Big Federal Data Centers Implementing FAIR Data Principles: ARM Data Center Example

TLDR

The Atmospheric Radiation Measurement (ARM) program, a multi‑laboratory DOE National User Facility, stores over 12,000 data products totaling more than 1.8 PB in the ARM Data Center, encompassing instrument, model, campaign, and PI‑contributed data since 1992. The paper aims to show how large federal data centers like ARM can apply FAIR data principles to enhance efficiency. They apply FAIR principles, emphasizing machine‑to‑machine interactions, to ARM’s modern, scalable data architecture.

Abstract

Atmospheric Radiation Measurement (ARM) is a multi-laboratory/multi-institutional, US Department of Energy Office of Science National User Facility. ARM's data is currently hosted at the ARM Data Center (ADC) in Oak Ridge, Tennessee. The ADC holds more than 12,000 data products, with a total holding of more than 1.8 PB of data that dates back to 1992. This includes data from instruments, value-added products, model outputs, field campaigns, and principle investigator contributed data. In this paper, we discuss how big federal scientific data centers, such as ARM, that use modern and scalable architecture apply findable, accessible, interoperable, and reusable (FAIR) data principles to improve overall efficiency. These principles mainly emphasize machine-to-machine interactions that are directly applicable to ARM because of its data volume.