Microsoft works to claim its own slice of the cloud

It would appear to be the most lucrative new platform in all of computing: the "cloud" -- the space on the Internet from which applications and services can be presented to customers without the need for physical location. It was once called the "grid," but the fact that companies other than IBM managed to effectively rechristen the idea speaks to its inability -- along with everyone else's -- to build a clear and concise message around just what the cloud is.

No one really knows, at least not completely. That's the message we've seen thus far from nearly every major vendor in the cloud space, and it's the message we're seeing from Microsoft as well. Even with a full business plan for Windows Azure, the company's platform for .NET Services in the cloud, its own platform evangelist admitted to Betanews that much of the projected purpose for the service still remains a mystery. Microsoft usually undertakes a platform buildout by leveraging its resources from other platforms (Internet Explorer from Windows, SharePoint from Office, Exchange from Outlook, etc.). And it would seem, at least on paper, that Azure would be leveraged from .NET.

But it's a stretch, and Microsoft knows it. So it's looking to developers to help fill in the blanks. In an interview with Betanews, Microsoft's Director of Platform Strategy for its Developer & Partner Evangelism team, Jamin Spitzer, explained not only Azure's marketing appeal to .NET developers in the sense of allure, but also Microsoft's appeal to developers in the sense of a plea: How should Azure work? He had some surprising answers, which included not only "It depends," but the fact that in some cases, it might not work at all.

SCOTT FULTON, Managing Editor, Betanews: The way an insurance company once phrased it -- one that's large enough to become a Windows Azure customer at some point in time: "When we get a call from a salesperson to come adopt their software or service or platform or cloud, the question we ask them is, how much can we use what we have? How can we leverage not just our talent but our existing assets to float a better system than the one we have, without completely amortizing our investment in everything that's come heretofore?" And what they notice with cloud offerings, to use that Damien: Omen II language, is that they don't appear to leverage very much. They ask the customer, usually, to start completely over. The exceptions being when a cloud service provider like Amazon comes along and says you can deploy what you've got now and leverage 100% of everything you've got in their cloud, which might make it less expensive but it doesn't make it any better.
So what are you leveraging here?

JAMIN SPITZER, Director of Platform Strategy, Microsoft: Part of the answer is, it depends. It depends on how the app was written. It depends on the languages that need to be supported for that app. It depends on the skills and the data schema, and I know the answer, "It depends," is not necessarily the best answer you can look for, but the answer honestly is, "It depends."
I know for example that Full Armor, one of our partners, was able to reuse the vast majority of their applications, and that's why they are one of the first early adopters of Azure that we're able to talk about. And that's because it's an ASP.NET app that took advantage of principles like decoupling and scalability and virtualization, perhaps when other people weren't necessarily thinking that way or writing that way. So I would say part of that is good forethought on the part of Full Armor, and I think part of that is, they put a bet down and they said we're going to build this way because we like the way this builds. And the market caught up to them.

SCOTT FULTON: Well, if the others weren't thinking that way or building that way at the time, then obviously, it would follow that this was an unusual case. Full Armor was already thinking well, well ahead of the field. And it therefore follows that the field is still stuck about three or four years back, stuck in 2005. So getting them into a position where they can leverage their assets to the degree that Full Armor did, is actually going to take...what would you call it, evangelism.

JAMIN SPITZER: It's going to take some evangelism, it's going to take some work, sure. The next generation of cloud computing and cloud computing apps is around taking advantage of that scale, that elasticity. When you look at what Amazon is doing...that provider-hosted model -- particularly the platform piece -- there's a lot of really good things going on there, a lot of things made possible by virtualization, by some of the efficient ways in which Amazon is doing what they do. But the app is still an instance-based app, right? And there are certainly a lot of things that you might be able to do with a highly efficient data center and virtualization that can make that app better and more cost-effective, but the Amazon model still requires a fairly robust set of systems admin capabilities.

We still don't really know that much about how application and services management is going to take shape, once you have lots of apps running lots of different places, and how you manage those. The scale of the app is largely the same; you virtualize the app, but it's still instance-based, you've separated the app from the hardware, but the app itself is still what we would call a scale-up app.

And to your point earlier, maybe you can save some money, but...in a lot of circumstances, they can't. I had the opportunity to meet with a financial services customer not too long ago, and he said that [his company] had a good couple of thousand apps running in its data center. And he said, for the vast majority of them, he did the math and determined that the cost benefits weren't there, not just in terms of what do I have to invest to get it to the cloud, but just the ongoing operational expenditures. He felt like he was running his data center at an efficient enough state, they were a large enough company with a lot of in-house skills. He felt fairly confident in the statement that most of his apps are better where they are.

Next: The whole "scale" argument, thoroughly dissected...