On Tuesday VMware and Salesforce.com announced their joint venture - vmforce.com. Both vendors offered informative blog posts about the announcement, and I recommend reading the following:
- VMforce and VMware’s “Open PaaS” Strategy (Steve Herrod)
- VMForce: Why? What? How? (Anshu Sharma)
- VMforce Provides Spring Cloud Platform (Rod Johnson)
These posts are also worth reading:
- Analyzing the VMforce announcement and PaaS portability challenges and the VMforce example (William Vambenepe)
- Duncan Epping’s wrap up post
The move makes sense for both vendors. VMware understands that existing as a virtual infrastructure platform for Windows applications – where Microsoft is a direct competitor – is a losing battle. Many vendors have tried and failed in their efforts to take down Microsoft. In the end, the problem for Microsoft’s competitors was always the same – integration across the Microsoft stack would always win out. Paul Maritz realizes this and understands that without challenging Microsoft’s dominance in the application space, VMware will eventually succumb to Microsoft as the preferred virtual infrastructure platform for Windows applications. Partnering with providers such as Salesforce.com is a logical move for VMware. The more VMware can help Salesforce.com and Force.com pick up customers, the more they may see Microsoft’s dominance in the application space erode. This isn’t a strategy that plays out overnight, but over the next 10+ years. Along the way, I expect VMware to partner with key Microsoft rivals such as Google. Such a move is too sensible for both companies not to happen at some point.
For Salesforce.com, the VMware partnership and integration with VMware’s SpingSource Spring Framework offers the potential to bring in a large number of new customers. In addition, the combined VMforce.com solution utilizes VMware vSphere backend infrastructure. As standard cloud API and data models emerge over the coming years, close alignment with VMware should remove concerns of VMforce.com being perceived as a niche solution, and will be one that leverages known and trusted back end virtual infrastructure. This is important when you consider factors such as regulatory compliance. Large enterprises turning to the cloud are wary of lock-in, and the more providers can offer secure solutions that are compatible across numerous cloud solutions, the more likely enterprises are to invest.
While I like the move, there are still many unanswered questions. For starters, making it easy for Java developers to run complex application stacks in the cloud is a good start, but what about the .NET developer ecosystem? If VMware and its competitors want to take market share from Microsoft, they’re going to have to entice .NET developers to come their way. I have said for some time that I see Azure evolving to be both a cloud and on-premise solution. So in time Microsoft will be offering even more compelling reasons to to come their way or stay with them, and Microsoft’s competitors will need a very good answer to counter. If VMware succeeds with “Open PaaS",” they may have the answer – a Java-based PaaS platform with choice on par with VMware’s vCloud ecosystem (1,000+ providers).
It’s also important to note that VMforce.com is not a pure infrastructure-as-a-service (IaaS) play. You can’t package any app as a VM and deploy it to VMforce.com. Instead, the solution is positioned to support Java apps, a Force.com backend database, and leverage VMware vSphere plumbing. As a joint venture, VMware couldn’t get into IaaS for the simple reason that it would position VMware as a competitor with the providers in its vCloud ecosystem.
Finally, if you go back to Steve Herrod’s original post on the VMforce.com announcement, I couldn’t help but focus on the fact that “Open PaaS” was placed in quotes. In the past, I have asked vendors if some of their “features” should be in quotes, insinuating that a particular feature was a bit misleading. If you’ve been in IT long enough, you already know that “open” is a relative word, with openness subject to the number of supported vendor or provider alternatives out there. In the case of VMforce, “open” does not have to imply open source, but rather a defacto standard for PaaS with a large choice of provider solutions.
The following statement from Steve Herrod’s post alluded to VMware’s plans for Open PaaS.
One thing in particular mention here should strike you … we will wholeheartedly enable deployment of these cloud portable applications to clouds that are not based on our underlying vSphere virtualization technology. This support is a key aspect of openness and will enable a broader and more competitive ecosystem of compatible Spring PaaS offerings. And this in turn will be the reason why developers will bet on Spring-based applications for maximum flexibility. Stay tuned as you’ll see many more announcements around this very soon.
Steve is implying that apps deployed to VMforce.com can be moved to non-VMware infrastructure without a Force.com backend, which I mentioned earlier when discussing Open PaaS. If VMware succeeds with VMforce and gets numerous other providers to support Open PaaS along the way, they will have a serious alternative to Azure, and one that is devoid of provider lock-in.
As James McBride once said, the details always tell the story. Right now, we have a story. Whether or not the story will be remembered depends on the details. Give us standard APIs, application models, metadata sets, security frameworks, and modernized service-level definitions that apply to emerging cloud-based architectures, and we have a very compelling story. So far, the book on VMforce.com has a nice cover, but the meaningful content remains a mystery.