I want to host a fairly large HDF5 file on the Internet, and I see Amazon S3 and / or EBS as a possible place to store it, and perhaps use an EC2 instance or two for processing in the background. I am looking for practical ideas on how this could be done or was done before and, if possible, a link to an actual example. I think this can be done in the same way as if someone was accessing a flat file in some other built-in data storage, but would like to avoid reusing the wheel if there is information available.
Marc source share