Virtualized network monitoring and service assurance are essential for operators to compete on QoS for 4G services and beyond.
For operators in northern and southern Asia, cost reduction and quality of service (QoS) from their LTE networks are a priority, albeit for different reasons.
Operators in Asia’s southeastern developing markets – such as Indonesia, Thailand, and Vietnam – are rolling out 4G to deliver broadband internet access nationwide. Meanwhile, operators in northern Asia’s more advanced and mature markets – including Korea, Japan and parts of China – face a different set of LTE network challenges.
They too must manage complex, high speed networks that support a broad range of connected devices, features and services, including VoLTE. But they must also prepare their networks for new and upcoming apps and services such as connected cars, remote home monitoring, and other IoT features. Furthermore, they must develop and put in place methods and practices that will support their imminent move to 4.5G, then 5G.
These challenges occur at a time when the number of mobile subscribers in Asia continues to grow, driven mainly by China and India. According to GSMA Intelligence, the number of subscribers will rise from 2.7 billion at the end of 2016 to 3.1 billion by 2020 – an addition of 460 million new subscribers – a 17% increase over four years
Smartphone uptake also continues to grow, thanks to falling handset prices. GSMA Intelligence predicts smartphone numbers in the region will rise from two billion at the end of 2016, to three billion by 2020. Meanwhile, mobile data traffic in the region is set to grow 8-fold over the next six years [PDF] from 3.7 Exabytes per month currently to more than 30EB in 2022.
However, operators’ revenue growth has slowed, as markets – especially developed ones – become saturated, and increased competition between operators puts pressure on revenues and margins.
We already see this intense competition playing out in India, where the launch of free 4G voice calls and data by greenfield operator Reliance Jio at the end of 2016 ignited a fierce price war. India’s other operators had little choice but to keep pace with Jio’s strategy.
The QoS advantage
In these circumstances, when everybody owns a smartphone, uses the same services, and pays the same price for their LTE data plan, it is QoS and user experience that becomes the key differentiator between operators.
But as quality of service becomes more important than ever, operators are struggling to make their LTE networks work properly under the strain of subscriber and data growth, without having to break the bank and invest in more infrastructure.
They’ve already invested hundreds of millions of dollars in spectrum and infrastructure to build a network capable of bringing high quality mobile voice and data services to their subscribers. Now they have to contend with rising opex, and also figure out how to stem customer churn rather than having to pay extra marketing and customer acquisition costs. This can be an expensive problem: in its 2016 annual report, Thai operator AIS reported customer churn of 12%. When combined with handset subsidies and promotional costs to acquire new customers, the cost totaled over $45 million.
4G networks require a complete rethink of how networks are built, controlled and managed due to the increased need for capacity, and contention for bandwidth between operator-owned and OTT services. As traffic approaches 100% utilization at busy times, these complex networks require automation, deep machine learning, and continuous feedback so that every subscriber, application, and internet-connected device has access to the network at all times.
Unfortunately, without being able to clearly see into the network, and identify which issues are affecting subscribers, operators will struggle to reduce customer churn and coincidentally have difficulty navigating their customer acquisition strategy.
There is an answer to the conundrum of how to improve 4G service and network and service performance without breaking the bank – virtualized network monitoring.
Today, powerful software virtualization solutions running on readily available commercial off-the-shelf (COTS) hardware can deliver 100% faster new cell site and service provisioning ; 25% faster fault identification and reduction: plus a 20% increase in network utilization – all at a tenth of the cost of traditional probe-based solutions.
Case study: Reliance Jio
To take Reliance Jio as an example, virtualized network monitoring has been critical in delivering greater capacity and capability for Jio from its network. By using monitoring to surgically manage the traffic over its network, Reliance Jio could launch attractive new services that simply weren’t previously possible due to bandwidth constraints.
These include mobile broadcast video, music streaming, video on demand services, plus other apps and services, all of which are transforming the Indian people’s access to media, remote learning resources, telehealth and remote commuting.
Jio’s virtualized monitoring resolves issues and optimize network quality and user experience in real-time. Jio can respond much more quickly and efficiently to changing network conditions and unexpected problems. For example, Jio once had to track down the cause of recurring VoLTE outages. Its monitoring system determined that the culprit was faulty router behaviour. Machine learning provided the input for the SDN network to retune itself, and act as “an early warning system” for similar potential future failures.
Case study: SK Telecom
In South Korea, operator SK Telecom has deployed a big data analytics-driven, software defined network over its nationwide network since 2015. The fully automated network uses real-time monitoring data and insight to continuously optimize itself, without human intervention.
As with Jio in India SK Telecom’s real-time monitoring delivers significant improvements in network utilization, and by extension, network monetization. Most operators run their networks at 50% utilization: but thanks to its performance-aware SDN, SK Telecom achieves 90% utilization. This in turn has had a profound impact on its network ROI.
Furthermore, SK Telecom manages a broad range of competing applications on its network, each with very different service needs – from burstability and ultra-low latency to availability, bandwidth intensity and, of course, mobility.
It’s only the detailed visibility provided by real-time monitoring that lets SK Telecom deliver the radically improved and optimized quality of experience necessary for its subscribers to receive the best possible user experience whenever they use their mobile apps and services. Each user effectively benefits from QoE-based network resource that’s allocated to them personally, according to the service they’re using.
From the 4G present to the 5G future
The need for continuous and accurate network performance monitoring is more pressing than ever for operators. Only real-time granular insight and continuous feedback into their networks’ behavior gives operators the full visibility they need for everything from managing increasingly convoluted network traffic patterns to supporting new applications and connections – and anticipating and identifying network problems before they occur.
Today’s 4G networks are sufficiently complex that they cannot rely solely on human control. Virtualized network performance and service assurance are therefore essential for operators to deliver high-quality 4G services to their customers in a way that is cost-efficient, future-proof, and reliable enough for operators to rely on QoS rather than price as their key differentiator.
And this need will only grow more pressing as operators in Asia take their first steps towards 5G. Korea and Japan are already planning to have live 5G networks in place for, the 2018 Winter Olympics in PyeongChang, Korea, and the 2020 Summer Olympics in Tokyo.
5G will bring denser networks, more complex traffic patterns, a hundred times more connections, and more bandwidth-intensive services, together with distributed edge computing and cloud-hosted applications.
More subscribers, more data, more connections, more services – the need for real-time network monitoring has never been greater, both now and for the future.