as phvt said, use fuse (http://s3fuse.sourceforge.net/)
i use it for daily backups, works very well.
the backup script mounts the s3 bucket, copies over the compressed backup (to avoid excessive transactions and save space) and when it's done it unmounts the bucket.
the current backup size is about ~ 250MB - 400MB and i had no problems so far.
when using a s3 browser i can see all the data in the bucket, no funny business with custom/proprietary storage or so.

that said, this works fine for daily backups but probably not if you need to serve large amounts of data and don't want to keep them on S3 (after all, it is slower than a local disk or NAS/SAN in the same datacenter and costs money for storage/bandwidth).

Actually, I was talking about the FuseOverAmazon "s3fs" and I didn't recommend any s3fs tool (I can't because I haven't used them). I have used Jungle Disk on both PC and Linux so I wanted to respond to hybinet's post.

Thanks for sharing your experiences with s3fuse, though. As your backups grow, you might run into the S3 limit of 5Gb. Your backups are small enough right now that I think you could use a much simpler system than a whole filesystem to put your backup onto S3.

my bad, FuseOverAmazon ( http://code.google.com/p/s3fs/wiki/FuseOverAmazon ) is what i'm using, not the other one.

Oh! I didn't realize its bucket data would be compatible with other tools. I'll check it out again. I wouldn't be the first person to get confused with all the 's3fs' projects out there...

I might use Google's about-to-be-announced GDrive:

http://blogs.zdnet.com/Google/?p=1247

James

Hmm, I'd like to see Google publish something like the Amazon AWS security whitepaper first.

I just want to say thanks to MaineCoon for his post. I'm currently using it to back up my cPanel linode and it works like a charm!