Earlier today, HPC in the Cloud examined Rackspace’s response to Amazon’s price reductions for their dedicated servers. So what are the ramifications of those reductions?
With the Rackspace’s announcement this month that they were getting involved with some of CERN’s cloud computing and the news of IBM acquiring SoftLayer last month, Amazon’s HPC instances may be facing some competition in the not so distant future.
To battle that impending competition, Amazon reportedly dropped their price to rent their dedicated servers, machines that operate solely for one institution. Specifically, according to Amazon, they implemented an “80 percent price reduction from $10 per hour to $2 per hour in any Region where at least one Dedicated Instance of any type is running.”
Further, they reduced the cost of on-demand dedicated instances amounting to a “reduction of up to 37% in hourly costs. For example the price of an m1.xlarge Dedicated Instance in the US East (Northern Virginia) Region will drop from $0.840 per hour to $0.528 per hour.”
Reportedly, Amazon is aiming at mid-sized enterprise customers with these cuts. Those clients would be the ones more likely to run HPC experiments in a cloud environment. A dedicated server, where partitions that inhibit performance do not factor into the equation, would be more ideal for high performance applications.
When that announcement was made at the beginning of the month, the markets responded as Rackspace’s stock dropped 8.5 percent. Rackspace responded to that development by noting how different their dedicated servers were from Amazon’s.
That difference could prove to be key as the two move forward in their quest to conquer the cloud market. While Rackspace utilizes a hybrid cloud structure, Amazon’s dedicated EC2 instances take place entirely in their datacenters.
“Dedicated Instances are EC2 instances that run on single-tenant hardware dedicated to a single customer account,” they note on their website. “They are ideal for workloads where corporate policies or industry regulations require that your EC2 instances be isolated from instances that belong to other customers at the host hardware level.”
Amazon’s argument from an HPC perspective is that their dedicated servers, even if they are all located in a public cloud where the latencies may be higher than a hybrid setup, is more financially feasible for those mid-sized institutions who would run such applications in the first place. Further, Amazon’s HPC instances are still where the majority of high performance cloud experimentation is taking place.
However, the debate is young and will assuredly deepen over the months and years to come.