Amazon Cuts Cloud Computing Prices
Amazon Web Services lowers prices 5% to 37%, reducing the cost of an entry level compute unit and enticing largest customers with volume deals.
Amazon Web Services has lowered the price of its basic unit of on-demand computing by 5% to 10% and the price of its reserved instances--the workloads that promise one to three years of steady use--by up to 20%.
Jeff Barr, Amazon's senior web services evangelist, posted a blog about the reductions late Monday, with notice of the blog going out at 10:47 p.m. In it, Barr placed the cuts in a context of Amazon's ongoing effort to lower the cost of various services. Adam Selipsky, Amazon VP of product management, previously told InformationWeek in an exclusive interview that price cutting was part of the company's retail DNA.
More Cloud Insights
- The Untapped Potential of Mobile Apps for Commercial Customers
- Secure Cloud: Taking Advantage of the Intelligent WAN
White PapersMore >>
But the leading cloud service supplier has also come under increased scrutiny recently for its pricing structure for some services, such as its S3 storage service. Art Wittmann, managing director of InformationWeek Reports, compared S3 pricing to the rapid decrease in the price per gigabyte of on-premises storage.
"In 2006, the cost of running a small website with Amazon EC2 on an m1.small instance was $876 per year. Today with a High Utilization Reserved Instance, you can run that same website ... at just $250 per year--an effective price of less than 3 cents per hour," Barr blogged. Amazon's entry level server, a small instance, is defined as a virtual CPU equivalent to a 2007 1-1.2 GHz Xeon processor with 1.7 GB of RAM and 160 GB of server disk storage.
[ Want to learn more about the pluses and pains of cloud computing? See Four Cloud Computing Pain Points That Still Hurt. ]
The price on an on-demand small server running Linux dropped from 8.5 cents an hour to 8 cents; Windows dropped from 12 cents an hour to 11.5 cents.
A large instance, with four times the CPU of a small instance, 7.5 GB of RAM and 850 GB of server disk, was formerly priced at 34 cents per hour for Linux and 48 cents per hour for Windows; they're now 32 and 46 cents, respectively, a decrease of 5.9% and 5.6%.
Extra large, with eight times the CPU of a small instance, 15 GB of RAM and 1,690 GB of server disk, was formerly 68 cents per hour for Linux, 96 cents per hour for Windows; they're now 64 and 92 cents respectively, decreases of 5.9% and 6.1%.
But the biggest price cuts came for users of large numbers of reserved instances--the customers who put lots of EC2 infrastructure to work on a regular basis.
The previous price drop in Amazon's on-demand instances took effect Nov. 1, 2009.
"If you own more than $250,000 of Reserved Instances, you qualify for a 10% discount on any additional Reserved Instances you buy. ... If you own more than $2 million of Reserved Instances, you qualify for a 20% discount. ... Once you cross $5 million in Reserved Instance purchases, give us a call and we will see what we can do to reduce prices for you even further---we look forward to speaking with you!" wrote Barr.
That may mean that Amazon Web Services has found itself with spare capacity on hand, as a major customer such as Zynga switched from relying 80% on Amazon infrastructure to 20%, while building out its own data center capacity. Or it may mean it can add more capacity than it used to at the same price.
The single largest price reduction was the example of an extra-large, multi-Amazon Relational Database Service run continuously. Customers can save "more than $445 per month or 37% by using a 3-year, Heavy Utilization Reserved Database Instance," Barr wrote.
Prices were also reduced on Amazon Elastic MapReduce, a distributed file sorting system; Amazon Relational Database Service; and Amazon ElastiCache, the scalable, in-memory caching system.
There were some variations by region in how the price reductions were implemented, Barr acknowledged. AWS has implemented service price reductions 19 times since it was launched in 2006, he added.
Find out how to move beyond server virtualization to build a more flexible, efficient data center in the new Private Cloud Blueprint issue of Network Computing. (Free registration required.)