Efficient access to publicly available big data sets is becoming a key issue for AI applications. This presentation will demonstrate the use of the FAB-cache service to efficiently retrieve large public data sets from the web, and will also show how to leverage FABRIC's various measurement services to track data movement performance.