Amazon S3: Utility Computing in Practice

Have you ever heard of Amazon S3? It’s the company’s Simple Storage Service that lets you use its computing infrastructure for your own purposes. Ok, that’s pretty cool, but here’s what makes it what they like to call a “game changer”: It’s all you can eat data storage with no upfront costs and no minimum fees. You pay only for what you use.

The other amazing thing is how cheap they’ve made it. Storing 1 Gigabyte of data on their servers costs only $0.15 per month. Uploading or downloading that GB of data costs you just $0.20.

This begs the question: If you’re a startup company with the need for gobs of storage space, why would you even consider spending tens of thousands on servers and hard drives up front when you can just use S3 and build your business on the cheap? That’s what SmugMug (a Flickr competitor) did when faced with rapid growth and a need for lots of storage space, very quickly. The need for $25,000 in computing power and data storage was taken care of for a couple of thousand dollars a month with S3. They currently have over 300 Terabytes of photos.

Computer experts have talked of utility computing for years, but this is one of the first examples I’ve seen of the concept in practice. Amazon S3 absolutely makes data storage as easy as getting electricity or water – pay for what you use, when you use it. Nothing more. As others have started to realize, this hugely levels the playing field for anyone building a business where they need lots of data storage capacity. Variable costs are always better for small businesses than fixed ones, especially if they’re tied to customer demand and revenue.

I have no doubt that entrepreneurs everywhere are dreaming up of businesses based around using S3 on their backend. Tunesafe, an iTunes music library back-up tool, is just one that I found already. It’ll be interesting to see what comes next. Maybe Apple can build S3 storage into Time Machine