When we think of cloud computing, most of us envision large-scale, centralized data centers running thousands of physical servers. As powerful as that vision sounds, it actually misses the biggest new opportunity: distributed cloud infrastructure.

The promise of edge computing

Today, almost every company in every industry sector needs near-instant data to be successful. Edge computing pushes applications, data and computing power services away from centralized points to the logical extremes of a network. It enables companies to put the right data in the right place at the right time, supporting fast and secure access. The result is an improved client experience and, oftentimes, a valuable strategic advantage. The decision to implement an edge computing architecture is typically driven by the need for location optimization, security and, most of all, speed.

New applications such as virtual reality and artificial intelligence, with requirements to collect and process massive amounts of data in real-time with extremely low latency, are driving the need for processing at the edge of the network. Very simply, the cost and distance of the hub-and-spoke model will not be practical for many of these emerging use cases.

Make no mistake: edge is not the end of cloud computing; it is the natural evolution. As more devices generate more data and demand for computing and storage, it becomes more efficient to push cloud capacity to the edge.

“Edge computing pushes applications, data and computing power services away from centralized points to the logical extremes of a network.”

Real-world applications

This was demonstrated at the OpenStack Summit in Boston last month by Beth Cohen, cloud technology strategist at Verizon. Verizon is leveraging OpenStack for its Virtual Network Solutions product, a massively distributed network-as-a-service solution, delivered via Verizon’s new universal customer premises equipment device, which is about the size of your home router. This device is being used around the globe to deliver cloud-based services like wide area network optimization, security and routing, with more services in the works.

Telecoms are pioneering edge cloud, and cell towers make an obvious entry point, but many industries including retail and manufacturing are following closely. From robotics-driven warehouses to oil rigs to self-driving cars and hospitals, the use cases are only beginning to be defined. Eventually, your mobile devices will be connecting to mini data centers in your coffee shop or even at the end of your street.

These pioneers are taking several different approaches as they define the technology stack for edge computing. Just like public cloud at scale, edge computing requires a convenient and powerful cloud software stack that can be deployed in a unified, efficient and sustainable way. Whether it’s a mesh or hierarchical architecture and regardless of where the control plane lives, automation is the key, and zero-touch provisioning is the goal. Ultimately, it will be a combination of open-source technologies that will drive this wave.

A hybrid approach

And even as the use of edge computing and private clouds grows, hyper-scale public cloud is certainly not going away. We are clearly seeing hybrid as the dominant approach. In fact, the more innovative cases we’ve seen are organizations who are making OpenStack work in a small footprint at the edge, while continuing to work with larger-scale OpenStack private clouds and public cloud providers.

This September, we are planning to convene a targeted working conference called OpenDev focused on defining the technology stack for edge computing. It’s intended to give users a forum to discuss how different technologies will work together to achieve this goal.

The cloud is rapidly expanding to the edge. As it does, new opportunities will emerge and new skills and technologies will be needed.