Cloud and edge are too often seen as competing alternatives, when in fact they're part of the same continuum of putting collection, compute and storage in the most effective place, done in the most efficient way.

Edge adoption will ultimately be driven by business value, whether that's cost savings from consolidated hardware that's easier to manage or the availability of applications and services that solve business problems better at the edge because that's where the users or the data (or both) are. If the edge is more complicated to manage or develop for than the cloud, cheaper hardware or lower latency won't be enough to attract users. While Kubernetes at the edge is one option, what many organizations want is the cloud services they're familiar with brought to the edge where the data they want to process is created, on their choice of infrastructure - and with the flexibility of cloud billing.

The original hybrid cloud promise was the power of public cloud running in a private cloud on your own infrastructure, but also coupled closely to public cloud infrastructure for any workloads that didn't need to be on your own infrastructure, because a private cloud will rarely offer the economies of scale and innovation hyperscalers can offer. As more (and more demanding) workloads have proved suitable for public cloud infrastructure, in many cases what's left are workloads that are best suited to the edge.

That's not redefining edge to be the same as on-premises computing, because this cloud edge is deployed and managed very differently. And while some of this cloud edge infrastructure might be in on-premises data centres, more of it will be in new edge data centres, embedded in edge devices or even built right into the telecom infrastructure.

This was the chapter on hybrid cloud and the edge that I wrote (with Simon Bisson) for the State of the Edge report 2022: you can read the rest of it here:

• edge

• cloud

• hybrid cloud

• Linux Foundation