> If a bunch of people shared some files via Dropbox, the company stored the files on Amazon’s Simple Storage Service, or S3, while housing all the metadata rel

> If a bunch of people shared some files via Dropbox, the company stored the files on Amazon’s Simple Storage Service, or S3, while housing all the metadata related to those files—who they belonged to, who was allowed to download them, and more—on its own machines inside its own data center space. I'm surprised at this. I would have expected the opposite: using AWS for the application and metadata but managing your own storage infrastructure. Much like how Netflix runs on AWS but streams video through its self-managed CDN. Can you shed some light on why it was designed this way?