Always read the fine print

When Amazon announced Glacier I took the trouble to read their pricing information carefully and wrote:

Because the cost penalties for peak access to storage and for small requests are so large ..., if Glacier is not to be significantly more expensive than local storage in the long term preservation systems that use it will need to be carefully designed to rate-limit accesses and to request data in large chunks.

Now, 40 months later, Simon Sharwood at The Register reports that people who didn't pay attention are shocked that using Glacier can cost more in a month than enough disk to store the data 60 times over:

Last week, a chap named Mario Karpinnen took to Medium with a tale of how downloading 60GB of data from Amazon Web Services' archive-grade Glacier service cost him a whopping US$158.

Karpinnen went into the fine print of Glacier pricing and found that the service takes your peak download rate, multiplies the number of gigabytes downloaded in your busiest hour for the month and applies it to every hour of the whole month. His peak data retrieval rate of 15.2GB an hour was therefore multiplied by the $0.011 per gigabyte charged for downloads from Glacier. And then multiplied by the 744 hours in January. Once tax and bandwidth charges were added, in came the bill for $158.

Karpinnen's post is a cautionary tale for Glacier believers, but the real problem is he didn't look the gift horse in the mouth:

But doing the math (and factoring in VAT and the higher prices at AWS’s Irish region), I had the choice of either paying almost $10 a month for the simplicity of S3 or just 87¢/mo for what was essentially the same thing,

He should have asked himself how Amazon could afford to sell "essentially the same thing" for one-tenth the price. Why wouldn't all their customers switch? I asked myself this in my post on the Glacier announcement:

In order to have a competitive product in the the long-term storage market Amazon had to develop a new one, with a different pricing model. S3 wasn't competitive.

As Sharwood says:

Karpinnen's post and Oracle's carping about what it says about AWS both suggest a simple moral to this story: cloud looks simple, but isn't, and buyer beware applies every bit as much as it does for any other product or service.

The fine print was written by the vendor's lawyers. They are not your friends.