Update: Dataverse 3.1.6 introduced three new Amazon S3 integration nodes; namely the S3 Get, S3 Put and S3 Delete nodes - which allow you to upload files to an S3 bucket, Download objects from a S3 bucket, and delete objects in an S3 bucket, respectively. Lavastorm recommends the use of the newer Amazon S3 nodes for current projects.
I've had a number of people ask me about connecting to Amazon S3. At one point, a few of my colleagues and I created a custom node that accomplishes this task.
The node itself leverages our data flow and interfaces to Java and HTTP to create an API connection to Amazon. An example of the node in use is attached to this post. However, the credentials will need to be updated to your environment.