The company this week announced its Reserved Instances for its EC2 service. The new service, to be based on Linux and Unix systems, will start at $325 for a one-year period but could drift up as high as $2,600, depending on how powerful a server you want. The price for a three-year deal starts at $500.
The appeal of the service, company officials believe, is it allows IT shops to reserve both processing and storage capacity ahead of time, instead of rolling the dice on whether it will be available when they actually need it.
The other element they hope will lure companies is the money it can save users compared with the existing model of charging per hour.
"Customers can choose to reserve capacity as if they owned the hardware, but unlike traditional infrastructure, with Reserved Instances, customers do not pay to maintain and operate idle hardware, and instead pay usage charges only when actually utilizing the instances," company officials said in a prepared statement.
Reserved Instances for Unix and Linux is currently only available to customers based in the United States.
IT shops that do sign up for the service do not need to buy their own servers, company officials said. Whatever capacity they reserve automatically comes with Amazon’s current service-level agreement of 99.95% availability for a 365-day period.
Reserved Instances will not replace the company’s existing, pay-as-you go model, but rather exist alongside it.
Amazon noted that payments made for unused or partially used Reserved Instances will not be refundable.
Before you jump into Reserved Instances, you might want to read about Amazon's S3 two-hour crash earlier this year. InformationWeek has published an independent analysis of this topic. Download the report here (registration required).