Open RAN and vRAN — along with multi-access computing (MEC) — are enabling the creation of highly-distributed 5G networks to support a plethora of new applications that require ultra-low latency, very high bandwidth, ultra-reliability and greater device density.
Unfortunately, history shows us that even the most reliable systems inevitably fail. What is different with 5G technology is its potential to enable new, never-before-seen applications supporting advancements in robotics, remote surgery, self-driving cars and more. In other words, the stakes have never been higher.
For example, what happens if there is a glitch, transient service degradation, or worse, outright failure in the network? What if hackers are able to penetrate this new RAN and mobile edge? The same 5G technology that will bring intelligence, data, ultra-low latency and ultra-high bandwidth to the corporate campus, the factories, the mines and the branch office could also put those products, services and brands at risk.
These examples are not chosen at random, either. Automakers anticipate a majority of businesses will rely on automated fleets over the next two decades. Similarly, a recent survey of executives at leading manufacturers like ABB, Audi, Bosch, GE and Siemens found that more than nine in 10 are integrating 5G into their operations, with 87% already seeing benefits. To realize the potential of 5G, network operators need to ensure the constant availability and security of these networks.
The 5G RAN is a patient that needs constant care — and new monitoring tools
The RAN or mobile network edge is the lifeline for these new applications as it contains the application intelligence and data for the 5G devices and services. Understanding how carrier systems are performing is vital to the applications that rely on them.
To use an analogy, imagine the RAN is like a critical care patient. Other patients or networks might get by with just a regularly scheduled check-up. Taking the weight and pulse (of traffic) is a basic need, but deeper diagnostic tools are needed that see the entire inner workings of more complex networks. Traditional assurance and security tools were built to assure and protect large aggregations of traffic at the data center and central office. The cost and provision of these tools were not designed for cloudified networks distributed out to the edge of the network. Instead, disaggregated and decomposed network elements and services must be “reconstructed” and seen in the chain that forms the interconnected service or services.
Much like a virus that has invaded a body, the symptoms of a runny nose, body aches and fevers may each be treated and somewhat ameliorated; however, the underlying cause remains and may only get worse. Until that virus is treated, or the body’s immune system kicks in, the system will continue to exhibit problems and may further decline in its operation. In this scenario, a physician may need more powerful tools or tests to diagnose and treat a patient. In the same way, network engineers need greater resolution – with tools that offer a composite view of the system, much like an MRI-type view that shows a slice of the body but also a completely reconstructed view of the body. And arguably, beyond that, a neurological exam that peers into the cloud of the nervous system is needed when more complex problems appear.
Open RAN/vRAN certainly adds a new level of challenge of diagnosing problems for 5G mobile network operators. Up until recently, the RAN was usually purpose-built by a single vendor. By opening it up, the RAN can now be disaggregated into its component pieces, including the Radio Unit (RU), Control Unit (CU) for signaling (CU-C) and user plane (CU-U), Distribution Unit (DU), PHY and the Radio Intelligence Controller (RIC). Keeping all these disparate, cloud-based components from degrading and failing is no small task. Successful deployment requires multiple rounds of testing and certification to ensure the interoperation of microservices and software components. Doing this at a nationwide scale — and then supporting upgrades and handling the change of vendors — is an ongoing process. Automation will help in the operation of cloudified networks going forward, though the task of maintaining interoperation and vendor management will always be necessary and requires greater end-to-end visibility to maintain reliable service.
On a final note, it’s worth discussing the potential for cybersecurity threat actors to hack 5G infrastructure. Putting so many disaggregated network elements into the cloud (or even different clouds) greatly expands the attack surface for hackers to gain access to the network. Additionally, carriers that rely on network components from multiple vendors (e.g., via vRAN) could expose themselves to vulnerabilities from uneven and poorly connected security between different vendor parts.
Somewhat worryingly, these vulnerabilities are not purely hypothetical, either. A 2022 survey conducted by Nokia and GlobalData found that 75% of communication service providers have experienced up to six security breaches of 5G networks within the past year, often coinciding with a service outage. Additionally, researchers have demonstrated that configuration errors and device vulnerabilities could allow attackers to move laterally across 5G network slices. So what’s the bottom line? Complex networks require complex solutions to maintain their health and security.
Fostering healthy outcomes with automation and network visibility
Automating the 5G network with machine learning (ML) and artificial intelligence (AI) for intelligent maintenance and operation of complex 5G networks is already in the works. This type of automation can act as a healthy immune system response to extend the earlier analogy. But simply applying ML and AI routines and algorithms to traffic will have limitations – especially until the technology matures further. After all, it’s hard to automate processes that haven’t yet been fully defined or well understood. Some aspects can be anticipated from experience with the past transition of 3G to 4G, of course, but others are very much unknown, and there are many more variables with cloudified networks.
Most carriers building cloudified 5G networks will continue to rely on a mix of new automation technologies and the practical experience of their network engineers built over decades to solve problems as they develop. Bridging this strategy together is a layer of visibility at the network, service and device level to detect and resolve performance issues quickly — whether they arise from the RAN or elsewhere.
Ultimately, carriers are right to be concerned about the reliability and security of their networks. No one wants or expects their network to fail but pretending that it never will does nothing to diminish exposure to that risk. There is just so much more that can go wrong with 5G networks, especially when discussing topics like Open RAN, vRAN and multi-access computing and the new applications that will run on it. To protect the health and vitality of networks, carriers must expect the unexpected.