Amazon tweaks EC2 pricing

This was originally published at ZDNet's Between the Lines.

Amazon has tweaked its Elastic Compute Cloud (EC2) pricing model to be more enterprise-friendly. The move is significant enough to sway IT executives to adopt more of Amazon's Web Services-especially when they have tight budgets.

Amazon on Thursday announced reserved pricing for its EC2 instances. Simply put, customers can reserve instances for one-year and three-year terms as if they owned the hardware. Enterprises can guarantee they have an EC2 instance for computing power they know they'll use and buy on the spot market to account for spikes at the usual Amazon rate. Under Amazon's model, customers only pay for the computing power they use even if instances are reserved.

Enterprises have begun to use Amazon's Web services more in recent months and the economic downturn has only accelerated those moves. However, enterprises view the Amazon EC2 as a way to add capacity for spikes and pilots. These corporations haven't been swayed to put more of their infrastructure on Amazon's platform because they are still evaluating the feasibility and return on investment.

Peter DeSantis, general manager of Amazon's EC2 business, said that the company's latest pricing may help the return on investment case for enterprises. By reserving instances under 1-year terms the savings are roughly 30 percent for customers. A 3-year term adds up to be about a 50 percent savings.

"We had a lot of the feedback from enterprise customers. They want to know that their data centers have boxes when they need them," explained DeSantis, who noted that there is a big menu of pricing options. "These broad options for enterprises appeal to what they are used to. The pricing model resonates to them. unanticipated needs. Best of both worlds."

In practice, technology executives can now have the comfort of knowing they have a set number of instances reserved without the added investment of adding upfront capacity.

Here's a look at the reserved vs. spot market EC2 pricing:

Depending on how enterprises mix and match these pricing alternatives, the savings can add up. A Linux/Unix box paid for hourly will run you 40 cents an hour for a standard large on-demand instance. That equates to about $3,500, assuming 24/7/365 coverage. If you reserve that instance over a year term Amazon will charge you $1,300 per instance plus the 12 cents per hour. That comes out to $2,351. Multiply those savings ($1,152) over 100 instances and you get some real savings: $115,000 a year or so.

A three-year option under the same assumptions yields an on-demand price of $3,500 a year ($10,500 over three years). A reserved instance is $2,000 plus 12 cents an hour. With around the clock coverage you get a price of $5,153 per instance over three years. Savings of on-demand compared to reserved per instance: $5,346. Across 100 instances, the savings comes out to $534,600 over 3 years.

Add it up and it's pretty clear that Amazon is using the downturn to be quite disruptive to entice enterprise customers. Recessions force enterprise buyers to get creative, and cloud computing is already on the radar. Even so, Forrester reports that only 5 percent of large enterprises have implemented cloud computing instances (but 46 percent are interested). Part of the problem is that the cloud model is fairly new and there are some uncertainties such as licensing, pricing, and the fact that services like Amazon EC2 are relatively new.

However, that's changing. Amazon has partnered with almost all of the big enterprise players--IBM, Capgemini, Salesforce.com, Sun's MySQL, and OpenSolaris, Oracle and Red Hat--and clearly is a player in the corporate cloud computing plug.

Meanwhile, Amazon is taking its approach to e-tailing--customers first, adjust to meet needs and keep innovation--to the enterprise market. As a result, Amazon is taking away the excuses to not try its cloud services.