Last fall I wrote a piece for VentureBeat on the contrasts between two key cloud operating models. My intention was simply to contrast cloud lock-in with cloud agility, and what that contrast means for IT teams, especially those who might be unsettled by the cloud. I also wrote at Seeking Alpha that the cloud agility market could be $60B. Yet today even the “cloud as taxi” metaphor is broken, when one considers the two distinct and dominant flavors of lock-in, intentional and incidental.
Cloud lock-in is essentially the extension of the tired hardware-bound lock-in into the cloud. The cloud service provider “owns” the customer in the same way that the hardware vendor used to exert control over customer environments. Training, expertise and complexity created effective barriers to competitors and enterprise development. Innovation was limited to the confines of vendor capabilities. Intentional lock-in is the creation of specialized services and APIs that limit agility, while unintentional lock-in is the high cost of deploying an app environment into any cloud.
Today OpenStack has particularly high entrance costs. While that is changing rapidly thanks to development efforts by many companies, earlier versions, especially those which have been modified, have unintentional lock-in. OpenStack certainly offers opportunities for enterprises with large IT budgets and dedicated developer teams; yet the price of entry and exit is not a strength.
Cloud agility is an emerging operating model; the cloud supplements existing assets in the data center and other clouds to enable higher levels of agility and synergy. Agility and synergy then enable higher performance through operating efficiency. Yet the challenge for achieving cloud agility is clearly in getting there and few service providers will offer ease of mobility between their environments and competitors. Today the dominant model is lock-in, whether intentional or unintentional.
I watched Bob Green (CTO at Dizzion) speak at a TechTarget Modern Infrastructure event in Boston last week and one of his slides showed a breakdown in operating costs in traditional versus private and public cloud environments for a particular application. There was a spike in the front of the public cloud (the cloud migration or integration expense) that he cited as an area of disproportionate influence on ROI.
That is why I’m pretty excited about what CloudVelocity (my employer) recently accomplished with an enterprise Oracle ecommerce stack (about 6TB in size), which was deployed into an AWS hybrid cloud in 7 days. After the proof of concept was deployed and tested it was left on per the request of the (Fortune 1000) company. The new environment turned out to be faster and easier for Dev/Test than the existing private cloud Dev/Test lab, without re-architecting or significantly modifying the app and stack. The stark contrast seen by company officials: a slow, vendor-driven environment with fixed costs versus an agile, on-demand environment with costs driven by usage.
It is very common to hear the “cloud as taxi” metaphor when talking about comparative costs, but that undersells the agility and synergy aspect, which might prove to be even more important. Think of the cab being available faster than your own car versus the car in the garage.
That takes us to another problem with the cloud as taxi metaphor: the upfront costs of getting into the cab today with existing cloud management tools and processes. The slide Green presented showed the spike in initial costs, perhaps the equivalent of paying $150 in advance for the rights to take a $50 cab ride. So what if another cloud provider charges only $40 or even $20 for the same ride? The $150 required in advance erodes the ROI for any change.
This is part of the OpenStack dilemma as well. As OpenStack matures and offers APIs similar to other offerings it reduces the upfront cost of entry. Fewer expenses and developers upfront combined with a more robust offering could tip the scales away from incumbent cloud providers. But this clearly has not yet happened and incumbents are investing heavily in new infrastructure (and innovation) at a hearty $50B annual clip.
That is why I think that hybrid cloud automation is perhaps the most strategic opportunity for growth and profitability. Make it easier to get multi-tier production apps into the cloud (without lock-in, including pre-cloud virtualization) and you accelerate cloud adoption. That makes the monetization of agility (versus control) perhaps the most significant opportunity in the cloud space; more so than slugging it out as an IaaS provider and competing with specialty APIs. Yet that would require a revolutionary shift in thinking, from control to agility.
See Will VMware or Microsoft Cash in…? for how I determined that the hybrid cloud automation (or agility) market could be $60B.