Optimizing the ROI of a geospatial platform in the cloud
Bayer Crop Science stood up a cloud-enabled open source geospatial platform almost two years ago. Ever since it went into production, use of the platform has steadily grown worldwide and the full value and potential of the platform has become abundantly clear. The platform is now servicing 300+ internal and external API clients regularly, with peak usage of 900 million calls per month while exposing 17 billion acres worth of global geospatial data assets to the enterprise. (You can find detailed information on the platform in “The Enterprise Geospatial Platform: A Perfect Fusion of Cloud and Open Source Technologies,” a talk by Naghman Waheed and Martin Mendez-Costabel at Strata San Francisco 2017.)
The engineering team quickly recognized the unaccounted increase in cost to operate the system, due to ever growing number of users and datasets. To address that challenge, the engineering and solution delivery team partnered to rearchitect parts of the platform with the goal of reducing costs while maintaining or even improving performance, SLAs, and overall platform health.
Martin Mendez-Costabel explains how Bayer Crop Science manages its geospatial platform and how it increased ROI. The data repository, which initially consisted mostly of EFS and Postgres, has now been replaced with low-cost storage components such as S3 and Apache Accumulo. The geospatial processing engine now includes not only Geoserver but also GeoMesa and GeoTrellis to account for performance improvements while optimizing cloud costs. The implementation of vector and raster tiles to compliment OGC standards already used before, plus the implementation of a multicloud strategy (AWS and Google Cloud) has resulted in increased cost optimization while also enhancing the user experience. The platform is also now integrated with data lake solutions to provide off the shelf and low-cost business intelligence alternatives. (For more information, see “You Call It Data Lake; We Call It Data Historian,” a talk by Naghman Waheed and Brian Arnold at Strata London 2018.) The result of these efforts has been that the yearly cost to operate the platform is almost at least 40% less in terms of yearly cloud costs.
Bayer Crop Science
Martin Mendez-Costabel leads the geospatial data asset team for Monsanto’s Products and Engineering organization within the IT Department, where he drives the engineering and adoption of global geospatial data assets for the enterprise. He has more than 12 years of experience in the agricultural sector covering a wide range of precision agriculture-related roles, including data scientist and GIS manager for E&J Gallo Winery in California. Martin holds a BSc in agronomy from the National University of Uruguay and two viticulture degrees: an MSc from the University of California, Davis, and a PhD from the University of Adelaide in Australia.
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires