Amazon basically gives every Kindle owner an email address to which they can email documents that will be converted to .mobi format and synced to their Kindle. I thought this was pretty interesting, so I started trying to think my way through how I could implement something similar, but quickly realized I was in over my head. How do you accept mail for tens of thousands of users, much less millions?

BTW, this service only allows emails to your address from authorized email addresses. I could see this being implemented as a procmail rule, except how do you scale the creation of individual procmail rules to this number of users? How do you even verify that the email truly came from that address in a scalable way?


  • Without pgp signing, there is no way to ensure that an email came from a certain address. Things like SPF help cut down on forgery, but in most cases, it's trivially simple to make an email appear to come from any address you want. – EEAA Jan 4 '12 at 4:21

You use a catch-all e-mail address that gets piped into their systems. No individual users or rules required - everything gets handled by the same set of scripts.

| improve this answer | |
  • I thought that made the most sense from a systems perspective, but I guess I'm inexperienced enough that I don't know how you have N email addresses that are all actually processed by the same queue processor. – Hank Gay Jan 4 '12 at 19:19

Not the answer you're looking for? Browse other questions tagged or ask your own question.