All cloud is hybrid cloud, or so it seems. Is there a single organization that has only one cloud? Hybrid has become a dominant model due to (among other reasons) its flexibility and scalability. In 2018, businesses combine a handful or more of cloud deployments into a single deployment to create a resource that dwarfs the power of a legacy data center.
It's clear, then, that hybrid cloud – and its close colleague, multicloud – has a bright future as a dominant model for enterprise IT.
Yet as commonplace as hybrid cloud is, confusion abounds about this complex cloud computing model. Even the very term 'hybrid cloud' is defined in myriad various ways. In some cases, vendors see hybrid cloud in a far different light than how it's viewed by customers.
To provide insight, I spoke with Dave Bartoletti and Joshua McKenty, both of whom are widely regarded as a top experts in cloud computing.
The topics we covered included:
- How customers and vendors view hybrid cloud differently.
- What customers are seeking in their hybrid deployments, and the degree to which they attain this goal.
- Where in the stack that businesses are most concerned about hybrid.
- The future of hybrid cloud.
(See the video and a transcript of highlight below.)
VP and principal analyst, Forrester
VP, Systems Advisory Group, Pivotal
Edited transcript of highlights:
How Do Customers View Hybrid Cloud?
D. Bartoletti: I answer this question all of the time. I actually have sort of a radical opinion about hybrid cloud from the beginning: There’s no such thing.
Hybrid cloud is not a thing – it’s a situation you find yourself in, when you want to use multiple cloud platforms and you want them to be relatively consistent. Most of our clients don’t spend any time trying to refine definitions between hybrid cloud versus multicloud versus the range of cloud problems that I have.
In fact, at Forrester, we’ve never had an inquiry where someone has called to ask “where do I buy a hybrid cloud, or source a hybrid cloud.”
J. McKenty: Yeah, I think Dave is spot on. Hybrid cloud is a great term for the problem that our customers have, as opposed to a strategy they hold.
And that problem is, whether they’re dealing with multiple cloud providers because they need a geographic reach, or they want to take advantage of specific capabilities, or they’re just trying to play them off of each other from a budget standpoint.
Even if they go ‘all in’ on one public cloud provider, they’re always going to have on-prem. So, they’re always hybrid no matter what. If they don’t think of it as that being the reality, they end up backed into that reality later when they realize, ‘oh, there are things we cannot put in a public cloud environment, or that it’s not feasible to move’.
Is the ‘Single Pane of Glass’ a Key Part of Hybrid Management?
J. McKenty: No. It actually ends up having very little utility. The Pivotal approach, especially because of our history around digital transformation and develop communities, is always to start from the business. From the application. From the workload. And say, what are you trying to build and who is it for? And then say, well, if it’s part of your e-commerce system then it needs to run as close to those customers as possible, so you need a distributed deployment model. Okay, well that kind of dictates where it’s going to be built, and then it also has to transact with your backend databases because you’re taking money and you have inventory, so it has to have connectivity back to where these systems live.
We work backwards from the customer experience. Now, if you have a business that has both factories and e-commerce – let’s say you’re a retailer of fashion. Treating those two things behind one pane of glass is absolutely valueless, frankly. It just does not deliver the value you are trying to get.
D. Bartoletti: It’s also an analyst drinking game, by the way. When we’re on briefings, ‘single pane of glass’… we all have to take a shot. There never has been one and there never will be one. People have wasted a lot of money on that.
Again, I think it’s a symptom of how we build software differently and run systems differently in the cloud era. In previous eras, the idea was I picked a monitoring platform, and then I bought adaptors for every new thing I bought and I plugged them into that monitoring platform, because there was one team that was just supposed to do monitoring.
And, as we break down the silos between traditional IT disciplines and they come together, now some developers are doing monitoring. Developers care what monitoring is chosen because they’re going to take advantage of it, too. It’s not just someone in operations that’s looking at monitoring.
Looking Ahead in Hybrid Cloud
D. Bartoletti: Speed is going faster in one direction. Agility is changing directions quickly, and that is still really hard. So, you might be able to get a new mobile banking app out of the door really fast, but can you throw it away if it doesn’t work? That’s really hard, and enterprises are not built to fail and throw things away.
When we tell folks on inquiry and they say we need to pick a platform for the next five years, I say no you don’t, you need to pick a platform for the next year and get your first app on it now, and pick one cloud platform to deploy it on. Pick one, please.
[Companies say] first we’re going to do contract negotiations with all three of the major public cloud providers because we don’t want to be beholden to any one cloud provider at this point. You know, lock in. Boy, are we scared of lock in. And we try to explain to them that here’s what’s different in this market. In previous software generations, companies that had an oligopoly or four or five vendors dominated a market, their innovation pace slowed down and their prices slowly crept up. Neither of those things is happening in the cloud. The prices are not going up and innovation is not slowing down. AWS will die if they don’t stay on a faster innovation pace than Google and Azure.
So, there are so many new things coming. The way I would think of hybrid is, I want to make sure that my approach is hybrid in that tomorrow when I look for a new solution I look for vendors that I’m not working with today.
J. McKenty: [Looking to the future of hybrid:] One, it’s dominated by networking, which means its dominated by distributed systems. And the second is that it’s dominated by people.
And this is what’s funny. People change much slower than systems. And it’s generational. The generations have pretty predictable characteristics. So, even though our market pace has hit this point where its never going back to normal, and the pace of change is accelerating, the mental model shift and the people involved is not.
So, what that means for our customer base, for the Fortune 500, is that most of what they grapple with over the next five years is all about people and communication. Most of what my team does, the system advisor group, is just about the people side of techno-social systems. Like, I can walk you through a diagram of how Cloud Foundry works, but what you need to understand is which teams use which interfaces and how they coordinate between each other. And that shift, what we call the shift to ‘product-centric organization’, the shift to ‘responsible autonomy’, all of the future of hybrid cloud is dominated by which teams are allowed to make those choices.