The Era of the Cloud has made many pie-in-the-sky promises about more consistent, efficient, and resilient IT infrastructures, and has done an admirable job of meeting much of that potential.

But there’s still a gap in what businesses want and need from a hybrid/multi-cloud infrastructure, and the industry’s capabilities to meet those needs. 

ONUG’s own Nick Lippis sat down with Adam Raymer (Head of Public Cloud Platform Architecture, JP Morgan Chase & Co.) and Shane Stakem (Head of Public Cloud Center of Excellence, JPMorgan Chase & Co.) to answer one deceptively-simple question: what in the world do we do with hybrid/multi-cloud IT?

Why Go Hybrid or Multi-Cloud At All? What’s the Point?

It would be simpler to have every IT component under one roof, everyone thinks. But in reality, say the experts, that’s not always possible (or recommended).

One main reason: organizations aren’t monoliths. They consist of different departments, with different priorities and requirements. They also have different customers and constituents with their own priorities and requirements. 

As Shane points out, “You’ve got, in many cases, interwoven priorities” in the business that don’t always align, and in fact may be very different. 

Only a hybrid/multi-cloud approach can provide the flexibility needed to serve all needs, based on requirements. When on-premises infrastructure makes sense, it should be used – but when it doesn’t, there have to be other options.

Of course, the inverse is true: that just because you can build a multi-pronged, multi-layer approach, doesn’t mean you should. Both Adam and Shane agree that simplicity is the goal (although not at the expense of adopting a “lowest common denominator” solution that is just okay for everyone). 

Finally, one virtue of having at least some of your infrastructure on premises is to do something that’s anathema to public cloud providers: destructive experimentation.In other words, sometimes, to improve a system, it’s necessary to pull the plug and break something every now and then – something that’s only really possible if you own your own hardware and software.

As Shane says, “Let’s keep pushing the button and try and figure out where it can fail. And let’s reward people for trying to figure that out. Because those inherently will come back to us.”

Facing the Convenience Vs. Control Debate: What’s the Solution?

There is a paradox between convenience and control. Everyone wants convenience, but obtaining it by incorporating public cloud into your IT infrastructure sacrifices control over assets (something that only on-premises access can guarantee).

This paradox raises the question: Is there a way to make use of public cloud and maintain some measure of control? 

Or are corporate and government entities resigned to making choices between two systems that just aren’t compatible?

For starters, it’s important to realize the nature of the challenge. “Public cloud, as well as private cloud, has gotten immensely complex,” says Shane, “and there isn’t a blueprint for how we do all of these things together,” given the shifting nature of regulatory requirements and compliance challenges. “That’s frankly not probably going to get a whole lot easier anytime soon.”

What’s needed to resolve the challenge, according to the panel, includes:

• More formal and hands-on education about the environment that leads to an up-skilled labor force
• A clear delineation of security requirements and a recognition that on-premises control may be needed for those (but not for things that can and should be externalized)
• Advancements in how public cloud providers handle security
• More industry-wide consensus on operating models, security procedures, and resiliency

A hybrid approach is the ideal way to address the challenge, as long as organizations, both public and private, know what they really need and not just what they really want.

Can Everything Be Virtualized – Or Is Hardware Still An Issue?

Cloud proponents would love to believe that any IT infrastructure can be virtualized completely and eliminate the need for on-premise hardware – or, indeed, any private-side hardware outside of computers and peripherals.

And yet, according to Adam, not everything can be virtualized – and in many cases, it shouldn’t be.

“Because not everything is software,” he says, “there are certain cases where you’ve got specialty hardware systems that you’re connecting. While the rest of the workload may make sense to have operating somewhere else, some physical things may need to remain behind to be able to work.”

Besides, as Adam points out, there isn’t one provider model that fits all. And even if there was, “You just can’t beat the speed of light.” For certain customers and constituents, proximity will always be a priority.

The bottom line: to get it right with a hybrid/multi-cloud environment, flexibility is the cornerstone of any successful operational model – as long as flexibility is based on understanding. 

For more information from the panel, view the session and see what experts think will be the future of IT in the era of the cloud.