Private 5G and Edge Computing for Industrial IoT

Enterprises are wrestling with transforming, from hosting all their workloads within their data centers to a hybrid, multi-cloud world where they use a combination of dedicated instances and leveraged As-a-Service functions. 5G and edge computing should be an aspect of that transformation, not another layer of complexity on top, and most certainly not a step back to managing a large of amount of disparate compute, storage and network resources in a bespoke manner.

Today’s Industry 4.0 customers are trying to take advantage of applications such as digital twins, video-based safety and efficiency analytics, and precision material and asset tracking. To accomplish these tasks requires reliable connectivity that is power efficient with high bandwidth andcompute power located in the right locations.

There are many flavors of private 5G – but which one is for me?

The private network space is crowded with solutions for industrial manufacturing, and manufacturers have several options, from connecting to a public 5G network, or outsourcing their own private 5G network to a managed service provider or hyperscaler, purchasing infrastructure while contracting operational support from a mobile operator, or building and maintaining their own 5G network.

This leads to confusion in the market about what private 5G means and how to approach it – the sheer choice in technologies and spectrum available as we inch our way closer to the edge requires multiple decision makers from C-level to the network engineer. One important consideration is that it is not an either this or that choice. Regardless, if you are connecting with Ethernet, LTE, Wi-Fi 6, 4G or 5G – the underlying network infrastructure should be agnostic. Underlying network architecture must be programmable and flexible to meet the needs of a dynamic network whether you are dealing with a fleet of automated robots on the manufacturing line or connecting multiple plants across the globe.

Here are the Top Five Commandments of Private 5G:

1. Think Edge. Critical manufacturing and automation will require latency below ten milliseconds. Unfortunately, the public cloud is currently only capable of sustaining latency radically higher than ten milliseconds. Edge computing is an absolute necessity in Private 5G, shifting computing power from the core network and a centralized cloud architecture, and distributing it to the edge. The goal is to lower the latency and improve performance for mission critical applications.

2. Programmability is essential in modern networks. Manufacturing must future proof systems with the ability to deploy, manage, and troubleshoot a network in real-time to solve for the evolving demand from different applications. Disaggregating the network control plane from the data plane is essential to ensure management of the network is no longer bound to the networking hardware. Programmable networks can be controlled “top-down” to install any functionality the user wants. They are dynamic.
3. Continue to break down the silos between different disciplines of the IT stack. We must put compute and networking together in a singular package. At the heart of this new edge architecture is a unified solution for distributed edge computing. Innovation and new thinking are required to succeed. For the longest time, network, compute, and storage capabilities were deployed as “silos” with each running its own control plane or operating system (OS). As much as automation has eased the deployment, configuration and management of these elements, the need for common services (i.e., zero-touch provisioning (ZTP), upgrades/downgrades, scalability on demand, etc.) has generated different OSs, including network OSs.
4. Cloud technologies are extendable everywhere. In the Industry 4.0 environment, agility is needed to respond to rapid fire queries for analytics and for real-time automated operations with AI. Solutions are needed that provide the ability to access the diversity of resources to respond to application needs not just network but also compute, storage, GPU etc. Cloud-native network fabric software increases the efficiency of space utilization at the edge by using containers to replace virtual machines. Using the same orchestration layer for these components make resources at the distributed edge and hybrid cloud better streamlined and optimized for rapid deployment of network resources for edge applications.
5. Demand is exponential; understand your needs and set up the right network from the start. Be wary of seemingly simple turnkey solutions that may not necessarily solve for all the complexities that will come. Any standalone Private 5G solution will need to be extensively customized to ensure that everything functions smoothly with your organization. It is clear the demands on wireless networks are only going to increase as new applications come online that demand bandwidth.

Deloitte predicts that over the next decade hundreds of thousands of companies will deploy private 5G networks. About one-third of the networking executives they surveyed prefer to purchase full end-to-end advanced wireless solutions, but the other two-thirds prefer to acquire best-of-breed components or a blend of components and full solutions, offering more of an opportunity to customize and control.

“Today’s reality is that no single vendor can offer full-stack, end-to-end solutions on its own. While some may already have the vast majority of the required components and capabilities, they still need to partner with others to acquire and integrate additional pieces to create complete solutions.” - Deloitte

Building 5G networks requires a fundamental shift that emphasizes support for edge-native apps and delivery of high performance at the lowest possible cost. The industrial cloud driven by private 5G will give enterprise the power to pick and choose network and application elements to customize the evolving framework that meets their needs. This will provide a viable environment for enterprises to innovate and transform.