As a quick FYI, I see in the comments below people comparing s3 pricing to the likes of a VPS instance storage from XYZ hosting company.<p>Given that you have no redundancy with that storage you could save an addition 30% and use reduced redundancy storage in s3 which still provides redundancy to a second AZ (just not all of them or across regions in the odd case of US Standard)
I really like that I can trust Amazon to keep AWS prices reasonable. I have used both AWS and App Engine, and while I love App Engine for ease of use, the pricing changes still make me feel skittish about using it for big projects even though they happened a year ago.
I wonder if dreamobjects being S3 api compatible and only 7 cents helped force the price decrease.<p>What I like about dreamobjects is no charge for get/put requests which can really add up with many small items being publicly hosted.<p>Maybe there's a remote chance of Amazon dropping "put" charges but I doubt they will drop "get" charges.<p>Ah I see Google dropped their fees first earlier this week so that was probably the pressure point. I thought Google came afterwards.
I believe the 'economies of scale' argument, but I also imagine that this move is related to the RedShift announcement. If Amazon wants users to use their offering for data warehousing and analytics, they'll need lots of cheap storage for hosting the data.
Sorry for my ignorance, but the pricing is a bit confusing for me. Let's say I have 1 GB of files, and I do 1 Terrabyte of downloads per month, how much would I pay?