Is it right for you?
A good fit
- OSPool jobs with input or output files over 1 GB, or inputs shared by many jobs
- Container images used by many OSPool jobs
- Reading datasets that are already published on the OSDF
- Projects that want to share a dataset with many compute sites
Not a fit
- Protected or confidential data (the OSDF is for open science data)
- Files that change often under the same name (caches can serve an old copy)
- A general-purpose place to keep your only copy of data
What it is
The Open Science Data Federation (OSDF) joins many data repositories (called origins) into one national delivery network, with caches placed near the computers that read the data. It is part of the OSG's services and runs software from the Pelican Platform. Files are named with osdf:/// paths, and the same path works from an OSPool job, a notebook or a laptop.
Three ways UCR researchers use it
- With the OSPool. Every OSPool account on an OSG-operated access point includes OSDF space for larger inputs, outputs and containers. See Moving data on the OSPool.
- Reading public data. Datasets published on the OSDF can be downloaded with the
pelicancommand without an account. - Publishing a dataset. A dataset in an Open Storage Network bucket can be connected to the OSDF on request to the OSG, so that jobs anywhere read it through the caches.
Data policy
The OSG describes the OSDF as suitable for non-public data from open science communities, but not for highly sensitive data. Non-public data is encrypted on the network, not on disk at the caches. Do not use it for protected data.
How to get started
Start with the researcher guide to the OSPool and OSDF. For help deciding whether your data fits, contact Research Computing.