Telecoms.com periodically invites expert third parties to share their views on the industry’s most pressing issues. In this Q&A Ken Gold, Director of test, monitoring and analytics solution at Exfo, discusses the role of AI in delivering on 5G’s promise.
We hear a lot about how 5G is going to change expectations not just for people, but also for machines. What does this mean?
The best way to explain this is to look at the ways 5G can be used and how it’s different from 3G and 4G. 5G comes with lower latency, which means we’ll be able to build applications that can do some very sophisticated things. We’re talking about being able to do highly critical and personalized processes using machines or “things.” For example, autonomous vehicles in a manufacturing area or being able to find routes and prevent collisions would require instant responses.
Whereas people might be tolerant if a video was slow to stream or a call was dropped on 3G or 4G, they simply be won’t have the same levels of tolerance if the situation is life or death. It’s a real shift in expectation. As a patient, you’d expect the machines involved to run on a stable network and the applications to have accurate, real-time data. As an industry, we’ve got to make sure that we can deliver not just on the 5G expectations people will have but also those of machines.
Is healthcare the industry that will embrace 5G more than any other?
The 5G use cases stretch across the consumer and business world, but it’s the business applications that will be adopted first because there is so much to gain in terms of operational efficiency and cost saving.
Healthcare will be one of the forerunners with manufacturing. In fact, the two sectors are predicted to make up about 40% of the 5G uses cases. The consumer side of things will be a little slower to get going, simply because handsets need to reach a critical mass and consumer “killer apps” need to be built. Ericsson studies show that the revenue growth from consumer will only be about 1.5%, compared to 35% in enterprise.
How will operators deliver on the expectations?
The answer to this question centers around the notion that operators must have full visibility across the 5G supply chain—that’s the core, transport and RAN as well as links controlled by partners, like app vendors and devices.
That’s going to lead to a transformation in how networks are designed and built, and how we assure they deliver what they are supposed to. Critical 5G services are only possible if discrepancies in expected behavior are predicted and detected in real time, and then fixed before there’s any impact.
Since the supply chain will be complex and made up of many different parties, there will be a huge demand for monitoring, processing, collecting and correlating data from many sources. That’s only going to be possible through partnerships.
Where does big data fit in?
It doesn’t. You can’t achieve the level of data management 5G demands with big data. Firstly, big data systems are not nimble or precise enough to keep up with dynamic service delivery and assurance. Secondly data repositories are generally operating in silos, which is counter-productive to the use cases we’ve been describing. And lastly, as 5G scales, compute and storage costs will become more and more prohibitive.
Actually, what you need is “small data” to deliver the right data, in real time, in context. This is where artificial intelligence (AI) can help. The network data, service-layer data and then the user-level data can be fed into AI tools and applications to align the data across the supply chain and raise alerts on anomalies.
These tools and apps only use what they need, when they need it, so all links—5G core, transport, apps, edge functions, RAN, devices and end users—stay in sync, so services operate flawlessly and deliver impeccable user experiences. In effect, AI is what will help create a smart 5G supply chain.
To support this new model, we’re already starting to see the emergence of system integrators that focus on specific industries, such as Manufacturing 4.0, and the emergence of hyperscalers to deliver the cloud infrastructure needed in the network core and at the edge.
Are we going to see a new era of data standards emerge too?
Yes, and it’s already happening. Take for example the Network Data and Analytics Function (NWDAF). It will standardize the collection of, and access to, network data. This will underpin the small data concept. It will also ensure consistent approaches to correlating and analyzing insights that inform the actions that need to be taken, in real-time, to assure high levels of application performance and user experience.
Success will rest on operators adopting cloud-native components that are easy to integrate. Choosing solutions with open APIs will be imperative because it’s the best way to ensure the network can adapt to future data needs.
Overall, I believe the mobile network operators that embrace partnerships and a modular approach to service assurance will be the ones that start to drive 5G revenue first. Indeed, these service providers won’t just meet the new expectations of people and machines—they will exceed them.
Ken has a diverse background in the telecommunications industry spanning more than 30 years. He is currently focused on understanding how virtualization and 5G will change the way people and machines communicate and how the carrier will leverage automation, artificial intelligence and machine learning to manage quality of experience in this new reality—which in the end is all that matters. Prior to joining EXFO, Ken held leadership roles in Solutions Marketing and Product Line Management with Nortel, Harris Broadcast, ADTRAN and Accedian Networks, focused on optical transport and Carrier Ethernet service assurance and performance monitoring.