Amazon's and Google's compute, storage and analytics cloud pricing just keep on dropping and dropping — so who's offering the better deal?
Low upfront costs are one of the benefits of using cloud services from Amazon, Google, Rackspace and Microsoft, yet in-house hardware — which at datacentre scale has also seen massive price reductions over the year — still undercuts the cloud for some enterprise workloads.
Nevertheless, all major cloud providers have been slashing prices over the past year, with the latest and largest cuts coming from Amazon in response to Google's own reductions the day before.
Amazon took the axe to EC2 cloud computing prices for both on-demand and reserved instances; as well as S3 cloud storage, with the biggest cuts to the smaller requirements; Amazon's Relational Database Service; and its cloud analytics platform, Elastic MapReduce. All price cuts will come into effect from April 1.
Amazon's new on-demand pricing for M1 (current generation), M2 (memory optimised current generation), C1 (previous generation) Linux/Unix instances on EC2 will fall between 10 to 40 percent, while its Windows pricing for those instances falls between seven to 35 percent.
Amazon's one and three-year reserved instances for Unix/Linux also saw overall price cuts of between 10 to 40 percent for M1, M2, C1 instance classes, and 30 percent for C3 and M3 classes. There are also smaller reductions of between 13 percent to 23 percent for Windows reserved instances.
Customers on light utilisation reserved instances should take note as Amazon appears to have removed that altogether, with only medium and heavy remaining after April 1.
Cuts to S3 pricing range from 65 percent at the 0 to 1TB range, bringing the cost to 3 cents per GB per month while storage requirements above 5,000 TB have been reduced by 36 percent to 2.75 cents per GB per month. (Customers in non-US regions should check their individual pricing.)
Elastic MapReduce pricing has also fallen between 27 percent for small instance types to 61 percent for larger optimised instances.
"With this price reduction, you can now run a large Hadoop cluster using the hs1.8xlarge instance for less than $1000 per Terabyte per year (this includes both the EC2 and the Elastic MapReduce costs)," Amazon Web Services said in a blog post on Wednesday.
Amazon RDS DB instances will see a reduction of an average of 28 percent.
But really, hearing about a provider's price cuts without being able to directly compare them against prices for a rival's equivalent products is pretty useless. Fortunately, RightScale, a cloud portfolio management company, has done this with two very useful tables.
Two handy observations RightScale makes are that while AWS matched Google's new on-demand pricing for standard compute resources and offered comparable pricing for higher memory and high CPU options, a few differences may play out when it comes to Amazon's reserved instances.
Since Amazon requires a one to three-year commitment on reserved instances, should it in future reduce prices, customers on these won't necessarily get the lower price. As RightScale notes, Amazon has a record for not passing on discounts.
"In contrast, Google sustained-use pricing is calculated as a percentage of the on-demand baseline rate. As Google’s baseline rates go down, the sustained-use prices will fall as well. As a result, all future Google price drops are passed on to all customers immediately when they take effect. Given the continual price reductions in cloud pricing, you may find that the ability to take advantage of future Google price drops might bring the price below the AWS heavy RIs," RightScale adds.
Here's the company's two tables for on-demand and reserved instances to make a like-for-like comparison.