Building Supercloud Starts with Multi-Cloud Networking

A funny thing happened on the way to supercloud. It started happening while we weren’t paying attention.

Let me explain.

First, let’s simplify the most excellent definition of supercloud from Cornell University that I shared last year to include the goal of supercloud, which is what helps guide the market in delivering an actual working solution.

“Supercloud is a cloud architecture that enables hybrid IT to seamlessly operate every layer of the IT stack across cloud providers, and on-premises traditional and cloud environments, and at the edge.” 

Now, you'll note the emphasis on the IT stack. That's intentional because even when we view supercloud through the lens of an overarching API, we do so with the intention of being able to provision, operate, and secure all the composite pieces of the stack. That’s basically data, infrastructure, and apps.

And when we look at that stack, the first thing we notice is that apps and data are dependent on infrastructure – notably the compute and network components of infrastructure. So, it would seem logical – or at least reasonable – to assume that before we worry about seamlessly operating apps and data across multiple environments that we’re able to seamlessly connect (network) across environments.

That’s why it’s no surprise to me that over the first half of 2023, the market has seen significant activity focused on what’s become known as multi-cloud networking. Our annual research found multi-cloud networking to be one of the “most exciting” technologies of 2023. That was not surprising, either, as organizations have been struggling with the complexity of cloud APIs and tool sprawl for more than five years now. Multi-cloud networking proposes to address at least part of that pain by simplifying the network and connectivity challenges that plague about every role within IT.

The thing is that supercloud starts with multi-cloud networking, but it can’t stop at the traditional network boundary. A solution that actually meets the definition of supercloud must continue to climb the stack, as it were, and enable seamless operation of applications and data, too. 

That implies an encapsulated – platform – approach to building the supercloud. Encapsulation solves for a unified control plane and the underlying network and connectivity complexity by taking a platforms approach that exposes operational control of the IT stack as services that require less domain-specific knowledge. This does not rely on a superset of APIs. All the use of APIs are encapsulated in the "supercloud," and only a small set of technical details are required for the platform to configure the infrastructure and app delivery services to connect and secure them across properties.

Encapsulation, aka "a platforms approach," offers the same simplification of interfaces that eliminates variances in APIs across cloud environments. But underneath is a platform that further encapsulates the complexity of networking and connectivity at every layer of the stack. The complexity is still there, but it's hidden by encapsulating it in the platform, so IT ops doesn't necessarily need to know how to make the networking and connectivity work within and across cloud properties; the platform does.

This means operators can effectively deploy an application in any public cloud and enable either private or public access without demanding a Ph.D. in networking or security. This is important because any implementation of supercloud must be able to support all three layers. Ignoring any of them introduces friction and complexity that reduces the benefits of the approach.

This is the way, and it’s what we’re seeing happen in the market right now. Supercloud will become a reality, and the path to get there starts with multi-cloud networking. 

Related articles: