Decisions at the speed of light: how AIoT will allow 5G to take flight

To deliver a truly resilient and customer-centric network, telco operators have no choice but to automate. The latest AI and machine learning technologies can monitor, operate and optimise the 5G networks of the future, delivering an unlimited connectivity service that’s stable, cost-effective and competitive.  

A big data problem

The anticipation surrounding 5G is enormous. By connecting everyone to everything, and across every sector, the technology is widely expected to kick off a new industrial and technological revolution.

Yet the advent of 5G invites a massive capacity and operational challenge. As more customers and organisations migrate, 5G networks are expected to grow to cover 65 per cent of the world’s population and carry 35 per cent of all its mobile data by 2024. For telecommunications operators, this data will be generated in unprecedented volumes. Telecommunications operators will need to constantly invest in their infrastructure to support the capacity increases needed. 

With the number of devices connected and communicating within the Internet of Things (IoT), and the high-speed, high-bandwidth possibilities of 5G, traditional data collection and analysis is no longer sufficient. For the sake of planning, running and optimising customer experiences, operators will have a limitless need for more intelligent decision-making. Decisions also need to be taken across domains and silos in a split second to meet customer expectations and honour service agreements.

Delivering on this promise would be impossible if we could only depend on human staff and operators to generate insight and make decisions. Fortunately, AI and machine learning have a vital role to play here. When you merge AI and IoT, you get the Artificial Intelligence of Things, or AIoT – a revolutionary combination that can transform industries, elevate customer experiences and accelerate business performance exponentially.

By 2022, Gartner predicts that more than 80 per cent of enterprise IoT projects will include an AI component, up from a mere 10 per cent today.

The best laid plans…

Telcos should seek to automate processes continuously during 5G rollout, with each step iterating on the last. By doing this, operators can sidestep major investments and overhauls by performing focused adjustments and expansions that both improve network performance and enhance the user experience.

However, before you maintain or enhance a 5G network, you first need to build it. Even at the planning stage there is enormous potential to make use of AIoT.

In today’s fast-moving, digital world, telcos are under tremendous pressure to deliver speed and scalability. To compete, they must be able to move resources, deliver connectivity and construct new capacity rapidly. However, it isn’t easy to respond with agility when, to supply connectivity to a new site of location, often requires the building of a new radio tower or data centre and significant civil engineering.

Network planning is often time and resource-intensive because the work is so data-heavy. Before a decision can be made, operators need to carefully review and analyse an assortment of population tables and network traffic. The objective is to have the cell site built and ready before customer demands start flooding in, but it’s challenging when the planning process is so complex and time-consuming.        

AIoT solutions are invaluable here because they can do much of the heavy-lifting for you. AIoT and analytics can process and produce insights from an immense amount and variety of data faster than any human can. With this information, operators can truly predict - rather than simply respond - to demand, ensuring they can build a cell site in the most lucrative location before the competition moves in.  

The customer comes first

AIoT really comes into its own, however, in network operations. 5G networks are micro in nature, made up of thousands of disparate and often siloed cell sites and data centres, each carrying and processing immense quantities of data. Keeping track and monitoring all of these sites, ensuring they are working properly and running efficiently, won’t be feasible without some form of automation.

You also need to consider the huge amount of data that needs to be analysed. It’s no longer feasible to move all of that data to a data centre or the cloud before you are able to detect network anomalies. This means that you need the ability to deploy AI out to the edges of the network, analysing the data in-stream, as it’s created. Actions can then be taken to fix problems before they disrupt the network, by detecting patterns of behaviour that provide an early warning of likely issues.

AIoT can predict network faults based on historical data and real-time, continuous analysis. It uses many different prediction models to work out the probability that a set threshold will be breached, alerting operators to the threat before it can snowball into a crisis.

The process is even more rigorous when machine learning is deployed for anomaly detection. Once a capacity issue is discovered, AIoT can search for the root cause and find the answer fast. From a business perspective, this enables automated monitoring and helps deliver on customer priorities like KPIs and SLAs.

Operators are still some time away from ‘closed loop automation’, where AIoT systems are able to automatically resolve network issues and provision resources. But the building blocks of technology to do this already exist. The challenge is capturing all the data that skilled workers use to make decisions, so that the decision-making process can be automated. For the time being, however, operators can benefit greatly from their workers acting in concert with AI solutions. They can work symbiotically, with the AI performing the analysis and providing the insight, while the human operators use this intelligence to solve challenges. Not only does this reduce human error, it frees the experts to focus on more tactical and valuable tasks. 

Ultimately, 5G brings challenges but also great opportunities to the telco market. It will introduce an unprecedented level of complexity and disruption, which can be mitigated by adding a layer of automation to the process. AI-driven automation is an opportunity for operators to meet their business goals and deliver resilient, on-demand services optimised to meet the needs of the customer.