AWS and Competitor Pricing Comparisons: Sound and Fury, Signifying Nothing

People are spending lots of time contrasting AWS costs with other cloud competitor costs in pseudo “true workload” comparisons and I guess I am a bit bewildered by it all. It seems like the comparisons are random and contrived. Even more, I question why they would matter even if they were valid.
I say they are contrived because the results can be easily skewed by the selection criteria. StackDriver ran a performance test with one set of results and now other groups claim the opposite results. This points to the ability of the tester to predetermine the outcome – and it makes sense when you think of the following three facts:

1. There are too many variables for a true comparison. Different instances types offer differing compute/memory/network capacities. I – or anyone else – can easily cherry pick instance choices along workload times and/or characteristics to make AWS win or lose the cost competition. The comparisons routinely choose a “standard” instance type from AWS while ignoring the optimal instance type.
2. AWS offers multiple pricing types. If you choose On-Demand as the AWS pricing type, AWS may occasionally lose the comparison. However, sophisticated users also rely on Reserved Instances and Spot pricing. These prices tilt the balance. Further, even in comparisons using Spot pricing, the Spot price fluctuations make the tester’s “choice of time” dispositive to the outcome.
3. Amazon’s core DNA is low cost. You see it in all their businesses and especially in their consistent price cuts. That is what they do. Add the facts that infrastructure providers benefit from economies of scale and that no one has a better economy of scale than AWS, and it just seems impossible that anyone can underprice AWS by a significant margin on a consistent basis.

So, for those reasons, I yawn at the pricing comparisons. They are random and show that if you hunt long and hard enough, you can some peculiar circumstance that fits your desired outcome.
However – let’s imagine that there really is repeatable scenario where someone else is marginally cheaper. Even if that was the case, I would say who cares? It is not important.
Why do I say that? Consider this:
These comparisons act as if all the cloud providers offer equal services. Well, that is obviously false. AWS offers the broadest variety of services and functionality. It has the richest ecosystem and continues to lead the way in innovation – and the competitors are not even close.
And – I also know why my company went to the cloud. Cost of running servers was important, but hardly decisive. In fact, like most business users, my company recognized the business agility that was gained by using the cloud.
We all know that there are business costs in time, effort, and money associated with hopping workloads from cloud to cloud in order to find the right service mix. Similarly, there are business gains from being with a versatile provider like AWS.
The comparison crowd does not want to count these business costs and benefits. Conversely, I think they are decisive and easily dwarf any differences of pennies on server price. My business needs the agility that is available from AWS’ myriad services. I am convinced that – for a large diverse workload – AWS is consistently cheapest. However, even if it was not, the AWS business advantages would still easily support my usage.
 

See CloudCheckr in Action

The largest enterprises, service providers, resellers, and government agencies trust CloudCheckr’s cloud management platform. Schedule a 30-minute demo to learn what CloudCheckr can do for you.