Federal agencies are starting to see significant advantages in reversing how they gather and analyze data to fuel their artificial intelligence efforts. Instead of pulling massive amounts of data into centralized data lakes for analysis, enterprises are starting to send analytics tools to the edge of their networks to parse localized data and then federate the results.
The change in approach represents a paradigm shift in how organizations are capturing and analyzing data, according to Al Ford, who oversees artificial intelligence alliances in the federal government market at Dell EMC. And it is providing agencies the means to dramatically reduce the volume of data transiting across their networks.
“We’re seeing tremendous examples of how you can reduce the overall network traffic [by] not using a centralized model,” says Ford in a new FedScoop podcast, underwritten by Dell Technologies. “If you’re using a federated model” — capturing pre-analyzed data from the edge to train AI models — “you’re not sending that data back to one central location.” That results in significant savings in time and network resources, he explains.
But Ford also points to second big advantage agencies are discovering by using a federated AI approach: the ability to provide inferencing capabilities almost in real time.
The shift to deploying containers with built in analytic tools “represents a change in trajectory in time-to-analytic,” says Ford.
“It provides a new roadmap for capitalizing on data and AI,” he adds. “This new roadmap allows for AI for more people, faster adoption, and increases the utilization of data that actually comes from the edge,” he says. That in turn, helps accelerate the “democratization of AI,” he says. “Not having to move their data is critical.”
Ford discusses the technology changes that are driving the shift in approach, citing the increased role of containerized workloads in today’s IT environments — and the ability to incorporate artificial intelligence analytics tools inside those containers.
Technology providers, like Dell Technologies, are able to build inference engines into containers, and deliver those containers from a centralized location to the edge. “So only the results of the analytics are what are shipped back,” he says.
Ford also highlights examples where federating analytics results, instead of transmitting entire data sets, is helping organizations like the Department of Defense, harness AI models more effectively. In one example, he illustrates how a commercial airline captures on-board analytics to alert maintenance crews ahead of flight arrivals of potential aircraft maintenance needs.
Listen to the full podcast conversation on federating analytics data on FedScoop.com. And hear more of our coverage of “IT Modernization in Government” on FedScoop’s podcast channels, wherever you get your podcasts.
Learn more about AI and data analytics at the edge from Dell Technologies.
Al Ford is responsible for artificial intelligence alliances for federal government at Dell EMC. He plays a central role in facilitating collaboration between federal agencies, Dell Technologies and Dell EMC’s technology partners to bring AI solutions to the federal government.