Mobile data: the challenges ahead

10 February 2021

Southern Asian telecom operators, like their counterparts across the rest of the world, have been preparing for 5G for some time. After all the hype, 5G is here. Yet the impact of the COVID-19 pandemic and the upsurge in traffic usage might prove to be a valuable learning experience as operators ready themselves for 5G.

The most obvious difference for operators came about as countries in Southern Asia imposed strict lockdown measures limiting people’s movement. From Bangalore to Bangkok and from Karachi to Kula Lumpur, businesses switched almost overnight to remote working, leisure and social activities went online and school children in many countries in the region joined virtual lessons – or worse still, had their education put on hold altogether. Many workplaces and other institutions adapted well to the measures, however, any hopes of a pre-COVID return to work appear less and less likely as infection levels are rising globally and new measures appear inevitable. Indeed, many businesses and their employees found remote working productivity levels easily sustainable, and cost-efficient.

Operators felt the effects of traffic usage instantly. In the Asia Pacific region as a whole, over half of operators reported traffic increases of 60 - 80%. Globally, one in five experienced a 20-40% increase while 13% experienced a surge of 80-100%, according to a recent research report from the Technology Innovation Council (TIC). For operators, COVID-19 has been a scaled-down version of the traffic demands a 5G future might hold. The long-term consequences of this shift in work patterns carries implications for each and every operator.


Cloud gaming driving demand

Data demand was already on the rise before the pandemic; COVID-19 simply accelerated the underlying trend. The main drivers of this demand have been video, gaming and collaboration platforms. Video streaming revenues alone are expected to hit $52 billion globally this year, according to Statista. Asia in particular is experiencing higher CAGR compared to the traditional high-growth markets for the likes of Netflix, Amazon etc. According to Statista, video on demand was given a boost by the pandemic, with Indonesia and India experiencing 18% and 17% increases in demand respectively.

As 5G networks grow over the coming years, growth will come from areas such as cloud gaming. Across Asia, revenues from cloud gaming are expected to top $3 billion by 2023, with smartphone cloud gaming representing the biggest single opportunity in the region. As network operators continue to roll out their 5G networks, balancing higher traffic demands will prove a tough challenge for them. However, it is great news for gamers, who can finally look forward to playing PC and console games easily from their smartphone.

The traditional barriers to entry are lifted with cloud gaming, enabling eager gamers to enjoy the latest offerings without the need for upfront investment in hardware. Even better, cloud gamers are no longer restricted to the limitations of their hardware, instead being able to take advantage of the limitless capacity in the cloud. All this is made possible when there is extremely low latency, which will be the case when 5G deployments are rolled out. Operators will, however, need to weigh up the differing demands in the highly-segmented cloud gaming market, which means understanding the usage trends of (for instance) the increasingly popular e-sports segment, which differ significantly from those of mobile gaming. In Asia, eSports is said to have generated over $13 billion last year.

By 2024, vXchnge says 5G networks will cover 40% of the world. What’s more, a study by the Mobile Video Industry Council found that cloud gaming could be 25% of 5G traffic by 2022. In other words, video and gaming use case are just the beginning. Perhaps the biggest changes will come in the form of the use cases 5G is uniquely primed to support. One such use case is IoT.


Powering the IoT

It’s often repeated that 5G is not just a faster version of 4G, it is a fundamentally different proposition. This is true, and a case in point is the IoT. Unlike 4G, 5G networks are made up of greater numbers of towers with smaller cells that transmit data over a different part of the radio spectrum. Having an increased number of towers in any given area enable 5G networks to support far more IoT devices than on 4G networks. To put things in perspective, it is worth remembering that 4G networks can support a few thousand devices per square mile, while 5G networks can support millions of devices over that same geography. Within these new parameters, large scale industrial IoT networks and the sensor networks required for smart cities begin to look not just possible, but within reach in just a few years.

KPMG estimates that by 2025, investment in automation including AI and machine learning could reach $232 billion. Combined with already existing strong investment in IoT solutions – estimated at $6 trillion in the last five years – operators need to be prepared for an exponential growth in data traffic - several orders of magnitude above what they have experienced so far. The era of IoT and machine-to-machine (M2M) communications involves not millions but billions of devices connecting with each other online.   

According to IDC, Asia will be a frontrunner this year for IoT deployments with over 8 billion IoT and device connections.

In industrial and manufacturing settings for IoT, fast response times with low latency connections are essential. Current 4G networks generally have latency rates between 50-100 milliseconds, but initial 5G rollouts have reduced that to less than 30 milliseconds. Similarly, autonomous cars rely on low latency for their GPS sensors and navigation systems to communicate with the car. 5G improves vehicle-to-vehicle (V2V) as well as vehicle-to-everything (V2E) connectivity and will additionally enable a consistent internet connection for passengers accessing the internet.

However, operators face several hurdles before they can satisfy the global hunger for data with 5G. Firstly, the new radio rollout will initially only have limited coverage. The radio technology used to fully harness 5G speed and latency selling points (mmWave), introduces new challenges to operators, as it requires lower range towers, which in turn increases the cost of deployment. Operators will need to combine mmWave with alternative access technologies such as Sub-6 and LTE for ubiquitous coverage. This scenario will favor continuous access handovers, and will mean almost inevitable latency, bandwidth and congestion will affect 5G’s early deployment phases.


Encryption – we’re seeing the tip of the iceberg

Encrypted traffic has already become a thorn in the side of mobile networks operators. OTT players deploying encryption techniques make it harder to regulate traffic and maintain QoE, which in turn lead to subscriber dissatisfaction. Not only does 5G promise more data, operators are also facing the prospect of new encryption protocols such as eSNI (Encrypted SNI) and DNS obfuscation through DNS over HTTPS (DoH) and DNS over TLS (DoT), some which will become a reality as early as next year. Major OTT players have already signaled their intention to move to DoH and eSNI as soon as these standards are approved. These new techniques threaten the efficacy of traditional traffic management solutions, which won’t be able to identify traffic based on the SNI/domain. It is hard to see a way for operators to manage network congestion in these circumstances, especially given early evidence of the recent news from China blocking any eSNI-based TLS 1.3 traffic in the country.


The future is changing

One way operators will be able get a handle on 5G network congestion is by integrating machine learning (ML) and AI technologies that can intelligently manage data delivery through the RAN to preserve QoE and use costly RAN resources efficiently. AI and ML will play an important part in increasing automation, speed, efficiency and event detecting anomalies, predicting faults and reducing site interventions. Today’s networks lack the capability to dynamically manage RAN automatically, but in future AI will be able to handle capacity allocation at times of peak demand.

Managing – and monetizing – mobile networks has for a long time been a little like a rollercoaster ride for operators. Technical advances, changing usage trends and concerns around privacy mean that mobile operators have had to learn to move fast and be nimble in order to continue to keep customers happy.