The cloud discussion has been percolating through IT for about seven years now. It shows every sign of now going to a full boil.
Most every IT leader I meet is now accountable for having an acceptable "cloud strategy" of some sort.
Up to now, I think it's fair to say that most IT leaders have been playing what might be charitably described as an edge game, largely by keeping cloud at the periphery of the IT landscape, and far away from the core.
Time is running out.
Sooner or later, cloud is going to crash into the core of IT.
Just like two planets colliding, the result will look very different for both. And the decisions you make today will greatly affect what the new world looks like when the inevitable happens.
The Many Faces Of Cloud
Fully appreciated, cloud has multiple aspects: a consumption model, an operational model, a service delivery model, even a business model for enterprise IT.
Most importantly, the advent of cloud has fundamentally changed the way we think about things, arguably for the better. Sorry, there's no going back :)
Just as most IT shops couldn't have made it through the last decade without some form of staff augmentation, the exact same could be said for cloud in the next decade.
Going it alone will become singularly unattractive and unpopular.
What's Your Core Cloud Strategy?
It's hard to have a customer meeting without the topic coming up in at least some form. While the responses vary, the pattern remains the same: 95% of "cloud strategies" I see always involve the periphery of enterprise IT, and keep cloud well away from the core.
We've gone to using Office 365. We've built a low-end private cloud to keep people from going to AWS. We're using a handful of SaaS applications. We're experimenting with a few apps with AWS. We're messing around with build-your-own OpenStack. Something along those lines.
All well and good, really. Most often, a measure of benefits are there, and valuable experience is gained. I'm certainly not critical of these efforts.
But, let's be honest, these are tentative first steps in a much longer -- and more important -- journey.
I've now started to follow up the cloud strategy question with a somewhat controversial supplemental: what are your plans for cloud concepts at the core?
You know, the applications and workloads that really matter to the business?
The Pregnant Pause
After a moment or two, the IT leadership folks usually come back with a reasonable set of responses, once you look past all the addressable security and control concerns.
They usually acknowledge their core IT environment isn't at all compatible with the usual popular choices, e.g. AWS et. al.. Lots of coding and conversion to get any sort of reasonable operational compatibility, even for small stuff.
The people who've seriously tried uniformly agree, it's a lot harder than it looks on the powerpoint. Standalone apps, maybe you stand a decent chance. Deeply integrated core enterprise stuff, no way.
This inherent architectural incompatibility is clearly a formidable barrier to adoption at the core.
Coming Back Around
You either believe that IT is evolving to a hybrid cloud model, or you don't. By hybrid, I mean that external resources are a seamless extension of what's running in the data center, and not another bag bolted onto the edge of IT.
If you do have this belief, you quickly realize that both ends have to meet in the middle: your core IT architecture and operations will have to mesh near-perfectly with your intended target public cloud. If not, it's going to be awfully hard to guarantee service levels, drive operational efficiency, etc.
In the face of inherent architectural incompatibility, the entire proposition comes into question.
Unfortunately, most (but not all) of the public cloud options weren't designed with core enterprise IT use cases in mind. There's not a full SaaS/PaaS/IaaS stack. There's not the option to consume the identical offering on-premises or externally.
And you'll see a disturbing lack of what I call "same-same-same": same technology, same architecture, same operations, same support, etc. Without this, attempting to deeply integrate cloud into the core becomes a morass of complexity and frustration.
If hybrid cloud is the new architecture for enterprise IT, these desires for a complete stack, flexible consumption options and inherent "same-ness" are hard to argue with.
Depositioning False Prophets
I look at many of the claims other vendors are making, and I shake my head.
Amazon's pitch is that you should put 100% of your workloads in their cloud. Goodbye, on-premises.
That's a fundamental non-starter on just so many levels, not the least of which is that you'll be essentially recreating your core IT environment from scratch. Oh yes, and you'll have to figure out what you want to do for integrated PaaS and SaaS.
VMware's pitch is that, once you virtualize with their hypervisor, you can move to any cloud you want. Assuming, that is, at some point in the future VMware can offer you a credible cloud.
Theoretically reasonable for IaaS, but -- again -- what about integrated PaaS and SaaS? And would you be comfortable with VMware's hardware-agnostic support model?
To their credit, Microsoft has a better understanding of what portions of the enterprise might need, but still serious gaps through the lens of end-to-end critical IT applications.
HP, IBM, Dell/EMC, Cisco, Google, et. al. non-starters in this particular game.
One interesting one you hear occasionally is IT's desire to be a "cloud broker". Much like people try to get the best deal for an airplane seat or hotel room (essentially commodities), IT can "play the market" by freely moving workloads between wildly incompatible clouds with no concern for data movement, operational compatibility, etc.
Maybe that's a theoretical possibility in the distant future for non-critical workloads, but I have a hard time seeing this be a reality for core enterprise IT in my lifetime. Who wouldn't want a Star Trek transporter?
Which leaves us with Oracle.
The basics?
Oracle Cloud offers full and integrated IaaS/PaaS/SaaS.
Oracle owns their entire technology stack: from app to chip. They can deliver the exact same experience on-premises, in their cloud, or any combination -- and -- by definition -- working seamlessly together.
Oracle's mission-critical IT credentials are impeccable. Not to mention, Oracle is uniquely able to support the complete stack at both ends of the wire.
If we're having an adult conversation about bringing hybrid cloud concepts to core enterprise IT, it's a completely unfair comparison with other industry pitches out there.
Preparing For The Inevitable
Vendor olympics aside, there are some inarguable truths to be grappled with in the world of enterprise IT.
First, sooner or later it will be demanded that cloud concepts find their way into the core of enterprise IT.
Simply doing a bit of cloud embroidery at the edges won't pass the sniff test.
Second, the architectural decisions you make today will dictate whether that's an easy thing to do, or a near-impossible thing to do.
I would argue that any vendor who can't show you a viable hybrid cloud option that mirrors the core of IT should be considered strictly tactical, and not at all strategic.
And for those of you quietly hoping something better will come along, let me remind you that hope is a terrible strategy.
---------------------------
Like this post? Why not subscribe via email?
Comments