The SpringForce/VMForce announcement April 29, 2010
Posted by Sacha in CloudBees, English, IT.
VMWare and Salesforce.com common announcement this week, VMForce, was really interesting.
First of all, it is yet another drum beat on the cloud march. The virtualization king (Market cap: 25B) and the CRM PaaS queen (Market cap: 11B) got together in what appeared to be a relatively ego-less offering: VMForce. One camp has the developer community; the other has the XaaS credibility (and is the biggest PaaS provider by far). This is not a lightweight marketing partnership: this is a pretty deep partnership under a unified brand umbrella.
This reinforces a message that will become a truism in the next few years: the cloud is new platform.
What is this really about? This announcement is about the PaaS leader, salesforce.com (CRM), offering itself a new programming model on top of its already successful platform, in order to extend its developer reach. CRM already has a strong set of services at its foundation: a relational-like multi-tenant database, pre-defined business schemas, REST APIs, etc. But all of those were only available from CRM’s own proprietary programming model. Thanks to VMForce, Spring users will have an easier path to Force.com.
But don’t be tricked, this doesn’t make VMForce the new “all-purpose”, generic PaaS offering: this remains a vertical-focused PaaS offering, aimed at Salesforce.com users. VMForce offers the advantages of its disadvantages: if you are a Salesforce.com user, this is going to boost your choices and flexibility by being able to leverage a substantial chunk of the Java ecosystem. Yet, if you are not a salesforce.com user (or if you are developing a non-CRM related application), VMForce represents a major lock-in into a platform that makes very little sense for your application.
So, who is the winner of that partnership, VMW or CRM?
The obvious beneficiary is Salesforce.com: it is already the PaaS market leader and through VMForce it keeps its lock-in while extending its developer reach. Brilliant. But does it necessarily mean that VMWare doesn’t benefit from that deal? While it is certainly a good financial deal for VMWare, it is almost secondary. The second beneficiary of this partnership is not the good old virtualization-focused VMWare, but VMWare’s middleware team: SpringSource. Acquisitions are never easy to pull correctly, especially when an operations-focused company acquires a development-focused one. Consequently, since SS’s acquisition last year, the middleware team has pretty much travelled in a black hole. Yet, VMForce (which should really be called SpringForce) will help SpringSource re-broadcast their message: they are after the cloud – and this is really what triggered their acquisition in the first place. So this announcement will definitively help restate their objective to the market.
But does it necessarily help SpringSource’s independent agenda as a PaaS provider? I am not at all convinced that it is the case. Lots of the PaaS’ hard problems have been solved in VMForce by simply leveraging Salesforce.com’s existing services. So, not only VMForce will not help SS’s independent agenda, it might actually represent a marketing/positioning challenge down the road when they’ll have to differentiate both offering.
Concerning the timing of the announcement, I was very surprised. It came out of nowhere, wasn’t announced as part of any cloud or developers conference, and the offering is not yet available, not even as a development offering. A beta developer release should be made available in the second half of 2010 with no known date for the GA production release. So, why so much hurry in the announcement? Both companies are known to be more reliable than MSFT’s vaporware PR.
Another element that hasn’t been announced yet is the pricing. This is probably fine for such an early announcement, but not even the pricing SCHEME was announced. And I bet there is a good reason for that: Salesforce.com is known to build by user (which has its own set of consequences and constraints for a PaaS) and VMWare is known to build by the deployment unit (CPU, core, VM, or whatever makes more sense). It will be interesting to see how they can resolve that pretty significant “philosophical gap”.
In any case, what I’ll remember from this partnership is how smart that was from Salesforce.com to make an ego-less move away from its own proprietary offering in order to extend its developer reach, and all this under a 3rd party branded offering, with limited impact on its own engineering roadmap. And congratulations to the SpringSource team for striking a partnership which impacts their mothership as a whole, this is probably their best way to escape from the acquisition black hole.