VMware's acquisition of SpringSource is not a match made in heaven. It's going to take an effort by both parties to make this marriage work. Still, it looks like one of the few responses VMware could make to counter Microsoft's dangerous invasion of its turf.These thoughts were prompted by an exchange with Salil Deshpande, a general partner of on the three largest investors in SpringSource, Bay Partners. "I don't think VMware can remain simply a virtualization vendor. Virtualization is just table-stakes, at this point," he wrote in an email response to my query on why VMware was making its $362 million investment.
Deshpande didn't mention Microsoft and he has no special knowledge of VMware's intentions. But with Microsoft offering Hyper-V as a feature of the operating system, what, in the long run, was VMware supposed to do? Stand by and watch its $1.3 billion virtualization empire get commoditized? How could it take advantage of current computing trends and ward off an incipient invasion of its customer base? One answer is to harness virtualization to a larger and higher purpose, one that might be difficult for Microsoft to match. It took its first steps in that direction with vSphere 4, a re-orientation of its product line from managing a local center of virtualization, its Virtual Center, into creating a data center operating system that would manage the majority of resources.
Managing virtual machines themselves is supposedly the big thing right now, but that's something of a mis-statement. Running applications effectively on virtual machines is the big thing, and VMware customers run a heavy percentage of Windows applications on their ESX Server virtual machines.
Microsoft's growth in the data center has been impressive, in part, because of the growing power of the x86 servers, thanks to multi-core processors, and the expanding legions of application programmers familiar with Windows Server.
In looking around for an alternative, it seems logical that VMware would gravitate to Rod Johnson and SpringSource, since Johnson is one of the few people who has understood that Java could be made simpler to use. Too many Java advocates love their complexity. Johnson himself has said he had to get out of the mindset of the Java programmer to see where it was going wrong, how it could be made easier to use.
In the future two new things will be required of applications that will give Java apps a continued, major role in the enterprise: they will need to be broken up into services and spread across many servers. Java applications lend themselves to running in multiple process threads, suitable for multi-core chips and multiple servers, taking advantage of a trend that yields high performance. In addition, the workload will need to be split between 1) the part running in a bare-bones data center that hums along furiously close to 100% capacity much of the time, and 2)the part that gets exported to the external cloud, which picks up the slack.
Java, or a combination of Java and one of the modern dynamic languages, can do this as well. The scripting language communicates upfront with the public cloud and tells it what's needed, setting up the core Java business logic.
SpringSource is one of the few companies that has had an appetite to pull these two attributes together, lightweight Java development and the flexibility and Web-worthiness of the scripting language.
VMware officials, wary of the looming Microsoft presence and eager to find a way to counter it, must have been intrigued by SpringSource's ability to keep grafting new capabilities onto what it had already accomplished.
Maybe VMware paid too much, as a few skeptics have said. But sometimes maintaining your ability to see into the future and navigate its shoals is worth a little extra.
"Cloud computing will be too pervasive and too important for VMware to remain only a virtualization vendor. It needs to do more. It needs to make many moves such as the SpringSource one. I think they will," said Bay Partners' Deshpande.