Startup idea: Amazon S3 / CloudFront cost distributor | Hacker News Reader

Startup idea: Amazon S3 / CloudFront cost distributor

5 pointsjeff185 comments

This seems too simple, so I may be missing something but here goes.

Right now Amazon charges $0.15 / GB of Amazon CloudFront data transfer. However, if you do a higher volume, it gets progressively cheaper, until at over 1000 TB, you are only charged $0.03 / GB.

Someone should create a competitive CDN that sells premium bandwidth for say, $0.05 / GB, dramatically undercutting everyone in the business by a huge margin.

How would one or two co-founders possibly compete with Amazon CloudFront in quality and price? The backend would actually be Amazon CloudFront! Since you would rapidly exceed the 1000 TB amount, you'd be able to pass the savings onto your customers.

To be 100% sure that you make a profit, you could wait until you get enough people to pledge to switch to your service before launching for it to become profitable from day 1.

Sign me up for 10 TB of usage.

This is essentially what EngineYard does with their Cloud offering and how they've been able to reduce prices recently:

We’re able to provide this new lower pricing because of a new capability from Amazon Web Services that allows us to group all our customer accounts into a single billing entity with Amazon. So, we’re now able to use a single virtual pool of reserved instance capacity from Amazon, and pass along these savings to customers. This means all of our customers get to reap the benefits of reserved instances without the need to pay the reserved instance costs up front. No other cloud platform provider has been as aggressive as us in slashing prices as our cost goes down.

FROM: http://www.engineyard.com/blog/2010/announcement-engine-yard...

Ah, arbitrage. I have long wondered what Amazon's justification is for this tiered pricing; their cost per GB should be the same for all customers.

Also see previous discussion: http://news.ycombinator.com/item?id=960123

My assumption is that they're working to recover fixed per-account costs, and anticipated support costs.

When I've done tiered pricing, that was my logic.

What's stopping you from doing this? If you can get enough people to sign to be profitable from the start it seems like you don't have much to lose.

I'm busy with my own startup. :D If I had more time, I'd like to though.

I already do this to some extent, on a small scale, e.g. I share a dedicated server with several other people, so it is much cheaper, but it would be great to commoditize raw bandwidth on a larger scale.

Reply on news.ycombinator.com