The good, the bad and the ugly: enabling communications across Africa

09 October 2023

James Gray, director of telecom strategy, PowerX Technology Limited

The magnitude of the task facing the telecommunications industry in Africa is profound. It’s the world’s fastest growing economic zone, home to 1.2 billion people, yet shackled by an infrastructure gap that presents unique challenges in providing access to regions, resources, communications, and markets.

The World Bank is unequivocable about the economic benefits of communications and connectivity. “Internet access can drive economic development through its impacts on both the supply-side and the demand-side of an economy. When infrastructure expands in developing regions, workers (…) gain higher wages or find employment. Digital connectivity directly affects the productivity of firms, workers, and other inputs in the production process.” And unique to Africa, mobile connectivity is the backbone of the pervasive M-Pesa branchless banking system – developed originally in Kenya – that allows users to transfer money, deposit, withdraw, and pay for goods and services with a mobile device, transforming local economies.

All these benefits rely on robust, high bandwidth networks – predicated on the ability of individual sites to reliably serve their subscriber base. In a typical developed market – say Germany – it’s not unusual to see an LTE base station to subscriber ratio of about 1:1000. Some markets have even lower ratios – Japan provides almost double the base station density for example, and Finland has an astonishing ratio of around 250 subscribers per base station. For Africans however, the numbers tell a different story. Take Tanzania, where around 3,500 subscribers fight for the bandwidth of every base station, or the DRC where this ratio climbs to around 6,500 per site. That’s over six times as many people being served mobile connectivity from an individual site compared to a median developed economy.

The challenges are familiar. Poor grid infrastructure with low quality power and outage issues, high transportation costs due to low quality roads and difficult terrain, high dependence on diesel (along with high fuel prices and shortages), operational leakages (including theft, vandalism, and diesel pilferage), lack of technical skills, and uncertain policy/regulatory environments.

So how can data science prevent or safeguard against these problems?

Much of the difficulty in maintaining and servicing remote rural sites is in the unpredictability of events that require intervention. Towercos – and the complex ecosystem of vendors, operations and maintenance suppliers, subcontracted electrical and mechanical engineers etc. that support them – are more often than not reacting to unforeseen (but not necessarily unexpected) problems: a malfunctioning battery, an unexplained disconnection from the grid, a missing rectifier module. These are the type of outage-inducing problems that must be fixed now – regardless of whether or not a bridge has been washed out, an armed conflict is underway, or a gang of fuel thieves is operating on the road to the base station.

But most of these emergency maintenance events can now be pre-identified – or flagged as highly probable – by employing the sophisticated data mining and analytical tools found in artificial intelligence (AI). Trawling through unimaginably vast troves of archived and real-time data, collected across networks that span thousands of base stations and towers, AI algorithms can identify anomalies and patterns never before accessible to engineers and operators. This unique insight into potential problems puts the towerco – for the first time – in the driving seat of preventative maintenance and problem-solving.

Now, instead of having a maintenance truck roaming a region stacked full of replacement parts that might be needed in the event of an unforeseen outage, a real-time feed can alert an operations and maintenance teams that a specific battery at a specific site has a high probability of failure within a certain timeframe. A targeted intervention can take place, under conditions and on a schedule controlled by central and regional operations management teams.

In the example of battery replacement – just one of countless components that require maintenance and/or replacing – there is a hard financial benefit. At present, batteries that are still operational are swapped out on a fixed schedule (usually every three years). But as any electrical engineer will tell you, they usually have a good couple of years service left in them, even though they’ve been removed ‘just in case.’ By using data science and AI analytics to more accurately predict when a specific one will fail, they can be left in the field longer – extending their lifespan to make significant reductions in maintenance costs.

This predictability has ripple effects through a towerco’s OPEX. As well as reducing miles driven to sites, maintenance fuel costs, wear and tear on vehicles, risks to human lives and potential losses due to environmental hazards, CFOs can stabilize and forecast cashflow better than ever before. It extends the window of financial predictability, which in turn benefits the roll-out of new regions and base station in-fill.

The benefits don’t end there. A key feature of AI oversight is the ability to reduce diesel fuel costs at individual tower sites, adjusting the switch from solar/diesel/grid/battery depending on the unique conditions at the site and machine-learned efficiency gains across the entire network. A reduction in diesel usage means fewer deliveries – and fewer opportunities for fuel trucks to fall foul of predatory gangs that have increasingly plagued many parts of the subcontinent.

Building out and maintaining the tower networks in Africa is important, essential work. The physical and logistical challenges this presents are formidable. But by turning to the technology of the future, towercos – and the complex ecosystem of suppliers and vendors that support their endeavours – can begin to do this work on their terms and under their chosen conditions. Predicting the unpredictable is within our grasp, giving a much-needed boost to the stability and growth of the African telecommunications industry.