Microsoft’s cloud partnership with Oracle gives the Redmond, WA giant a new license to distribute Java technology — something it hasn’t had since 2004. With this license, Microsoft can finally offer Java as a preinstalled, first-class environment on Windows Azure as well as its Hyper-V virtualization software. Microsoft and Oracle announced their cloud partnership on June 24.
Developers can and do run Java applications on Azure, but strictly as a “bring your own” exercise; they were not able to get usage rights to and support for the technology from Microsoft. Without a license, Java would always be a second-class denizen of Windows Azure. Now enterprise customers can get the legal air cover they require for crucial technologies. As a result, we expect to see much wider usage of Windows Azure for a wider range of Java applications. 
For both Microsoft and Oracle, life is about the enterprise. And Java is a foundation of enterprise applications. In Forrester’s Forrsights Developer Survey, Q1 2013, Java support was a top-five reason to select a cloud platform. Enterprise C# developers will naturally come to Windows Azure, but Microsoft needed to attract the Java shops as well. The most popular cloud platform — Amazon Web Services — has no problem with Java; neither does Google. And in the real world, most large enterprises use both .NET and Java, so why should developers need different public clouds for each?
The license also gives Microsoft protection from the sorts of legal actions that in 2004 cost it a $2 billion settlement with Sun Microsystems (since acquired by Oracle). 
For the vendor behind .NET, acquiring a new Java license is an indication of commitment to the technology important to many developers. Still to be revealed is how much effort Microsoft will put into helping Oracle and the Java community develop future versions of Java. Will Microsoft join OpenJDK, Oracle’s open-source project for the Java language and runtime, and commit engineers? We doubt it. 
We believe that Microsoft wants Java for enterprise completeness but is happy to let Oracle support and drive it. Which brings us to the quid pro quo in this partnership for Oracle. How does revenue flow to Oracle from this deal? Here’s what we think:
  1. Windows Azure is the No. 2 cloud platform in Forrester’s surveys and so provides a new channel for growth for Oracle’s database and middleware products in the era of cloud. Oracle already has a distribution agreement with category leader AWS. Oracle’s leaders recognize that the company’s big-game enterprise sales team will not sell cloud services to developers. And so it will rely on Microsoft’s extensive Windows Azure marketing to create an indirect channel for its products. Expect more of these channel deals in the coming months. And draw the obvious conclusion about Oracle’s ambitions as a primary provider of cloud services.
  2. We also believe that Microsoft will allow Oracle to capture the majority of “support” revenues associated with the Java platform on Windows Azure. 
  3. Microsoft likely will pay Oracle an annual license fee for the Java runtimes. Historically, those licenses have generated only small amounts of revenue.
We don’t think these sources will add up to a lot of revenue, but it should be highly profitable revenue and revenue that should grow. 
We say “should grow” because Oracle’s products remain overpriced for most of the developers working in cloud today. Oracle’s partnership with Microsoft opens the possibility for customers to take their on-premises licenses to Windows Azure, but they’re still premium-priced licenses. Our developer survey shows a marked preference among cloud developers for open-source software, for obvious cost reasons. Oracle’s bet is that its customers will stick with its database and middleware as they adopt cloud for more and more workloads. We’ll see. 
In the backs of our minds, a little voice is whispering that there must be more money than we can see in this deal for Oracle. Larry Ellison doesn’t “bet on the come”; he bets on big pools of revenue. We don’t yet see large new pools of revenue in this deal — but we’ll keep looking.
Jeffrey Hammond and I collaborated on this post.