Examples
Worked examples
- Is an instance
A research group uses Globus to transfer a multi-terabyte genomic sequencing dataset from an HPC cluster's scratch storage to their institutional data repository at project closeout, with the transfer automatically resuming after an overnight network interruption.
- Is an instance
A multi-institution collaboration creates a shared Globus collection so an external partner lab can pull the latest simulation output directly from the originating institution's storage, authenticating with their own university credentials rather than requesting a new account.
Counter-examples
Looks similar, but isn't
- Not an instance
Emailing a spreadsheet as an attachment (too small in scale and lacks Globus's endpoint/authentication model - this is simply file sharing, not a managed transfer workflow).
- Not an instance
A certified data repository like Zenodo or Dryad on its own (a repository provides long-term deposit, citation, and preservation; Globus is the transfer/access layer that can move data into or out of one, not a repository itself).
Editorial commentary
Globus is a non-profit research data management service, built on high-performance GridFTP-based transfer technology, that lets researchers move, share, and manage data across storage systems at different institutions, high-performance computing (HPC) centers, and cloud or repository platforms without manually handling low-level file-transfer protocols. It originated from the Globus Toolkit grid-computing research led by Ian Foster and collaborators at the University of Chicago and Argonne National Laboratory, and was relaunched as a hosted service (originally ‘Globus Online’) in 2010; it is now operated by the University of Chicago through its Computation Institute partnership with Argonne.
Core services
- High-performance file transfer – a managed, fault-tolerant transfer engine that automatically retries and resumes transfers interrupted by network or system failures, checksums data for integrity, and can move datasets ranging from megabytes to petabytes far faster than manual SCP/FTP methods.
- Globus Connect endpoints – institutions, labs, and individual researchers install Globus Connect Server (for shared institutional storage, HPC clusters, and repositories) or Globus Connect Personal (for a laptop or lab workstation) to turn a storage location into a named ‘endpoint’ or ‘collection’ that Globus can transfer to and from, without exposing the underlying filesystem directly to the internet.
- Sharing and collaboration – once an endpoint is established, data owners can create shared collections and grant specific collaborators or groups access to a folder without transferring ownership of the underlying storage, using Globus Auth for federated identity so users authenticate with their home institution’s credentials.
- Automation and flows – command-line and API access, plus a flows/automation service, allow research groups to script recurring transfers (for example, moving instrument output to a repository on a schedule) rather than triggering each transfer manually.
Role in research data management workflows
Within an RDM workflow, Globus most commonly sits at the movement and access-control points of the data lifecycle rather than at long-term storage itself: moving raw instrument or simulation output from an HPC system to an institutional or disciplinary repository at deposit time; transferring large datasets between collaborating institutions during a multi-site project; or giving external collaborators controlled, time-limited access to a dataset that is too large to send as an email attachment or upload through a conventional web form. Because it operates independently of the storage system underneath it, Globus is typically used alongside, not instead of, a certified repository – the repository provides long-term stewardship and citability, while Globus provides the transfer and access layer that gets data there or moves it between systems.
References
- Globus (globus.org), ‘About’ and ‘Data Transfer’ service pages.
- University of Chicago Department of Computer Science, ‘Globus Celebrates Ten Years of Connecting the Research Universe.’
- Argonne National Laboratory, coverage of Globus development milestones.
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="Globus"
vocab-term-identifier="https://casrai.org/dictionary/term/globus" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/globus",
"name": "Globus",
"identifier": "https://casrai.org/dictionary/term/globus",
"description": "A non-profit, hosted research data management service - built on GridFTP-based transfer technology and operated by the University of Chicago with Argonne National Laboratory - that provides high-performance, fault-tolerant file transfer, institutional 'endpoints' (via Globus Connect Server/Personal), and controlled sharing between research storage systems, HPC centers, and repositories.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/data-infrastructure#set",
"url": "https://casrai.org/dictionary/term/globus",
"sameAs": [],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"dateModified": "2026-07-18T06:30:53",
"inLanguage": "en"
}
