Let’s face it. Sometimes being an “enabler,” is admirable. However, if you’ve seen an episode of Intervention lately, being an enabler is not always a good thing. VMware’s IaaS strategy was to enable its partners to offer vCloud services and give it’s customers near unlimited (>9,500 partners) choice of cloud providers. There was a big issue with this strategy – it assumed that VMware’s cloud partners would be A-OK with allowing customers to come and go. At the end of the day, that didn’t meet VMware’s provider partners business model. No one wants to race to the bottom of a commodity market and providers rightfully should be concerned with their ability to differentiate with competitors and show value while sharing a common VMware stack.
Today’s news shouldn’t come as too much of a surprise. Nearly two years ago I blogged that this day would eventually come. The market would force VMware to be a provider, and it has.
Forget about the talk of “open.” At the end of the day, every vendor and provider is in the business of doing whatever possible to lock customers in and make it tough for them to leave. Providers have always wanted high degrees of extensibility so that they can add value to a cloud offering and in the end offer enough customized services to make it tough for customers to leave.
If we look at today’s IaaS announcement, VMware is trying to have greater control of the “choice” it’s customers get. Choice will mean a VMware-hosted offering that in theory will make it easy for customers to move VMware-based workloads in and out of the pubic cloud. The aim is an “inside-out” approach where workloads between a private data center and a public cloud operate seamlessly. The trick here, however, is how important mobility and choice will be to customers. Workloads that go straight to cloud and have few traditional enterprise management needs can go to any cloud. Front end web servers are a great example – static data, built to horizontally scale, and no backup requirements.
VMware’s challenge going forward will be to differentiate. If VMware is the “enterprise alternative” to Amazon, it better launch it’s IaaS solution with enterprise features (AWS isn’t perfect but it has tons of features that large enterprises are now taking for granted). Redundant data centers, enterprise storage, networking, backup, and security are a must. In addition, it must offer serious tools for developers; the time for VMware to show the results of its investment in Puppet Labs should be when the public IaaS offering launches. Otherwise, Amazon and other providers will continue to win on features and the ease of experience that developers have on its platform. Granted, this can’t all happen overnight, but VMware needs to show value quickly in order to gain momentum.
VMware also needs to make customers understand that the VM is the easy part. Management has always been the challenge in hybrid cloud models – most organizations running hybrid clouds have at least two management silos – one for public cloud assets and one for private cloud assets. Failing workloads over to different infrastructure and different hypervisors isn’t a challenge because of converting a VM. The challenge exists because the operational software stack deployed to the VM (e.g., backup, security, performance management) may have hooks into a particular hypervisor’s APIs. So moving a workload often can entail considerable QA work to ensure that the production workload runs and is managed properly in the new environment. This is an opportunity where VMware can leverage its management assets both inside the data center and in the public cloud to allow customers to redeploy workloads and not have to worry about the infrastructure or management stack. That can significantly reduce complexity and opex overhead for organizations looking to operate seamlessly across both public and private clouds.
That said, another bottleneck to VM mobility in the cloud is software licensing. VMware can’t control the licenses of the software that runs in its VMs, but VMware must make it easier for customers to license VMware management software in a hybrid cloud environment and allow licenses to move with VMs between cloud environments.
Finally, let’s not forget that VMware isn’t blazing new ground by being a provider and enabler. Microsoft is taking the same approach with Azure. In the end, there should be enough room for partner’s to differentiate by industry vertical or geography, for example.
In the end, becoming a cloud IaaS provider is a move that VMware had to make. VMware was losing cloud mindshare to its competitors and its only choice to keep up with their rapid innovation pace was to be a provider itself. Entering the market late as an IaaS provider places VMware in the unenviable place of playing catch-up. VMware’s strength is its enterprise data center dominance and it’s ability to make a compelling case to its clients for hybrid cloud will undoubtedly determine its success. Telling that story is about way more than the technology. Businesses don’t care about technology details – they want agility and value. VMware must carefully craft a message that speaks to the business needs while convincing customers that a proprietary VMware stack is in their best interest. Many clients I speak with are trying to be more standardized and service provider-like in how they deliver services. This translates to fewer vendors in the data center which leads to lower opex costs and often an easier ability to automate IT processes. There is certainly an opportunity, but VMware success is far from guaranteed.
What do you think about the news today? Is it too little too late or will you give the VMware IaaS offering a serious look?