Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Cloud prices: How low can they go?

Brandon Butler | April 1, 2013
Competition has led to 29 price reductions in 14 months among big-four cloud providers

"Compute and storage, the most mature services, saw smaller reductions than newer cloud services like database," says Kim Weins, RightScale's VP of marketing. "This could indicate that cloud providers have already squeezed more efficiencies in these areas, and future price reductions may be smaller." Still, Weins predicts that cloud prices will continue to fall for the foreseeable future, especially as hardware prices continue to drop, scale increases and providers become more efficient in running their services.

As a comparison, AWS's default virtual machine (VM) size costs $0.06 per hour for Linux machines and $0.15 per hour for Windows machines, and come with 1.7GB of memory and 160GB of local disk space. Rackspace has a 1GB VM with 40GB of local disk for $0.06 per hour, or a 2GB memory machine with 80GB of local disk space for $0.12 per hour. Google Compute Engine's default VM has 3.75GB of memory and 420GB of local disk space and costs $0.138 per hour, while Microsoft Azure has a standard 1.75GB VM for $0.08 per hour.

Perhaps the cloud wars hit a head late last year where during the course of 11 days three of the major cloud providers each dropped their storage prices by one-third. On Nov. 26 Google announced a 20% reduction in its storage costs, on the eve of Amazon's first-ever user conference, named re: Invent. During the conference three days later, AWS dropped its storage costs by 25%. That same day Google announced another price drop of 10% more. Not to be outdone, a week later Azure dropped its prices 22%. The back and forth points to the jockeying cloud providers mangle with one another, and the fight for headlines when one company makes an announcement.

RightScale released the study in conjunction with a new cloud price forecasting feature the company launched this week named PlanForCloud. Customer input their expected cloud usage and the system predicts how much it will cost across different providers.


Previous Page  1  2 

Sign up for CIO Asia eNewsletters.