Globus Connect Server

Globus Connect Server is software that is meant to be installed on multi-user machines, and on machines where multiple users have data (even if they can not log in directly). Users can access their files through Mapped Collections, and can create Guest Collections to share with others who do not have access.

Globus Connect Server natively supports transferring data to other Globus Connect Server collections, as well as to Globus Connect Personal (like the one which might be running on your laptop right now). Globus Connect Server also supports transferring data through a user’s web browser, though it is not fast.

Globus Connect Server requires:

• A supported Linux distribution that still receives vendor support.

  Globus Connect Server 5 supports running on Debian, Fedora, OpenSUSE Leap, Red Hat Enterprise Linux 7 through 9, SuSE Enterprise, and Ubuntu.

  For Enterprise Linux derivatives, Globus Connect Server 5 supports CentOS 7, 8 Stream, and 9 Stream. Also supported are versions 8 & 9 of Alma, Oracle, Rocky, and Springdale Linux.

  If you use Red Hat Enterprise Linux (including derivatives), you should expect to pull some packages from EPEL.

• A Public IP address.

  Even if you are only doing data transfer within Stanford, your system must be able to accept unsolicited connections from Globus HQ, running in the Cloud.

• Preferably, a server that you can dedicate to doing data transfer.

  Even with a single transfer, if configured appropriately, Globus will happily saturate a Gigabit or 10 Gigabit link.

Once you have the above items, you should read the Terminology Guide, as GCSv5 brings some new terms compared to GCSv4 (unfortunately, this page does not exist yet). After that, you should read the Pre-Installation Planning guide to prepare your environment for GCSv5.