One of the advantages of a being a relatively ancient industry analyst is the ability to look back on over 25 years of innovation and spot the real paradigm shifts amidst the updates, revisions, and otherwise mundane changes that attempt, but never deliver, something radically new.Basking in the nerdacopia that is Microsoft’s Professional Developers Conference this week, which has marked Windows Azure’s coming out party, it’s clear that Azure is one of those shifts, on a par with just about anything Microsoft has done in my decades of covering the gang from Redmond, particularly with respect to enterprise software. The trick will be in getting the rest of the world to understand, and, true to the Microsoft model, prove Microsoft right by building the apps the justify, and thereby define, Azure’s ascendancy.
The only Microsoft products that I believe are comparable in their enterprise impact are SQL Server, which Microsoft didn’t even develop but licensed from Sybase in the mid-1980s, and then spent a decade or two trying to scale up to enterprise class functionality; and the way-too-prescient multimedia PC, which Microsoft co-created in 1990 without any concept of its “killer app” (the original demo showed a children’s video game, as-if), and then went on to glorious realization once the World Wide Web arrived to justify the by-then dead standard. (Okay, there may be other big products, but we’re digressing.)
Why Azure is so important boils down to two compelling needs in the enterprise software market: the need for an enterprise-class cloud to host a growing number of net-new enterprise applications emerging in the market, and the need for a means to leverage the existing Microsoft skill sets in the enterprise (and ISV community) towards the combined goals of lowering total cost of ownership and making cloud, and particularly hybrid cloud/on-premise computing, safe and reliable for the enterprise.
We’re beginning to see inklings of these net new applications running on Azure: some of them were shown last week to industry analysts at the Dynamics Analyst Summit in sunny Redmond (for a few brief hours, anyway). The basic takeaway from the Summit was that the Dynamics gang understand that Azure isn’t just a place to host existing Dynamics functionality (and definitely not the place to run AX in the cloud: one can rent Dynamics AX in a hosted model from a Microsoft partner, but it won’t be running on Azure), but a platform for a new class of enterprise application that can be simultaneously on demand and on premise, and deliver net-new functionality that can’t be readily delivered in an pure on-premise mode. It’s a good start.
But there’s going to more to cloud computing than deployment choice, or even hybrid on-prem/on-demand apps. As I’ve been writing a little obsessively, the cloud, particularly the Azure cloud, will reach its apogee as a platform for next-generation, on-demand applications that deliver functionality to the enterprise that can’t be delivered for love nor money on premise. This new generation of application aggregates data, process, and services in the cloud, and in doing so for its many stakeholders, becomes a self-improving, self-appreciating asset that grows in value the more data, processes, services, and stakeholders it accumulates.
This happens in two ways. The first is that the aggregation of data, services, processes, and stakeholders in the cloud lowers the cost of the transactions it supports as it simultaneously increases their potential value. For example, a supply chain in the cloud is cheaper to manage the more its stakeholders are using the same transaction platform, even as the accumulation of stakeholders makes it possible for participants to achieve economies of scale that also improve performance and functionality across the supply chain. That’s already possible, and already in place, albeit in a few lonely examples.
The second, perhaps even greater value, is that the data the accumulates in the cloud becomes more and more valuable as it grows in quantity and quality. That supply chain in the cloud begins to accumulate a tremendous amount of data about the behavior of the stakeholders, data that can begin to drive predictive models about how the supply chain functions. Those models can be used to improve performance, lower rates, create new service offerings, and otherwise make the cloud a crucible for process improvement, best practices, and new functionality that wouldn’t be possible without the aggregating effects of the cloud. A supply chain in the cloud would thus also be able to use statistical modeling to predict the behavior of the stakeholders, and in doing more carefully define the risk profile of the supply chain, its stakeholders, and, even, an individual order. Those risk profiles would have a genuine monetary value: if you could prove that your supply chain cloud raised your perfect order rate, you can take those data to the bank and get a lower loan rate, and then take the data to your insurer and get a better insurance rate too. Repeat ad infinitum across the entire supply chain.
That’s only one example in a single industry: I think the analysis of the behavioral data in these commerce cloud sites would have a significantly greater and more genuine value then all that click-stream analysis that Google (and many others) now performs. The difference lies in the direct correlation of this cloud behavior to specific commercial transactions, as opposed to the much more flaky notion that “page-views” equal actual views, which is the basis of so much of the analysis of consumer activity that is driving the over-evaluation of Google today.
Back to Azure: as a platform for enabling this kind of application, and its ensuing value-add, Microsoft has several distinct advantages. One is an understanding of enterprise-class computing. The Azure gang gets that security, reliability, high-availability, and all those other things that are important to IT are built into Azure. Remember, unlike Amazon and Google, Microsoft lives in the IT department, products like SQL Server power some of the world’s largest databases, and this experience infuses Azure with a sense of enterprise purpose that the consumer cloud providers simply can’t find when they sequence their relatively short DNA strands.
The other advantage for Azure comes from Visual Studio, and every other dev tool and service that falls under the somewhat ubiquitously confusing moniker of .NET. I sat in several rooms this week at PDC and watched that geekiest of geek crowds absorb the relative simplicity of converting their existing skillsets to Azure. To say Azure is a couple of clicks away from most of the several million .NET apps developers is only the slightest of exaggerations: while they all won’t be able to build the kind of app I’ve outlined above, assembling a team of developers to build a next-gen Azure app will be trivial compared to, say, assembling a crack team of Apex developers to build something comparable on Force.com. And when you add the growing library of existing Microsoft services available on Azure – such as SQL Server, as well as Dynamics xRM, and many more on the way you’ve got the cloud platform Marc Benioff thought he could dreamforce into existence (and couldn’t, and won’t.)
So, a few thousand words into this post, let me return to the initial premise. Microsoft is sitting on top of the next generation application platform the one that, like the PC, before it, will define not just the future of Microsoft but the future of enterprise software as we know it and can only envision it for some time to come. This won’t be 1985 all over again, when the debut of Windows presaged an unparalleled, and so far unbeatable, monopoly on the desktop. Others can, and will, figure how to do an enterprise cloud a la Azure. But when you put it all together enterprise-class cloud, value-added services, millions of developers it’s a pretty impressive, and hard-to-beat combination. Now it’s up to the ISV and developer community to make those killer apps real, and prove Microsoft right. The future of enterprise software hasn’t looked this interesting in a long long time.