Mosso challenges Amazon on cloud storage
The assumption made by many who follow cloud computing is that Amazon has changed the game with unmatched pricing and service offerings. However, Mosso today released a study that questions how true that is for storage and CDN.
On their blog today, Rackspace's cloud division, Mosso, shows off a study they did where they compared the costs and performance of Amazon Web Service's S3 storage service and CloudFront Content Delivery Network (CDN) against Mosso's combination of CloudFiles and their partnership with CDN provider, Limelight Networks. The blog post presents five common use cases, and compares the cost of CloudFiles/Limelight with the Amazon offerings, both with and without Amazon's support option.
I spent some time on the phone yesterday with Mosso co-founder, Jonathan Bryce, and Senior Cloud Architect for Rackspace's cloud division, Erik Carlin, discussing what they found. The short-short version is that, for the five use cases they analyzed, they claim (not surprisingly) that Mosso beats Amazon's offerings in simplicity, cost and performance, especially when support is taken into account.
From a simplicity standpoint, one simply needs to take a look at the pricing charts used in the study's calculator:
While Amazon may be trying to be flexible for a wide variety of customers, the complexity of their a la carte style pricing may be a disadvantage over their smaller, customer service centric hosting-turned-cloud competitors.
From a cost perspective, Mosso's post speaks for itself. I won't dive into it in detail, but I will point out that Amazon's choice to charge extra for support works against them in a big way here. Here's an example from one use case:
As Bryce noted in our conversation, "Support is intrinsic to enterprise cloud computing, and you shouldn't have to pay extra for it."
John Keagy, CEO of GoGrid, agrees, as he notes in an interview with Robert Scoble posted yesterday. Personally, while I guess there are those that won't want formal support, I think that cloud computing is a move from a product to service mentality. Providing a number to call if something goes wrong just seems like the heart of such an offering in enterprise IT. I'm surprised that Amazon hasn't already announced a "no additional cost" basic support tier.
One final thing about cost. I was surprised to learn that Mosso gives you the CDN and support services at no additional cost above using the storage services alone. No additional per-Gb charges, no additional bandwidth charges. Again, simplicity of pricing compared to a completely ala carte scheme.
The performance numbers are perhaps the most interesting aspect of the report. If you were wondering if Amazon could compete with the technical focus of an Akamai or Limelight, here's your answer. You can read the report for the details, but essentially Mosso ran the five use cases through basic performance tests, and found dramatic advantages for Limelight over CloudFront.
I am a big fan of Amazon Web Services, and I think they are constantly defining what is necessary in a hybrid IaaS/PaaS offering to meet a wide variety of cloud computing demands. However, while I admire their pioneering ways, I am aware that there are alternatives out there., now Mosso makes a clear case that, for storage and CDN, you really should shop around.