In the last month, two Platform as a Service (PaaS) vendors launched Infrastructure as a Service (IaaS) offerings designed to attack the near monopoly currently held by Amazon Web Services. Microsoft was the first, announcing the addition of Linux and Windows virtual machines and a number of other supporting offerings to their Microsoft Azure cloud. This past week, Google followed up with their announcement of Google Compute Engine.
The introduction of these two cloud offerings places a magnifying glass on hole in another company's cloud strategy: VMware. Just over a year ago, I wrote a blog entry that spoke about how VMware was the first company that appeared to be actively pursuing what I call a "whole cloud strategy". In other words, they were developing a unified story across SaaS, PaaS, and IaaS and across public and private clouds. Unfortunately, the approach they have taken has been very VMware (taken from the playbook of the old Microsoft) and ignores the reality of world outside of VMware.
Here's my perception of the VMware cloud strategy:
- Encourage the migration of enterprises to a cloud infrastructure friendly to existing VMware environments in the form of vCloud
- Provide a platform in the form of CloudFoundry to give vSphere customers a cloud-scale target for greenfield applications
- Develop a portfolio of SaaS offerings that can support enterprise SaaS needs
Over time, it's become clear that the VMware marketing arm is actively engaged in FUD aimed at discrediting cloud-scale infrastructure solutions (public and private) rather than develop their own cloud-scale infrastructure solution. I am guessing that the reason they have chosen this approach is that the development of such an infrastructure solution significantly reduces the friction of the vSphere solution and enables customers to more easily explore VMware alternatives.
In short, the VMware strategy has been based on two premises:
- The enterprise can be enticed to stay in the vSphere sphere for their existing applications by using vCloud. With vCloud, they can see some cloudy benefits without having to re-architect applications to be "design for failure". This helps blunt the impact of cloud services like Amazon and competing cloud-scale infrastructure offerings like CloudStack, Eucalyptus, Joyent, Nimbula, and OpenStack.
- The enterprise can be convinced to skip IaaS for new applications and go straight into PaaS via CloudFoundry. CloudFoundry on top of vCloud can hide all of the technical debt in the VMware platform and make an application operate at cloud scale.
The theory then goes that there's no reason for a customer to adopt any kind of cloud-scale IaaS (public or private) solution. The Microsoft and Google announcements, however, have exposed the flaw in this thinking.
In short, Microsoft and Google moving into the IaaS space is the clearest signal that Platform as a Service just isn't ready for the big leagues yet. While their respective PaaS offerings have proven popular among developers, the level of adoption of PaaS services is a rounding error in the face of IaaS adoption. The move of Google and Microsoft into the IaaS space may ultimately be a sign that PaaS isn't the grand future of cloud everyone has been predicting, but instead just a component of a cloud infrastructure—perhaps even a niche component. If the world isn't ready for PaaS, then VMware has a huge hole in its cloud strategy because vCloud just won't cut it for cloud-scale infrastructure.
For example, it's also telling that all of the recent cloud IaaS offerings on the market aimed at high-scale operations are not vCloud-based. They run either on an internally-developed IaaS platform or one of the vCloud competitors I mentioned above. Where you see providers who support both vCloud and non-vCloud options, the vCloud option is aimed squarely at the Virtualized Data Center (VDC) v2.0 market, and the non-vCloud option is aimed at the cloud-scale market. If VMware wants to have a whole cloud strategy, they are going to need to develop a story for operating a cloud-scale IaaS and stop publishing FUD around "design for failure".