Amazon is offering a new service: Amazon S3 – Simple Storage Service. Before I get into that though, a side note, who (or what) is generating URLs over at Amazon? The URL for their new S3 service:
http://www.amazon.com/gp/browse.html/ref=sc_fe_c_1_3435361_1/104-7254936-4303935?%5Fencoding=UTF8&node=16427261&no=3435361&me=A36L942TSJ2AJA
Easy to remember right? Good grief. With that out of the way, back to S3.
The prices seem reasonable: $0.15 per GB-Month of storage used and $0.20 per GB of data transferred. As others have already pointed out, this boils down to $15 per month for 100 GB of storage plus traffic fees. Who cares about 100 GB though, this puts at $150 per month for 1 TB plus traffic fees. Now were are talking storage. Of course uploading a TB worth of data from one place would take quite awhile, even with a 10 Mb connection, so this isn’t without problems.
Instead of launching this as an app first and then coming out with APIs, this is an API only service. If you’ve been reading this blog for awhile this shouldn’t come as a big surprise. I wrote about the idea of API only services back in July 2005. Amazon obviously has a huge amount of back end resources already to keep their current web offerings up, so focus on what you do well and create and API that allows others to build apps on top of it.
So what sort of apps are we talking about? I’d really like to see rsync gain the ability to speak to S3. I could rsync my notebook to S3 from time to time and if it was destroyed I could rsync the data from S3 to my desktop. Same idea for my desktop. I really like rsync.
Undoubtedly someone will create a web based file manager that will use S3 as the back end. This is just plain obvious. I’ll be surprised if we don’t see at least one show up in a fairly prominent way with in the next two months. After that I’d expect to see things like an app to allow you to map a Windows drive letter to your S3 account. A similar thing could be done in the unix world, perhaps via FUSE. Hopefully other web based apps will pick on this, imagine Flickr being able to pull images from S3 or writely being to read and write to S3. There is a lot of potential in this area.
Perhaps the biggest surprise out of all of this is that Google didn’t do it first. There have been rumors about Google offering a similar service, Gdrive, but it certainly isn’t available for use yet. I suppose Yahoo could have done this too.
For now I’m going to put the Amazon S3 PHP code sample on my list of things to review in the near future.