With SpringSource buy, VMware constructs cloud platform
- 18 August, 2009 03:48
The talk of CloudWorld this week was VMware's acquisition of SpringSource. The top-of-mind chatter focused on the price: US$400 million plus, a very large sum for a company doing perhaps US$25 million in revenues. Certainly there was a good bit of envy in this type of conversation. And, of course, the fact that SpringSource is an open source company further makes the number even more eye-watering.
However, looking beyond the envy and valuation puzzlement, what is the implication and meaning of the acquisition?
I must say that, to me, it looks brilliant--and it takes a lot for me to make that assessment. Having suffered a few disastrous acquisitions at companies I worked for, and observed at remove many, many more, I've come to view these events pretty skeptically. They're always announced with fanfares and group hugs among the management teams, and 18 months later, after the acquiree's top team has left and the struggle to integrate the products into the acquirer's portfolio seems doomed, everyone shakes their head and wonders "what did company X see in company Y?"
This acquisition seems different to me, and it holds the potential to really give VMware a premier position in marrying enterprises and cloud computing. Here are the reasons:
It provides a complete cloud offering: vSphere can offer a single product that offers both IaaS and PaaS, configurable at the application level; in other words, within a single cloud, some applications could operate in an IaaS environment, while others operate in a PaaS environment. Unlike other cloud offerings, which are single option -- either all applications that run in that cloud have to be based on IaaS or PaaS, the VMware offering can provide the opportunity for applications to assume an IaaS environment or a PaaS environment; moreover, this should be granular and configurable per application. So those applications that require total environment flexibility can use vSphere to provide empty VMs into which a complete stack--OS, middleware, application--can be poured. Those applications that can leverage a Java framework can achieve the higher productivity (and perhaps higher quality) of using a VM which already contains an OS and programming framework. The fact that a single offering can support both IaaS and PaaS, granular at the application level, is a real benefit.
It offers support for *the* enterprise language: Java is the default enterprise language for complex, highly-scalable applications. Offering a Java framework in a cloud environment is going to be very, very attractive to enterprises that want to implement cloud computing and want high developer productivity. I'm a bit dubious about some of the comment on this acquisition that explains the acquisition as following the Microsoft "get developer support and IT will follow" playbook. Commenters who support this analysis point out the fact that VMware's CEO comes from Microsoft, so, supposedly, he's just executing the same game plan he knows.
I think this misses the point. While that strategy may explain Microsoft's success, Java is in no need to gain acceptance at this point. And implementing cloud computing is not going to be driven on an app-by-app basis--it's a higher-level decision and one rooted in operations. The question is more likely going to be: "we've got a ton of Java apps and would like good support for them in our cloud implementation--which is the cloud best suited to support the Java apps we've already got (many of which are built on Spring) and that we'll do in the future? And, by the way, it would be great if that cloud choice also supported non-Java apps as well." From this perspective, the combined vSphere/SpringSource offering will be appealing.
It delivers a great offering for vSphere service providers: A lot of the hosters and outsourcers that are getting into the cloud game will love this offering. They're trying to move into cloud computing, driven by the realization that hosting machines or offering virtual machines is a less-than-desirable strategy in a world that wants the agility and scalability of cloud computing. By the same token, many of these providers, companies like Verizon Business, AT&T, etc., have been wracking their brains about how to compete with the low-cost Amazon offering. Offering a more complete and flexible cloud platform that supports both IaaS and PaaS will give them some competitive ammunition against commodification. Of course, they'll still struggle with the question of how they'll differentiate among themselves when they all offer vSphere, but that's tomorrow's struggle!
It supports the shift from licensing to services and monetizes open source the right way: I've always felt that cloud computing was going to be the way many organizations consume open source, given the seeming reluctance on the part of many organizations to embrace it in a packaged form. Moreover, most open source companies have struggled to achieve revenue levels commensurate with their product adoption levels--after all, why should people pay for something they can get for free? That goes away when what's delivered is not bits, but service: what you're buying is not software, but execution, so the licensing conditions of the bits is irrelevant. Likewise, I've felt that a transition from proprietary licenses to value-based service is the future of software companies. I know that VMware is a true-blue license company and that seems to negate the argument. However, it may be the last great packaged software success story, since most software today seems to be offered in a bits-and-service bundle. In any case, much of how users will consume sPhere will be via service provides who will transmute VMware licensing fees to usage-based cloud services.
Of course, no guarantee exists that the reality of the VMware/SpringSource mashup will prove successful. Many mergers offer brilliant opportunities and then founder on poor execution, organizational rivalries, shifting competitive environments, and plain bad luck. Here are some things that could challenge the happy consummation of the acquisition:
It could founder on the reality of enterprise Java applications: A common framework is great, but the reality of enterprise Java applications is that they are a mish-mosh of product versions. Every application requires a specific stack of OS and middleware, and even applications that have a common stack often require different versions of the individual components within the common stack. A vSphere/Spring stack that offers a consistent framework is great from an operations efficiency perspective, but may be unworkable in the face of individual applications being written to varying version and patch levels. Of course, this might not be as critical for home-grown applications that can be written to a consistent stack, but commercial applications define configurations based on the vendor's decisions, not the deployment operating environment's configuration. To be fair, this is a challenge for any PaaS environment, and I expect that Azure will face it as well. Nevertheless, corporate IT operations environments are messy places filled with varying application stacks, and a common framework may be a theoretical ideal that poorly maps to the pragmatic deployment reality.
It may be too complex, even though architecturally elegant: The jury is still out on internal clouds and this offering may prove too difficult for enterprises to implement. Getting a highly-automated virtualized environment to work properly is formidable, given the complexity of knitting together all the piece parts of the computing infrastructure. Making it more challenging is the cost pressure that most IT organizations are under, making investment in operations kit and personnel very difficult. Trying to overlay a complex cloud environment on top of complicated infrastructure may be too daunting for most IT organizations in a time of budget squeezes. Of course, this might mean that IT orgs will turn to service providers which, as noted, may be eager offerors. This question cannot be answered today, but gives thought for the future prospects of the offering.
It might be too late: Integrated product offerings are notorious for delays and an extended delivery timeframe for the IaaS/PaaS product might render its relevance moot. One need only look at the Oracle Fusion debacle to see how imposing integrating disparate products is. It was three or four years ago that, one year post-PeopleSoft acquisition, Oracle announced with great fanfare (I seem to remember the Oracle President announcing with pride "It's great when a plan comes together") that is was half-way through integrating all of the product portfolio with Fusion. Even allowing for Oracle math, this illustrates how hard it can be to get different software products to work together. While the vSphere/SpringSource integration might be very attractive, the pace of cloud computing adoption might mean the product misses a window of opportunity. An attractive product offering can freeze the market for a while, but customers will not wait forever before moving forward--especially on something as important as cloud computing. Again, this question cannot be answered today. Unlike the previous issue, however, this one would affect both internal cloud implementers as well as service providers.
I hate "on the one hand, on the other hand analyses" and wish I could provide a more definitive opinion on how this will work out. My sense is that it's an execution question, and it's impossible to predict how the two companies will come together and execute. The attractiveness of the potential product is obvious, and I think the strategy behind it is brilliant, not a word I overuse and definitely not one I apply to most mergers. There is such innovation and opportunity in cloud computing that this product could be a big winner. We'll just have to wait and see how the story turns out.
Bernard Golden is CEO of consulting firm HyperStratus, which specializes in virtualization, cloud computing and related issues. He is also the author of "Virtualization for Dummies," the best-selling book on virtualization to date.
Do you Tweet? Follow everything from CIO.com on Twitter @CIOonline.