Private 5G is a non-public mobile network that uses licensed, unlicensed or shared spectrum. This is best suited for the areas where the spectrum provided by a mobile network operator (MNO) is restricted for government use only or is not available. Many organizations around the world have either deployed a private 5G network or are in the process of deploying one to secure data inside their premises.
Edge computing is a distributed computing model and open information technology (IT) architecture that moves workloads and compute operations away from centralized points closer to the data source. This data is processed on a device or by a local computer or server instead of being transmitted to a data center. This helps reduce the roundtrip latency and improves performance.
Red Hat's approach to edge computing focuses on three use case categories: enterprise edge, operations edge and provider edge. There are multiple business use cases around private 5G and edge computing. I have captured just a few here.
Consider the case of a clothing manufacturer who wants to fully automate their facility. It starts with opening the gate for a loaded truck of raw materials entering the facility and continues until the end products are delivered to their respective cubbies. There are several additional steps involved to get the end products ready for shipping.
Private 5G and edge computing can play a crucial role in fully automating the facility. All the devices—including video cameras equipped with sensors, robotic arms, autonomous vehicles, conveyor belts, chutes and so on—are connected through a private 5G wireless network, and a 5G tower inside the facility keeps the network connection active and accessible. Each device has an edge application with an embedded AI/ML model that helps in inferencing and making quick decisions.
The video cameras at the gate are equipped with sensors and are connected through private 5G. These scan the truck’s license plate that the AI/ML model then recognizes and the edge application makes decisions accordingly. Using the private 5G network, the robotic arm and autonomous vehicle are notified to get ready. Once the truck enters the facility, the robotic arm—instructed by the equipped AI/ML model and edge application—will offload the raw materials, and the autonomous vehicle will carry it over the respective aisle and shelf as per the coordinates provided by its edge application. When the end product is ready, it is sent to the respective chute via conveyor belts, which then take it to its respective cubby.
The edge classification in this use case is a typical example of operations edge since the AI/ML model inference and edge application make a decision at the site itself. In this case either little or no data is stored at the facility and the model continues learning and improving over time.
Consider the case of a hospital that would like to automate their facility to provide 24/7 care to their patients, especially those admitted in ICUs. Every ICU room will be equipped with medical devices such as oxygen monitor, heartbeat monitor, blood pressure monitor, nurse station, etc.
Let’s see how private 5G and edge computing can help in automating this facility. All the medical devices (oxygen monitor, heartbeat monitor, blood pressure monitor) are connected to the hospital’s private 5G wireless network and the 5G tower inside the facility is keeping the network connection active and accessible. Each of the devices has an edge application with embedded AI/ML model that helps in inferencing and making quick decisions.
In a situation when the oxygen monitor indicates that the oxygen level has gone down, the AI/ML model interprets the data provided by the monitor and the edge application quickly sends an emergency alert to the nurses' station as well as a page in case the nurse is elsewhere. In this critical situation, there is no room for latency.
Another situation involves the maintenance of patient records so medicine is prescribed based on the patient's medical history. During certain emergencies, these records need to be accessed quickly, so they have to be maintained in a centralized local database.
In this use case, the edge classification is an example of operations edge and enterprise edge since in the first situation the AI/ML model inference and edge application makes a decision at the site itself, and in the second the patient records are maintained inside the local database.
Telecommunications, media and entertainment industry
Consider the case of a service provider that would like to expand their 5G network in the area where the spectrum is unavailable or restricted to only government use.
A private 5G and edge computing solution is a good way for them to get the unlicensed network available to the organizations in that area. They may either get the in-house open radio access network deployed at the site or purchase it from provider companies like Mavenir. Open RAN is the critical component of 5G that processes the incoming and outgoing radio signals.
This use case is an example of provider edge since the 5G open RAN service is provided by the service provider.
Red Hat 5G edge solutions
Red Hat Enterprise Linux (RHEL) is designed to meet the needs of the hybrid cloud environment and is ready for you to develop and deploy from the edge to the cloud. It can run your code efficiently whether deployed on physical infrastructure, in a virtual machine, or in containers built from Red Hat Universal Base Images (UBIs).
Red Hat OpenShift is a hybrid cloud container orchestration platform used to deploy and operate a hybrid mix of containerized applications. It is well suited for running the private 5G and edge workloads since the critical component—single root I/O virtualization (SR-IOV)— is available as an operator that creates the virtual functions (VFs) on Peripheral Component Interconnects (PCIs) to make the multiple virtual machines (VMs) access the same PCI.
Red Hat OpenShift Data Foundation is a software defined storage service for containers that can be deployed on edge, on-premise, or on public cloud and has the capability to provide file, block and object storage classes.
Red Hat Advanced Cluster Management for Kubernetes is a platform that manages the multiple clusters from a single pane of glass including the cluster deployed on edge. It also deploys the applications and enforces the security policies across multiple clusters.
Red Hat Ansible Automation Platform extends a consistent automation experience across cloud, datacenter and edge, enabling organizations to scale automation in heterogeneous environments. With a common platform organizations can build, run and manage the entirety of their highly distributed systems with security-focused automation workflows delivering resilience, across a wide ecosystem and even to remote locations where network connectivity may be intermittent.