so -- i've been building a piece of software to work with amazon s3, which is a web service that allows you to put and retrieve files from amazon's servers. the idea here is that you can get data redundancy. since amazon has to support a geographically dispersed computing environment anyway, they have decided to allow you to purchase storage space on their servers. since their scale is large, they can provide a lot of space cheaply. they are charging $0.15 cents per month per gigabyte with $0.20 per gigabyte transfer fee.
this means off-site backups with multiple redundency for a pittance. for instance, i am a developer and i have about 10 gigs of non-replaceable data (i.e. source code, emails, etc...). i can store that for $1.50 a month plus $2.00 to get it up there and some fraction up to $2.00 to retrieve it, if i have to.
currently the options are:
- get a hosting account with a lot of space and store the files there
- advantage: pretty cheap. you can get hosting accounts for a couple of bucks that offer you several hundred gigs with no transfer fees.
- disadvantage: requires ftp to get the files there and retrieve them or create your own web service
- use an existing service
- advantages: it exists already
- disadvantages: flat fees cause "use it or lose it" syndrome
- create s3nas
- advantages: it will work how i want it to (which is the best way, for me)
- disadvantages: must trust my own code with important data
so, the minimum charge for working with amazon s3 seems to be $0.01 per month. if they have 100,000 developers experimenting with s3, that continues to be an extremely small chunk of change.