We've now moved beyond the vision and enabling technology discussion with private clouds. We're seeing more and more IT organizations start to put adoption strategies in place.
And, with large numbers usually comes inherent patterns.
In this post, I'd like to talk about the three primary adoption models we're seeing, and some of the context that leads to their choices.
Rewinding ...
A private cloud is a next-generation IT infrastructure environment that IT controls. It may use internal resources, external resources, or any combination.
What makes it a "cloud" is three aspects: (1) it's built differently -- dynamic pools of virtualized resources, (2) it's operated differently -- highly automated low-touch and zero-touch models, and (3) it's consumed differently -- convenient consumption, sometimes pay-as-you-go.
What makes it "private" is that the IT organization is under complete control -- how it's used, how it's protected, how it's managed, how it's secured, etc.
Of all the various enterprise IT cloud models being proposed, I would argue that the industry is quickly gravitating to the one I just described. Virtualization is the specific enabling technology that's making all of this possible.
But How Do You Get There?
Last year we all spent a lot of time in three areas.
First was the private cloud model itself -- what it was, how it compared to other cloud models, and why we thought this was going to be the best approach for most enterprise IT organizations.
Second, we spent a lot of time on enabling technology -- what was there today, what was coming soon -- in all the relevant disciplines. Compute, network and storage. Hypervisors, orchestration and security. Legacy and next-gen applications running in these environments.
Third, we also teed up a number of "softer" discussions: how the operational model had to change, and how that was going to present difficulties for some. How the consumption model had to change as well, and how that was going to be perhaps just as difficult.
We didn't have simple answers -- we wanted to be very honest that we as technologists could only go so far in creating private clouds -- at some point, the organization itself had to adapt and adopt.
So, what are these adoption models that we're starting to see? And why are people making the choices that they are?
Adoption Model #1 -- Incremental Progression
There are many IT organizations that view this evolution as yet another gradual industry transition, very similar as to others we've seen before. UNIX, for example, started at the periphery of IT, and gradually worked its way into the core of IT operations. In many cases, the progression was slow, gradual and thoughtful.
These people tend to have multi-year horizons. Technology decisions come up every day ; they publish a desired end-state, and try to align tactical technology choices around a bigger picture.
Organizational alignment is achieved through working teams making recommendations, followed by slow and gradual organizational alignment, including the sourcing (whether internal or external) of new skills needed for the end-state. Consumption models are done through patient negotiations with business users and finance, and -- more often -- with technology suppliers.
The appealing part of this adoption model is that it's a linear extension of an existing initiative, e.g. virtualization of servers. One can easily see a private cloud model farther down the road of progressively virtualizing more and more of the environment.
Change is incremental and predictable -- the pace is determined by the comfort levels of the IT organization with various new technologies and the newer operational models that they imply. Risks tend to be minimal and relatively easy to mitigate -- simply because you have plenty of time.
And that's the primary downside -- time! These strategies take many years to unfold, and the business may not be satisfied with the pace of change. Additionally, there's a "critical mass" argument as well -- enough of the pieces have to be in place at the same time to really kick-start the model.
But, in all fairness, not every enterprise needs to drive their IT organization to get to the next level, especially in regards to next-gen infrastructure. There are plenty of IT organizations where safe and steady is the order of the day, and I for one will not argue to the contrary.
We, as vendors, need position our offerings in two modes: tactical, immediate benefits to today's problems -- as well as creating a strategic foundation for the future.
Adoption Model #2 -- Accelerated Change
Some people in IT leadership are in a hurry. They want to accelerate change in a big way.
They realize that to gain the visceral impact, several things have to change at the same time, and do so in such a way that organizational thinking starts to orient around the new thing, rather than the legacy.
This is where things like Vblock and Acadia-developed services through partner come into play. In once sense, Vblock is nothing more than a vehicle for accelerated change.
To put this in perspective, let me share with you a scene that's played out for me at least a dozen times so far.
I put Vblocks in front of the IT infrastructure team.
The server team says "well, that's not how we do things". Ditto for the network team, the storage team, etc.
And their boss responsible for all the disciplines says "that's exactly the point!" and presses to move forward.
The IT leader realizes that not a lot of value is created by the server team spending a lot of cycles to decide on the very best server, and the storage team to decide on the very best storage, and the management team to decide on the very best management tools.
Now, the individual technologies in a Vblock can certainly pass muster considered separately -- after all, it's the very best from VMware, Cisco and EMC respectively -- but the point isn't the individual technologies, it's how they come together to deliver a service -- a vending machine for VMs.
Trying to migrate the legacy forward will take too damn long. Let's find a big hunk of "unmet need" within the IT organization, put the new technology in place, put the new operational model in place and the new consumption model in place -- and get on with it!
Sometimes there's a gap of in-house resources (people and money) and that's where Acadia-developed services through partners closes the gap. You can rent people to implement the operational model. You can get a pre-fab consumption model if you want.
Once the environment is up and running, people come to appreciate its benefits -- the KPIs are stellar compared to alternative approaches. Internal resistance lessens. And we have accelerated change.
The pros of this model are simple -- accelerating change to a next-gen model. You're essentially buying time-to-value -- both in the solution itself, but more importantly in accelerating organizational change.
The challenges? You need a decent-sized portion of the estate to get started -- trying to do this with a few dozen VMs won't show the economies of scale -- both resources and operational model. You need enough organization sponsorship to try something that's a clear departure from what's been done in the past. That can be harder than you think.
And you've got to get people weaned away from thinking about individual engines, spark plugs and transmissions, and thinking towards "great car".
Adoption Model #3 -- Work With Compatible Service Providers
There are a very small number of IT organizations that see the end game clearly. They don't think they really need to own a lot of internal IT infrastructure in the long term.
So they've invested in seeking out and meeting with compatible service providers -- much in the way most IT organizations work with technology vendors.
They evaluate the services available. They provide direct feedback as to what they like, and what they don't like. They are very clear as to what they're willing to spend if a vendor meets their requirements.
Now, truth be told, many of these people tell me they don't find a lot to like out there today in the various IaaS, PaaS and SaaS spaces. But they're committed to starting somewhere, and helping the market along using their direct influence.
I'm meeting people who've already moved out various forms of collaboration and productivity applications, as well as certain forms of digital records management. Test and dev is getting increasingly popular, as is serious interest in business analytics.
Everyone of these organizations has some sort of "external cloud governance model" that provides a high-level summary of what's needed from a service provider for different workloads to migrate outside of the data center.
You can see EMC IT's cloud governance example here.
Large Organizations -- A Hybrid Approach
Larger IT organizations are starting to do all three.
They've usually created a long-term architectural and organizational vision, and starting to align tactical efforts along a bigger strategic picture (incremental progression).
They're typically very interested in standing up an internal private cloud around a big unmet need within IT to gain comfort and experience with the technology, operational and consumption models.
And they're aggressively working with compatible service providers to help create new alternatives to simply owning the tech themselves.
EMC's IT organization, for example, is doing all three today. I could name several dozen "name brand" IT shops who are in the same league.
And, the more I look, the more I see of this happening sooner than later ...
Comments