Last Updated: May 12, 2016

Table of Contents

1. Introduction

1.1. Globus Data Publication Overview

The Globus data publication service makes it easy for communities to establish searchable collections with customized metadata schema, curation policies, licenses, and access controls.

To submit a dataset for publication, authorized users follow a web-based workflow that employs fillable forms for collection-specific metadata and the Globus transfer interface for dataset assembly. A dataset in the Globus data publication service is composed of data file(s) and metadata. The data files may be of arbitrary size and format, allowing any type of data (e.g., text files, formatted documents, images, source code, executables, raw data from sensors, control parameters) to be associated and published in a single dataset. Depending on the collection’s policy, a curator may need to review and approve a submitted dataset before it is published.

Every published dataset is represented with its own landing page, which provides a description of the dataset and links to the data on the collection owner’s storage. A persistent identifier (PID) is generated for every landing page, facilitating citation. Globus provides search and browse capabilities, allowing authorized users to discover and retrieve datasets of interest. Globus retains a copy of the metadata to facilitate the discovery process; another copy is kept with the data on the owner’s storage.

1.2. Production and Trial Services

The production Globus data publication service is located at Only those datasets that are intended for long-term availability should be published with this service, as entries are citeable and will be maintained indefinitely.

In addition to the production Globus publication service, a "trial sandbox" has been set up at to let individuals and groups explore what Globus publication offers. Any Globus user can submit datasets to both the "Curated Trial Collection" and the "Uncurated Trial Collection" that are found in the "Open Trial" community on the trial service. All data on the trial service will be removed periodically and should not be cited elsewhere.

1.3. User Guide Overview

This guide is aimed at users who want to publish datasets in existing collections, including the trial collections. Administrators setting up and configuring collections should consult the Data Publication Administrator Guide.

This user guide introduces the Globus data publication web interface, and steps through the submission and curation workflows that make up the data publication process. Although the publication search and browse features are not covered in detail, access points to them on the publication home page and dashboard are documented. A glossary is included at the end of the guide.

Screen shots and links in later sections of this guide reference the trial Globus data publication service, as it is open to all Globus users. The instructions in this guide also apply to the production Globus data publication service, with workflow details varying from collection to collection.

2. Prerequisites

You must have a Globus identity before you can submit datasets to the Globus data publication service. Familiarity with the Globus transfer interface will also be helpful, as it is used to assemble datasets in the submission process.

If you are new to Globus, this how to guide will help you sign up for a Globus account and become comfortable with the Globus Transfer Files page. For the impatient, skip the details and use the Sign Up page to create your Globus identity.

3. Getting Started

3.1. Publication Home Page

The Globus data publication home page (for the trial service) is shown here:

Publication Home Page
Figure 1. Publication Home Page

The publication home page is displayed when you first access the publication site. It can also be reached via the Browse & Discover link that appears on most publication pages after you log in.

The publication home page is a starting point for browsing communities and their collections, and for discovering published datasets of interest through facet-based filtering (author, subject, and issue date). The Search box located near the top of the page is another way to find datasets of interest—keywords entered in the box are used to match descriptive text of published datasets.

Globus makes sure you see only the communities, collections, facet values, and datasets that you have permission to access.

3.2. Logging In

To submit a dataset for publication you must be logged in to the publication service with your Globus identity.

To log in, select Log In on the navigation bar at the top of the publication home page. If an active Globus session already exists, you will immediately be redirected to the data publication dashboard. If no session is active, a Sign In page will prompt for your Globus username and password (you may use any linked Globus identity and credentials) before displaying the data publication dashboard.


On occasion, you may notice that the Account menu is displayed in the navigation bar, indicating that you are logged in, but when you interact with the service you are asked to log in again. This typically occurs when you have been inactive for awhile and your session has expired. Since your browser window is not automatically refreshed, the menu continues to display. This behavior can be confusing when you first encounter it.

4. Data Publication Dashboard

The data publication dashboard is the starting point for most operations.

4.1. Empty Dashboard

The first time you use the Globus data publication service, and other times when you do not have any pending work or submissions, your dashboard will be empty:

Empty Data Publication Dashboard
Figure 2. Empty Data Publication Dashboard

The navigation bar at the top of the dashboard has Publish highlighted to show that you are using the Globus data publication service. The navigation bar menus and options provide quick access to many other Globus features outside of the publication service.

Just below the navigation bar, the Browse & Discover option will take you to the data publication home page that was discussed earlier. The Data Publication Dashboard option is not selectable while you are on the dashboard page, but from other publication pages it will take you to the dashboard. The Communities & Collections option will take you to an alternate interface for exploring the publication service’s participants and their holdings.

The Search box works exactly as on the home page—keywords entered in the box are used to match descriptive text of published datasets. With all browse, discover, explore, and search operations, Globus makes sure you see only the information that you have permission to access.

Located in the Data Publication Dashboard area of the screen, Submit a New Dataset will launch a new submission workflow, and View My Published Datasets will open a display showing all of your submitted datasets that have been accepted for publication.

4.2. Populated Dashboard

When you have pending work or submissions, the dashboard will be more fully populated:

Populated Data Publication Dashboard
Figure 3. Populated Data Publication Dashboard

In this figure, the dashboard area provides updates on all of the datasets in the curation and submission stages of the publication process that are relevant to the logged-in user. The details will be explained in later sections as you step through the publication process, but briefly:

  • Curation Tasks Waiting to be Performed: Lists datasets you previously agreed to curate but have not yet approved for publication.

  • Curation Tasks Waiting to be Claimed: Lists submitted datasets that are waiting to be claimed by a curator.

  • Dataset Submissions in Progress: Lists your datasets that need further work before being submitted.

  • Datasets Submitted for Approval: Lists your datasets that have been submitted but not yet approved for publication.


Collections exist in communities. Collection names must be unique within a community, but need not be unique across communities. To fully identify a collection, both the collection and the community are displayed using the notation "Collection Name << Community Name".

Before leaving this section, notice that the dataset "Monday Curated Dataset" in the collection "Globus Demo Collection « Globus Demo Community" appears in two sections of the Populated Data Publication Dashboard figure. Typically, a user would not be offered the option of curating their own dataset. This option is allowed for the "Curated Trial Collection", so that users can take a dataset through both the submission and curation stages of the publication process.

5. The Submission Workflow

5.1. Overview

The submission workflow, described in this section, is the first stage of the publication process. The submission workflow is used to create a new dataset and submit it for publication in a collection.

Collection administrators have the ability to customize the submission workflow, specifying which steps to include and the order in which the steps are performed. Submission workflow steps can include License, Describe, Assemble, Verify, and Complete. Collection administrators also control the license terms that must be accepted in the License step and descriptive (metadata) information that can and must be entered in the Describe step.

In this guide, a dataset will be submitted to the collection "Curated Trial Collection" in the community "Open Trial." The workflow steps and metadata requirements for that collection are representative of those found in many others. The general submission process is the same, regardless of the collection-specific details.

Pages in the submission workflow have some common elements that help you navigate through, and exit from, the workflow.

A progress bar at the top of the page shows the steps of the submission workflow:

Submission Workflow Progress Bar
Figure 4. Submission Workflow Progress Bar

In the progress bar, completed steps are colored light blue, the current step is highlighted in dark blue, and upcoming steps are shown in grey. You can return directly to any completed step to review or revise your entries by clicking on the (light blue) step name in the progress bar.

Many of the submission workflow pages contain navigation buttons at the bottom of the page:

Submission Navigation Buttons
Figure 5. Submission Navigation Buttons

Back takes you to the previous step in workflow, Continue takes you to the next step, and Save & Exit saves the workflow and returns you to the publication dashboard.

The save and exit option can be particularly helpful if a long-running operation, such as a large transfer, is taking place and you want to exit and return later to check progress and continue. The saved submission workflow will be displayed in the Dataset Submissions in Progress section of your dashboard. From there, you can resume the submission process, view the dataset submission in its saved state, or remove the submission from the system.

5.3. Select Collection

To begin a new dataset submission, choose Submit a New Dataset from the data publication dashboard, discussed earlier. You will be asked to select the collection where you want to publish your dataset:

Select Collection
Figure 6. Select Collection

The collections that you have permission to submit (and publish) to are shown in the dropdown list. After you select a collection and leave this page, the collection cannot be changed. If you accidentally specify the wrong collection for your dataset and continue beyond this page, you must Save & Exit this submission workflow, remove this dataset submission (via the dashboard), and begin again by choosing Submit a New Dataset.

In this guide, the collection "Curated Trial Collection << Open Trial" is selected.

5.4. Accept License Terms

Collections typically require you to accept license terms that govern the conditions under which the data is being shared with the owner of the collection. When such a license is present, you must accept the license terms during the submission workflow. Normally, this is done as the first step of the collection’s workflow so that the license terms are clear early in the dataset submission process.

Accept License Terms
Figure 7. Accept License Terms

Since you cannot continue with the workflow until you accept the license terms, the usual navigation buttons are not present on this page.

If you choose I Do Not Accept the License Terms, the dataset submission will be saved and accessible via the Dataset Submissions in Progress section of your dashboard. This gives you the opportunity to clarify the license with the collection owner and then resume the submission if you are ready to accept the terms, or to remove the submission if you are not.

Choose I Accept the License Terms to proceed to the next workflow step.

5.5. Describe the Dataset

This step in the submission workflow is used to enter metadata that describes the dataset you are submitting for publication. You will be presented with a fillable form where you enter the metadata. Each collection defines its own metadata and controls the prompts that appear on the form.

If the collection’s form contains many fields, it may span multiple Describe pages. The submission workflow progress bar indicates how many pages the form spans for the selected collection (three in the example shown). Like other parts of the workflow, you may save your work at any time and resume later. This may be helpful, for example, if you do not have all of the required information when you start the submission process.

The figures in this section show two of the three pages in the metadata form defined by the administrator of the "Curated Trial Collection". Different collections may have different forms, but the general process for entering metadata is the same across collections.

Describe Dataset
Figure 8. Describe Dataset, Page 1

In the form, fields marked with an asterisk (*) are required. Some fields, such as "Authors" in this example, permit you to enter multiple values. When multiple values are allowed, the Add More button is shown to the right of the field entry. Click this button to display additional rows so more values can be entered. Values can be removed by clicking Remove Entry next to the entry to be removed.

Describe Dataset
Figure 9. Describe Dataset, Page 2

The second page of the form provides an example of additional types of metadata. You can see that the Globus data publication service allows for considerable customization in terms of metadata schema and entry prompts and fields.

Enter as much (or as little) descriptive metadata as you want on pages two and three of the Describe workflow step. For the "Curated Trial Collection", none of the fields on pages two and three are required by the collection.


In some cases, the metadata field entry prompts do not exactly match the labels shown on the more generic publication pages. For example, "Issue Date" and "Publication Date" refer to the same metadata field, and "Subject", "Subject Keywords", and "Keywords" are used interchangeably.

5.6. Assemble the Dataset

The assemble dataset step encapsulates the task of copying the data files that will be included in your dataset to a storage location allocated for the submission.

Assemble Dataset
Figure 10. Assemble Dataset

Click on the Assemble Dataset button to open the Globus Transfer Files page (also referred to as the Globus transfer interface) in a new browser window or tab. You will use the transfer interface to copy your data from one or more Globus endpoints to the submission’s allocated storage location.


In some browsers, if the Globus transfer interface is already open in another tab when you click "Assemble Dataset", the browser will not automatically raise and shift focus to that tab. Look for the tab labeled "Transfer Files | Globus" and manually select it if necessary. You may not notice this behavior the first time you assemble a dataset, but you may run into it later.

Globus Transfer Interface
Figure 11. Globus Transfer Interface

The right pane of the Globus transfer interface is pre-populated with the submission’s storage location (endpoint and path). Do not change this pre-populated location. Use the left pane to navigate to the Globus endpoint and path that has the data files you want to include in your dataset. Transfers are initiated using the right-pointing arrow.

You can use the transfer interface to initiate multiple transfers, starting from various endpoints or paths, to complete assembly of your dataset. The files can be of any type and size, allowing you to assemble and publish a dataset with wide-ranging content.

In the figure above, we chose to transfer three small files from the universally-readable "Globus Tutorial Endpoint 1" endpoint and the "/share/godata/" path. The three files are copied in a single transfer task and they are the only data files we will include in the submitted dataset. You can also use one or more of these files for your trial submission, or choose other files you have access to that contain artificial data—you should not publish real data in the trial collections.

After all transfer tasks have been initiated, you can close the transfer interface to return to the Submit: Assemble Dataset page. Alternatively, you can leave the transfer interface open and return to the submission workflow by selecting the "Assemble Dataset | Globus" tab or window in your browser.

Your assemble dataset window can be refreshed periodically to monitor the progress of all transfer jobs associated with the submission:

Assemble Dataset Transfer Status
Figure 12. Assemble Dataset Transfer Status

You can check the details of a transfer task by clicking the link showing the status of the transfer ("SUCCEEDED" in the previous figure). Globus will also send you an email indicating the completion or other status of your transfer(s).


When transfers are long-running, it is often desirable to save the current workflow and return later to check progress. As described earlier in the navigation section, the state can be saved, and the dataset will be listed in the Dataset Submissions in Progress section of the dashboard. The progress bar of the submission workflow allows you to access the assemble page directly to view the updated status of the transfers for the dataset submission.

Before continuing past the assemble step of the workflow, make sure that all of your transfers completed successfully. You can re-open the Globus transfer interface (via the Assemble Dataset button) to transfer additional files or to confirm the contents of the dataset submission’s allocated storage location—a screen refresh may be needed.

5.7. Verify the Dataset and Finalize Submission

The verify step of the workflow gives you the opportunity to review the contents of your dataset before submitting it for publication. This is your last chance to correct any problems with the metadata or data files that make up the dataset. After you finalize your submission, metadata cannot be changed and access to the allocated endpoint and folder is restricted so assembled files cannot be altered.

Verify Dataset
Figure 13. Verify Dataset

All of the metadata values, as well as the status of all transfer operations for the dataset, are displayed on the verify screen.

If you need to make changes, you can use the progress bar to return to any step in the workflow, or the Back navigation button to return to the previous step. Alternatively, the Edit buttons that appear with each group of metadata can be used to go directly to the relevant Describe page to make corrections. Finally, the Review Files button will open the Globus Transfer Files page, allowing you to review the data files that have been assembled for the dataset and to add (via new transfers) or delete files if necessary.

After any updates, the Globus service will re-create the metadata file that is included in the dataset—this operation must complete before you can finalize your submission.

When you are happy with the metadata and data files in your dataset, select Finalize Dataset Submission to submit the dataset for publication. You will see the following screen:

Dataset Submission Complete
Figure 14. Dataset Submission Complete

If the collection you submitted to requires curation prior to publication, your dataset submission will be queued for review by a user with the curator role for the collection. The "Curated Trial Collection", used in the example shown, does require curation. While your dataset is awaiting curation, it will be shown in the Datasets Submitted for Approval section of the publication dashboard:

Dashboard with Dataset Submitted for Approval
Figure 15. Dashboard with Dataset Submitted for Approval

If the collection does not require curation, your submitted dataset will be published immediately and made available to users who are allowed to access the collection.

6. The Curation Workflow

6.1. Overview

The curation workflow, described in this section, is the second stage of the publication process for those collections that are configured to require curation. The curation workflow is used to review, and potentially update or reject, a dataset submission prior to publication.

Curation for a given collection is performed by a group of users identified by the collection administrator.

6.2. Become a Curator for the Curated Trial Collection

Special provisions have been made to allow any Globus user to join the group of curators for the "Curated Trial Collection" (the collection used for demonstration purposes in this guide). If you submitted a dataset to the "Curated Trial Collection", and would like to try out the curation steps for your dataset, follow the instructions in this section to become a curator.

First, select the Groups button on the navigation bar to move to the Globus Groups page.

Navigate to Globus Groups
Figure 16. Navigate to Globus Groups

On the Globus Groups page, shown below, raise the "Search" tab on the left side of the page, and enter ‘data publication open trial’ in the search box to find the group you need to join. The group you are looking for is "Data Publication Open Trial Curators".

Search for Data Publication Open Trial Curators
Figure 17. Search for Data Publication Open Trial Curators

Click on the "Data Publication Open Trial Curators" name when it is displayed in the search panel to see a description of the group:

Join Data Publication Open Trial Curators
Figure 18. Join Data Publication Open Trial Curators

Click the Join Group button to request membership to the group. You will be prompted to select an identity and provide identity information. Then click Submit Application to Join.

Join Data Publication Open Trial Curators
Figure 19. Select Identity

You should see confirmation that your application to the group has been submitted. Once approved, the "My Groups" tab on the left side of the Globus Groups page should list "Data Publication Open Trial Curators" (you may need to refresh your browser window):

Member of Data Publication Open Trial Curators
Figure 20. Member of Data Publication Open Trial Curators

At this point, you are finished with the Globus Groups page and can return to the trial Globus data publication service.

6.3. Claim Curation Task

If you have a curator role for one or more collections, the dashboard will inform you if there are any pending curation tasks waiting to be claimed.

Claim Curation Task
Figure 21. Claim Curation Task

In the figure above, there is one task to be claimed in "TestMDF << Materials Data Facility". As a courtesy to those working in the "TestMDF << Materials Data Facility", please curate the datasets you submit and not those submitted by others. Normally you would not have the option of curating your own datasets.

Click the Claim button for the dataset you want to curate.

6.4. Preview Curation Task

After you claim a curation task, the Preview Task page will open. It displays more information about the dataset submission and provides the option of performing the curation (Accept Task), or deciding not to claim the task after all (Cancel). The location of the dataset is shown, but it cannot be accessed until the task has been accepted.

Preview Curation Task
Figure 22. Preview Curation Task

6.5. Perform Curation Task

If you chose Accept Task on the Preview Task page, the Perform Task page is displayed:

Perform Curation Task
Figure 23. Perform Curation Task

At this point, the location of the dataset in the "Endpoint and path to dataset" area of the screen is a link that will open the Globus Transfer Files page, giving you access to the data files that are part of the submitted dataset.

Depending on how curation is configured for the collection, the curation operations Approve, Reject, and possibly Edit Metadata will be available to you. The Do Later button lets you save the current state of the curation activity and resume it later via the Perform button next to the dataset in the Curation Tasks Waiting to be Performed section of the dashboard. The Unclaim button lets you return the submitted dataset to the pool of curation tasks waiting to be claimed.

As a curator, you should inspect the metadata presented on this page and use the location link to examine the data files. For collections that allow curators to edit metadata, Edit Metadata will take you to the same workflow that was used during dataset submission. In some cases, the metadata entry form will contain additional fields that were not presented to the dataset submitter. After completing the workflow, you will be returned to the Perform Task page.

If you Approve the dataset submission, a persistent identifier will be assigned which links to the landing page for the dataset. This identifier is displayed on the confirmation page shown after approval. Additionally, the user who submitted the dataset will be notified by email that the dataset has been added to the collection; the email will include the persistent identifier for the dataset.

If you Reject the dataset submission, you will be prompted to enter descriptive text providing feedback to the user who submitted the dataset. The text you enter will be included in an email to the user informing them that their dataset submission was rejected. The rejected dataset will appear in the Dataset Submissions in Progress section of the user’s dashboard. From their dashboard, the user can resume work on the dataset to address deficiencies identified in the curation process and resubmit, or remove the dataset submission.

7. Dataset Landing Page

Every published dataset is represented with its own landing page, which provides a description of the dataset and links to the data on the collection owner’s storage. Authorized users can access the dataset landing page via the persistent identifier that was assigned when the dataset was published, or discover it through any of the Globus data publication search and browse features.

The landing page for the example dataset submitted and curated in this guide is shown here:

Dataset Landing Page
Figure 24. Dataset Landing Page

The landing page shows the distribution license terms for the dataset, which are set by the collection. A downloadable citation that includes the persistent identifier is also provided on the page.



A collection in the Globus data publication service holds a group of published datasets with common characteristics: metadata schema and requirements, access privileges, publication and distribution licenses, curation requirements, PID type, and storage. Every collection exists in the context of a community, and a combination of community and collection policies determine the publication requirements for the collection.


A community is the representation of an organization or institution in the Globus data publication service. An administrative group is associated with every community, and members of that group have the authorization to create or delete collections or sub-communities within the community.


Curation in the Globus data publication service is the process of reviewing, and potentially updating or rejecting, a dataset submission. Curation provides a degree of verification prior to publication of the dataset in a collection. Typically, curation will be performed by a group of users whose domain knowledge is sufficient to insure that the submitted dataset is of appropriate quality for the collection. Collection policies control the curation requirements for a given dataset submission, and not all collections require curation of submitted datasets prior to publication.


A dataset is composed of data file(s) and metadata that are submitted and published as a unit in the Globus data publication service. The data files may be of arbitrary size and format, allowing virtually any type of data (e.g., text files, formatted documents, images, source code, executables, raw data from sensors, control parameters) to be associated and published together.


In the context of the Globus data publication service, metadata represents information provided by the user to identify or describe a dataset they are submitting for publication. The metadata is entered via forms in the submission workflow and conforms to the schema and requirements defined by the collection where the dataset will be published. The Globus publication service automatically includes the metadata in the dataset when the dataset is published.

Persistent Identifier (PID)

A reference, often a URL or DOI, to a dataset or other resource that is intended to be valid for a long period of time. PIDs are a good form of citation because they provide a stable way to reference a resource, even if the location of the resource changes. The Globus data publication service automatically generates a PID when a dataset is published to a collection.