I've been hosting a few large files like the screencast demonstrating how to install and configure CaRP on a secondary server to prevent overburdening my main server or overrunning my bandwidth limit. This morning, I was unable to reach the secondary server, preventing me from even being able to view some of my webpages. So I decided it was time to start using Amazon S3.
If you're not familiar with S3, it's a place where you can host files (either publicly available, or private) for web hosting, backup, transfer, etc. Amazon itself uses S3 for business critical storage, so you know they've made it extremely robust (they build it with the expectation that hard drive and other equipment failures will occur regularly -- I imagine they use redundant storage, raid and such to ensure that hardware failures cause no disruption to service).
They charge 15 cents a month per gigabyte of storage, plus a little for bandwidth usage. Based on my current needs, it should cost me in the neighborhood of $2.50 a month -- ie. it's almost as good as free. As cheap as it is, I'll be more motivated to create more screencasts and perhaps do some podcasting. I also plan to set up my primary server's backup system to offload the backups to S3 so that I don't have to download them immediately to my own computer to keep my data secure.
It turns out that my secondary server wasn't down -- there was just a routing problem somewhere between me and the server (somewhere close to my ISP -- if I SSH into my main server, I can reach the secondary server from there), so the problem wasn't causing a complete disruption to my business, but it's good to be switched over to a more reliable service anyway.