartner analyst Mark Driver wrote in a March 2017 report. “It’s happening more and more and I think it’s going to be a trend,” Mendez-Costabel says.

Related digital transformation articles: