1. 1

First in a series about Dropbox switching to it’s own infrastructure from AWS without interrupting service.


  2. 1

    For now afaict most of the public technical information that’s available on their design is in various parts of an HN thread, where a few Dropbox engineers were commenting; look for the posts by jamwt and james_cowling. For example, how they do data redundancy, and why they chose rust for this project. Though presumably elaborated versions of all that will show up in this blog-post series.