How Farmers.gov really launched USDA’s next generation of cloud services

Casey Cook, the cloud architecture branch chief for the Agriculture Department, said seven mission areas are taking advantage of more than 300 APIs to make data...

The Agriculture Department found a way to stop reinventing the cloud computing wheel.

Starting with the revamping of the Farmers.gov platform, USDA figured out how create a repeated process through the use of application programming interfaces (APIs) to modernize systems.

Casey Cook, the cloud architecture branch chief for the Agriculture Department, said this approach is helping USDA reduce its technical debt.

Casey Cook is the cloud architecture branch chief for the Agriculture Department.

“The idea really got to be, ‘well, we could do this together and adopt a pattern that’s repeatable and reusable,’” Cook said on Ask the CIO. “That was 2017 or 2018, and now we support seven mission areas out of the eight in the department. We have a few hundred production APIs deployed.”

Cook described APIs as the plumbing of how you move data between systems. He said APIs help mission areas get a hold of the data that is valuable to helping them make decisions, and do it in a way that minimizes the need for human interaction.

The initial test case of applying these APIs was with Farmers.gov, for which USDA won $10 million in 2018 from the Technology Modernization Fund, to integrate data across three layers — business, process and experience — using this approach. The TMF website says USDA ended up only receiving $4 million for Farmers.gov and the agency is now in repayment status.

Cook said APIs helped integrate multiple systems and modified the data delivery to bring all the information together through the portal. USDA ended up consolidating seven platforms and 150 other web resources to create Farmers.gov

“You’re connecting multiple system APIs to execute data enhancement or data revision to get you another business outcome. That’s really the experience API; that is the piece where it goes to  Farmers.gov, where they took these multiple systems, integrated them and then modified it to give it the payload to the API on what Farmers.gov was expecting,” he said. “They were able to have that pattern even though the systems weren’t so tightly coupled. We could potentially bring in a modernized solution because the experience API still needs that same outcome and we only have to modify the bare minimum and increase that reuse. That really increases our ability to modernize because we can do a system-by-system modernization without impacting our whole API network case.”

Cook said Farmers.gov runs on the Salesforce platform through a software-as-a-service instance, while the data resided on both cloud systems and on-premise data centers.

“We were able to leverage the experience API platform to aggregate that information from all these disparate sources and provide the Salesforce front-end one integration for all of its data. We had this dependency that we created so we didn’t have to open every USDA port and protocol system to allow access to Salesforce. We’re able to aggregate it through this method to help ensure some consistency, some visibility as well as some cyber protections,” he said. “Then we’re able to really keep a good eye on some of those integrations and how the authorization models and things of that nature work, as opposed to having to independently manage each of those 200 source systems. That’s some of the real power that you start getting, is we’re able to manage it using a public key infrastructure (PKI) type system or JAVA web token authentication or other various technologies that we have. That flexibility to adapt the experience API system and those APIs to fit the needs of the various different sources can really let the mission areas bring whatever they have and we can help tie it together for them.”

Reducing data duplication

The success of using APIs for Farmers.gov opened the door to other modernization efforts in the cloud and with legacy systems.

Cook said once you define the API patterns, build dependencies on them that are known interactions, USDA can achieve its goal of a more data driven modernization effort. The experience API is reducing the amount of data duplication happening across the department and creating that “single source of truth” for information.

“APIs are really one of the vehicles that help mission areas adopt cloud in a more effective manner, while still having legacy systems and having that ability to have that hybrid approach of keeping the data where it is. It has really been one of the most attractive pieces of how APIs have been adopted in the department,” he said. “We need to do a new thing, but we have to integrate it with all these older systems. We need to have a way to marry up all these different programs and efforts, and that API pattern really enables that … creating a flexible and reusable pattern.”

The added benefit of creating that flexibility and reusability is the consistency APIs bring across disparate mission areas and that can lead to better overall customer or citizen experience.

Cook said mission areas from Rural Development to those bureaus that do mortgage underwriting all can use the API to publish or obtain data.

Business outcomes drive modernization

“In the future, I expect USDA to start having data sources and APIs that are publicly available that we could publish for the American public to pull data in the right use cases where mission areas might have that need. We could publish those so user communities could then interact with us and start having those interactions with apps on various systems that they may need on the other side,” he said. “We’re really trying to be sensitive to what our mission areas need, and try to predict what new capabilities they may need. I think it’s really going to start interacting more with artificial intelligence and machine learning and that intelligent automation, robotics process automation areas, where we can really start increasing the capabilities and decreasing the data spread. The data domain is really valuable to the organization and to a lot of our customers. Decreasing that spread and really enabling the consumption of that data in a more systematic way is probably the best for the agency and our customers.”

Cook added his office already has been working with some USDA customers to move toward a containerized platform for AI.

As other agencies develop an API-first strategy, Cook said they should keep in mind that the business outcomes must drive their efforts.

“What do they need and how can you help them accomplish that outcome? Don’t choose the tool and force everybody to use it. That’s not our goal. We were partnered with a vendor, we’re using that, but that was based out of other selections, and it drove higher consumption because it was a tool we had that we knew how to use, and we could increase usage and decrease our platform costs,” he said. “I believe we’re still somewhere around we’re saving around a million dollars a year because every mission area is not managing the platform. They donate to it every year. We continuously update it based on our customer’s needs. It really is changing and evolving based on how our customers come to us and partner with us.”

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.

Related Stories