This is an implementation of the PyPI server for hosting your own python packages. It stores the packages in S3 and dynamically generates links to them for pip.
After generating the S3 urls, pypicloud caches them in a database. Subsequent requests to download packages will use the already-generated urls in the db. Pypicloud supports using SQLAlchemy, Redis, or DynamoDB as the cache.
Pypicloud was designed to be fast and easy to replace in the case of server failure. Simply copy your config.ini file to a new server and run pypicloud there. The only data that needs to be persisted is in S3, which handles the redundancy requirements for you.
Code lives here: https://github.com/mathcamp/pypicloud