We are setting up something like a document library on Amazon AWS. The architecture is basically 1. AMI server running MongoDB 2. A bunch of PDF documents on elastic block storage 3. Some server-side scripting that pulls and parses metadata into MongoDB, and finally, 4. some front-end services that query the database. The end goal is a website that facilitates search and discovery of the PDF library (e.g. show me all PDFs written by X).