Is the NBIA Data Retriever Open Source?

Yes!  The NBIA Data Retriever, like the back end server that it talks to, is licensed under the 3 clause BSD license, sometimes referred to as the 'new' or 'modified' BSD license.  The BSD license is very permissive, business friendly, and approved by the Open Source Initiative.  The source code is available from Github in the CBIIT/NBIA-TCIA repository, specifically in the software/nbia-download-manager subdirectory.  Developers are free to review or fork to create derivative works.  Of course, we would appreciate, but do not require, 'pull' requests back to the repository if you have created nice new features or bug fixes that others might benefit from.  They would be reviewed by the NBIA development team for inclusion, possibly after some tweaks, back into the main repository.

If you do create a derivative work, please keep in mind that the RESTful API that the NBIA Data Retriever uses to communicate back to the NBIA services is not standardized.  The API could change in future versions of NBIA or the NBIA Data Retriever, though we do try to maintain some level of backwards compatibility.  The same is true of the format for the Manifest files that drive the NBIA Data Retriever.  It is up to the maintainer of derivative works to stay 'up-to-day' on the RESTful API and manifest file format.

Do I have to use NBIA Data Retriever to pull data from the archive?  Or are there alternatives?

Data in the archive is available through other methods.  For example, a person versed in scripting or programming could easily pull data from the archive using TCIA's or NBIA's RESTful interfaces, described at 

How do I know which version of the NBIA Data Retriever I am running?

...