Software Delivery Case Study
Increasing Field Efficiency During Harvesting
Do we need some additional content here? Idk, it might be cool, but not entirely what we should put in this very important sentence.
THE PROBLEM
Today’s status quo agricultural practices won’t feed the world of tomorrow. A Midwest agricultural leader set out to create more efficient and intelligent field operations.
By 2050, farmers will need to grow around 70% more food in order to feed a growing, global population of more than 10 billion people. If that wasn’t ambitious enough, it becomes an even tougher target to hit as the number of farmers and farmland decline each year.
THE OPPORTUNITY
Feeding the population of 2050 will require farmers to make their farm operations as efficient and data-driven as possible.
While producers are capturing thousands of data points related to their farming practices every day, much of this data remains unused. By transforming this raw data into reliable, real-time insights, farmers can increase operational efficiency, and ultimately, maximize crop yield.
THE SOLUTION
The Midwest agricultural leader partnered with Lean TECHniques to deliver an elastic cloud software solution that could ingest, process and share data back to the machine within three seconds, allowing on-board displays to be updated in near real-time and accessible across machines.
Crop producers are able to see where all machines are within the field and view load capacity information, enabling faster unloading times and smoother harvesting logistics.
The solution was built using AWS Simple Queue Service (SQS) and EC2 instances within elastic scaling groups. EC2 autoscaling rules allowed the system to easily scale up or down as needed to handle fluctuating data loads, creating both a responsive and cost-effective solution. Additionally, the ELK stack was used to enable a high level of system observability to detect and resolve issues before customers would be impacted.
Accelerating Speed to Market
In order to keep the risk of impacting the system and customers low, changes were kept small and a continuous delivery pipeline was implemented. The CD pipeline enabled validation to occur for each change automatically before deploying it to production where similar validation occurred.
Real-Time Data Processing
In order to create a tool that has maximum adoption, it needs to be easy-to-use and respond quickly. AWS Kinesis-based architecture was already in place, but load testing and real-world simulations demonstrated that the current architecture would not allow for data to be ingested, processed and sent back to the farmer within three seconds.
Because of that, a proof-of-concept was delivered using AWS Simple Queue Service (SQS) and EC2 instances, which successfully met the project goals and helped avoid the start-up overhead of AWS Lambda, achieve low latency and provide nearly unlimited capability for parallel event processing.
Ongoing Performance Monitoring
Performance and uptime is critical when it comes to making informed decisions while farming. In order to uphold quality, simulators continuously run against the production environment and trigger alerts for any potential outages.
THE IMPACT
Crop producers can now quickly see which parts of a field have been tilled and planted, how much has been fertilized, and what has been harvested. It’s a line of sight that wasn’t available previously and is enabling farmers to make more informed decisions, improve efficiency and help feed more people.