The case of the missing hockey stick
There are days when I wonder whether the old joke about 201x being the Year of VDI will be reused to talk about mass migration of legacy enterprise applications.
One of the great advantages of Virtual Clarity not being a born in the cloud company (apart from not having the word 'cloud' in our name and hence immediately differentiating ourselves from the herd) is that we have an institutional memory that stretches beyond the last five minutes. That memory proves useful when called upon to wonder why it is that we have yet to see the great and much heralded mass migration of enterprise applications to public cloud.
Virtual Clarity was born in 2008. We just enjoyed our ten year anniversary in fact. In 2008, VMware ESX had a bunch of years and revisions under its belt and was finally, unquestionably, enterprise ready. More importantly, migration of workloads and applications from horrifyingly under-utilised physical machines to new virtual infrastructures offered an iron clad business case. There were good reasons to transform applications during migration. There were even better reasons to transform the IT operating model to change the way IT was delivered. Critically, for VMware and the CIOs whose cost base they improved so dramatically, it was sufficient to just move workloads from physical tin to virtual tin to reap huge rewards. Many, perhaps most, of the organisations that jumped boats first into virtualisation were able not only to turn off fleets of physical machines but whole data centres. The business case for virtualisation was so strong that most organisations didn’t bother creating business cases at all. The result was the hockey stick like growth in virtualisation, the eclipse of EMC by its daughter organisation and, today, a forlorn expectation that we would see the same wild growth in migration to public cloud that we saw for virtualisation a decade ago.
However, it simply isn’t happening. The explosive growth of AWS and Azure, and somewhat less explosive growth of GCP, has emphatically not been because of the bulk migration of legacy enterprise applications. This is despite the fact that, in many ways, the business case for the use of public cloud is even stronger than it was for virtualisation. The strength of that business case relies on the new capabilities offered by public cloud to a far greater extent than the business case for virtualisation did. When a typical organisation picks up an application from their data centre and migrates it as-is to public cloud, the immediate saving is likely to be zero and may even be significantly negative. The pure-play lift and shift migration that offered such great savings in virtualisation often lacks a real business case in public cloud.
There are a few reasons for this. In most cases, the applications that are moving are already running on efficient, virtualised infrastructure whose capacity cannot usually be released until all applications are evacuated from a cluster. During that time, the customer is now paying twice – once for their on-premises infrastructure and once again in the cloud. To a far greater extent than was true for the virtualisation revolution, organisations need to pay attention to their operating model. At Virtual Clarity, we have found that the key to a positive public cloud business case is to consider infrastructure, application and operating model transformation in tandem. Doing so can deliver business cases that unlock the value of public cloud.
Infrastructure transformation uses the public cloud provider’s capabilities to create a modern platform for enterprise IT. Minimising abstraction of the cloud provider’s services offers better value than building layer on layer of tools and code to make services portable. Application transformation is a continuum that starts with optimisation and ends with rebuilding applications using cloud native models. At VC we find that considering each application in the context of the provider’s services and the underlying enterprise IT cost model is the right way to make decisions about how each application should be optimised or transformed. The future state operating model both drives those decisions and is itself influenced by them. The degree to which database platforms are optimised to use the provider’s managed services, for example, helps determine how databases will be managed in the future. Looking at these dimensions simultaneously can seem to be a task of daunting complexity, but, at VC, we have built a set of tools (Decision Support Engine and Migration Director) and methodologies (ClearFoundation, ClearInsight and our value management approach) that we use to cut through the complexity and create clarity. We call our approach precision guided transformation.
Considering the enterprise operating model as a whole is the way to create real value from a move to public cloud. A transformed operating model eliminates whole areas of unproductive work; it manages technical debt rather than simply kicking the can down the road; and it ensures that responsibility and accountability for the future of technology in a firm is shared by stakeholders across the organisation.
The hockey stick graph of public cloud migration is missing - but we know why it hasn’t turned up yet and what needs to happen to make it appear.