Could S3 be an end-user product?

Thursday, December 20, 2007 by Dave Winer.

I've not made much of a secret of the fact that I've been working on a new product, and am getting close to offering it as a public pre-alpha thing for Mac users only.  

It's fairly Flickr-centric, sucking photos down from Flickr in a variety of ways and pushing up photos in others. Like Radio 8 and Google Reader, it has the ability to maintain an output feed of stuff you want to pass on to friends and associates. 

As I was developing this I wished that Flickr had the ability to store simple text files, because I needed a place to put an RSS 2.0 feed with enclosures on behalf of each user. Of course I used Amazon S3, but I had to implement my own lightweight identity system so that Juan couldn't accidentally or intentionally replace Alice's feed. If only every user had a place where they could store stuff that's net accessible so that once and for all we could stop inventing new places. 

I was inspired to write about this when I read an Uncov review of Pownce where they reminded me that they were reselling S3 storage at a big markup. What if the users had their own S3 storage that they paid for themselves?  

Then it occurred to me to ask if people thought S3 could be an end-user thing. I'd like to find out, so if some non-technical users who have Amazon accounts would like to try setting up an S3 account, I'd be interested in hearing how it goes. Here's an idea of how you get started. 

1. You must already have an Amazon account. Nothing special about it, if you buy shirts or books or stereo equipment from Amazon, you use the same account for S3 storage. Already that's pretty easy, millions of people have Amazon accounts, right? 

2. Visit this page on Amazon, ignore all the stuff about objects and buckets. In the right margin is a big button that says "Sign up for this web service." Click the button. A very familiar page appears, asking you to sign into your Amazon account as if you were going to buy something (you are!). 

3. From there, I'm not sure what you see, because I have already enabled my account. But the end result of signing up is that you get two strings with weird names: Your Access Key ID and Your Secret Access Key. Any software that would save a document in S3-space on your behalf would need these two strange strings. In return each document would have a URL just like any other document on the web. Nothing strange about that.  

4. You could also use the space to store stuff using an FTP-like app that runs in Firefox called S3 Organizer. It's about as hard to use as the Mac Finder or the Windows Explorer, i.e. it's no challenge for a moderately geeky user. The cool thing about it is that you're able to share anything you upload into S3 space with anyone else. You can even use BitTorrent to access any file to save you bandwidth bills and distribute the load round the net. It's all very easy to do, imho. 

As a developer who has to pay for his users' storage needs I would very much like to see users learn how to use S3 to store their stuff, so I can focus on writing software and fixing bugs instead of paying to store your stuff.