Thank you for the spirited discussion and the opportunity to rebut.

A lot of the hype around the Cloud is energized by the myth of “something for nothing” or nearly nothing. Do you really think Amazon isn’t “in it for the money?” The “dirty” secret around cloud computing is that cloud providers make their money off base load while hyping the savings of peak load, and that most users are base load users, not peak load users. Companies with instant 10x peaks or hockey-stick grown are the extreme minority, only about 2-3% of our customer base. How do you think that Amazon had the ability to charge a fraction of a dollar per instance hour when they had a huge inventory of servers standing by – much larger than the actual servers making money, at least in the early days? They did it by buying the market and setting their prices below their costs. I know this because I do the costing for ENKI and no economies of scale could possibly lower Amazon’s costs to 1/10th or less of my costs if I kept that sort of inventory (just the power is a killer!). They were investing, expecting a return, just like any publicly held company. And the expected return is determined by the stock market, not their altruism or even their profit motive. They expected the return to come once a large percentage of their resources were *not* idle, but supplying base computing load to customers. I hope, for everyone’s sake, that this is happening for them because it would be bad indeed for the cloud market if their gamble didn’t pay off.

The issue I was bringing up with my comments is that the price-based cloud which Amazon pioneered is focused on the developer. Go to any cloud computing conference and it’s all about the developer. As a result, there’s no end-to-end consideration of how to provide value to the enterprise, most of which don’t develop but rather purchase and deploy software. And there’s no discussion about how Amazon’s opaque architecture creates up to 200ms latencies between instances (and S3) that completely eliminate the ability to host whole classes of applications, or about the continuous stream of customers who I see coming to ENKI that just can’t get enough service from Amazon to get them over the hump to getting an application deployed. That isn’t any way to capture the interest of the C-level folks who are concerned with the value cloud can provide to their enterprise, not dollar cost per instance or even whether provisioning is automatic or manual. It’s not that Amazon is wrong or even bad, it’s just that the cloud market is separating and segmenting into different types of clouds aimed at different customers. Which is why I took issue with your one-size-fits-all definition.

But in the final analysis you’re trying to make a virtue of complexity, and i would argue that I am trying to make a virtue of simplicity. Somewhere in the middle is the value statement.

Honestly- I did’nt find Staten’s definition to be all that compelling, its a very data center centric view – and the cloud is going to shape that up. its not all going to be on premise private clouds.

Now I look at Cloud with some of the same jaundiced eyes. ASP models can deliver terrific value, but somehow they have become cloud based. All with no significant changes. I wonder how they managed that?

Behind that lightness there is a serious comment. That is, that for all computing that we need to have done, we need to select the most cost effective platform for it. Sometimes it is in-house and hosted. Sometimes it is ASP, where the functionality and the platform are externalized (think Domain as a Service), and sometimes it is cloud based where a couple of forks occur. We can develop applications specifically for the cloud – so the cloud is just the base platform, or we can run purchased apps in the cloud – essentially ASP with a different deployment model.

Listen, I think the cloud very quickly outgrows the “proof of concept” stage of super simple demonstrations of the cloud’s power, and moves into a technology advancement phase, where more advanced solutions are possible, at the cost of more complexity. Typically, this game is played by both consultants and start ups, resulting in plenty of cr*p, but also the occasional revolution.

Its all a complex adaptive system, James, and we are all just agents along for the ride.