Globus
Globus is a non-profit service for secure, reliable research data management. It allows users to transfer their data between established "endpoints" through either a convenient web interface or a comprehensive command line interface (CLI) that is suitable for use in scripts. Endpoints must run a Globus Connect server, so you cannot use Globus to transfer data between two arbitrary computers. Globus provides the following benefits over other data transfer mechanisms:
- Speed - By optimizing the transfer settings for you and using parallel streams, Globus achieves transfer speeds that can't be matched by other common utilities.
- Reliability - If a transfer is interrupted for some reason, it will be retried and resumed automatically. Additionally, you can optionally choose to perform validation after a transfer is complete.
- Security - While your credentials are always sent securely, your data are not (for performance reasons). However, you can optionally have your data transmitted in an encrypted form, making the data security comparable to SSH.
- Ease of Use - The web interface used to initiate and monitor data transfers makes it easy for new users to get started with Globus. And the web interface is available everywhere, without the need to install software.
- Cloud Transfers - Organizations with a Globus subscription can additionally pay for cloud "connectors" that work with many popular cloud platforms.
Learning to Use Globus
Globus is described in detail in two other topics at this site. The introductory topic, Globus Data Transfer, provides information about:
- How Globus transfers work.
- Creating a Globus ID and linking it to other organizational accounts, such as for research institutions like ACCESS and universities.
- How to use the Globus web interface to perform data transfers and other operations.
A second topic, Advanced Globus Data Transfer, describes some more technical aspects of using Globus, including:
- Using the Globus command line interface to issue data transfer commands from a terminal.
- Best practices for using Globus CLI commands in shell scripts, which allow you to automate your common data transfer work.
- How to install and use a "personal" Globus endpoint on your own computer.
Exercises
In order to perform this exercise, you must first generate the test data on your local computer. If you have not already done so, create a Globus account. Then, either identify two computers that are Globus endpoints (such as TACC's Stampede2, Frontera and Ranch) or download Globus Connect Personal to your computer to make it serve as a Globus endpoint. Detailed instructions for these steps can be found here.
Use the Globus Web Interface
Use the Globus web interface to transfer the three files from your "source" computer to your "destination" computer and back, noting the amount of time each transfer takes. Are the transfer times longer or shorter than when using scp or sftp?
Use the Globus Command Line Interface
Install the Globus command line interface on your source computer and use it instead of the GUI to transfer the three files to your destination computer and back again.