Eric,

Thanks for your comment. You are correct that Amazon needs to keep some capacity available for burst; however, this capacity can be as low as 20%. In addition, as your pool gets larger the % of reserved capacity shrinks. At Amazon’s current scale they likely need only 15% or less excess capacity in each availability zone. Also, FYI, I have spoken to many of the original EC2 team and the excess capacity is nowhere near 66%. It’s roughly 20%, as it is for most cloud providers.

I do not know Amazon’s margins, but I suspect them to be at least 50% or higher. In addition, the common trend I’m seeing right now is that hosting providers do the math and realize that cloud computing is a much, much, better business that dedicated servers. There are major efficiency gains for both capex and opex. In capex and opex, because using homogeneous hardware allows buying in bulk and also reduces costs associated with stocking parts and equipment for many configurations. There are major advantages to the hosting providers as well in the costs associated with running the cloud compared to a traditional hosting facility.

I’m confused about what you are describing. You’re comparing managed VPSes vs. unmanaged VPSes. A managed VPS is more of a managed service product than a straight VPS product. If you abstract away all of the management then of course customers won’t care. Regardless, I know for a fact that Amazon is eating away at the VPS market and that larger players see this and are responding to it.

I agree with your premise about the value of managed services and abstracting the underlying cloud further. It’s absolutely true that there is major value here and likely an area of ENKI’s focus, I presume. Regardless, customers will choose what they want and while many want the problem to just go away, others are looking for the ability to get into the guts because they have special requirements. These are typically not the same VPS customers however, but larger web application businesses that need secret sauce.

Best,

–Randy

I also disagree with the high value of control. We have quite a few erstwhile VPS customers in our cloud and what they want is not control but the opposite: freedom from having to be in control. They want us to manage everything: administration, auto-scaling, incident response, etc. Software has progressed to the point where you can run a high tech business successfully if only you could get away from the whole administration racket, but here we have Amazon training a new generation of “cloud administrators.”

I think VPSes will be here for a long time, both due to their price advantages as well as the fact that they have more mature support models.

The same thing applies to VPS vs Cloud. The VPS “product” doesn't have all the features of a cloud. Even if the customer doesn't want to invest in auto-scaling/code-as-infrastructure/fault-tolerence/etc right now, the customer may be willing to invest in “easy migration path to auto-scaling/etc down the road”. The customer pays an extra $10-$20 per month right now, but saves the pain of migrating from a VPS to a cloud provider next year.

Worse news for the VPS vendors is that they are fragmented (tools, walk-thrus, how-tos, resources, etc.) compared to the “one” AWS. That means customers may be willing to pay a premium for AWS. Are we ready to say “nobody ever got fired for choosing AWS”?

I totally agree that AWS hasn't even started to tap that market (and they could/should/will).