Amidst growing network complexity, telecom networks are becoming increasingly susceptible to disruptions of all kinds. This is a serious concern as network troubles can be detrimental to enterprises.
The most common network issues include network congestion caused by unbridled data traffic, link flapping (that is, a network’s oscillation between being connected and disconnected), faulty hardware such as routers and firewalls, software hiccups like firmware bugs and protocol mismatch, and cybersecurity threats. Self-healing networks (SHNs) are emerging as an innovative solution to address these challenges.
SHNs can identify and rectify faults without any human intervention. They have evolved over time from automated fault diagnosis and recovery to artificial intelligence (AI) and machine learning (ML) algorithms, network functions virtualisation (NFV) and software-defined networking (SDN). NFV facilitates dynamic reconfiguration of network resources to address issues while SDNs equip SHNs with a centralised control layer that can regulate network configurations based on real-time insights. These intelligent zero-touch infrastructure solutions illustrate phenomenal adaptability and resilience.
SHNs provide the insights necessary for autonomous monitoring, analysis and problem-solving. They also learn and evolve from each incident and can pre-empt catastrophic problems and offer remedial measures by assimilating historical and real-time telemetry data on device health and network performance from various sources such as routers and firewalls.This data is used to identify performance metrics such as latency and bandwidth utilisation. A digital twin modelling the network is then generated to test changes/remedies before they are applied to the actual physical network. These remedies could range from network traffic rerouting, reconfiguration and traffic filtering to detecting equipment degradation, spotting unusual activity and handling cyberattacks.
SHNs can also help organisations by running business operations smoothly, bringing down costs and reducing their IT downtime, thereby allowing IT employees to focus on more high-value chores. According to an industry study, SHNs are 75 per cent faster than traditional network systems in terms of time taken for ransomware detection and have a 90 per cent data recovery success rate.
Content service providers (CSPs) around the world are increasingly venturing into SHNs, though there are regional differences with regard to the level of network autonomy. A report by Capgemini states that currently, European telcos, particularly France with its 8 per cent Level 4 automation, are the front runners in autonomous network mobility driven by the need to reduce costs.
It is followed by their North American and Asian counterparts, which are at Level 2 automation (23 per cent and 25 per cent respectively) and Level 3 automation (14 per cent and 13 per cent respectively). Within Asia, Japan and India stand out in terms of their Level 2 automation (50 per cent and 33 per cent respectively) and Level 3 automation (25 per cent and 17 per cent respectively).
Notably, the Philippines’ leading telecommunications provider, Globe Telecom, announced that it is investing heavily in infrastructure capabilities to make room for Level 4 autonomous networks. This will enable the CSP to enhance network stability, use AI for proactive incident detection and prevention, and enhance customer experience.
Other players like the MTN Group, which is among the biggest telecom operator groups in Africa, are leveraging SHNs for network assets and platforms. The company aspires to scale network automation to Level 4 by 2025. To this end, MTN has successfully tested the sleeping cells self-healing solution (which allows for automated surveillance, discovery, analysis and recovery of sleeping cells) and the intelligent internet protocol private line solution (to develop a new cloud network architecture).
Another interesting development is that New Zealand’s largest telecommunications and digital services company, Spark NZ, has rolled out the second generation of the optical transport network. Functions like real-time analytics preventively identify faults such as those caused by weather-related calamities. The solution also enables troubleshooting through optical restoration and redirects network traffic.
Malaysia’s Digital Nasional Berhad (DNB) too is also not far behind in this race. DNB caters to six telecom service providers by adopting closed-loop (that is, real-time monitoring of network conditions) AI-based automated and data-driven processes to efficiently manage these complex multi vendor diverse-technology networks.
To add to this, earlier this year, DNB partnered with the Swedish multinational networking and telecommunications company Ericsson to explore the use of intent-based operations. This unlocked avenues for DNB to boost profitability and open routes to differentiated services with discrete slices without compromising on quality.
MEO (formerly Portugal Telecom) also recently decided to deploy SON for optimisation and network assurance.
Closer to home, Bharti Airtel and Reliance Jio are also deploying SONs to enable their customers to enjoy world-class high speed data and voice experience. Further, in May, Airtel entered into a strategic alliance to expedite cloud adoption, network modernisation and generative AI integration for its customers. A suite of solutions will be available to Airtel, including geospatial and voice analytics and marketing technology solutions.
Meanwhile, Vodafone Idea (Vi) is also planning to launch the IoT Smart Central platform to promote seamless integration and autonomous management of internet of things (IoT) devices. The platform offers various services, including self-care digital experiences, life cycle management for SIMs, real-time visibility of IoT devices, and end-customer real-time charging, billing and invoicing.
Hurdles to adoption
Despite their numerous benefits, SHNs present their own set of challenges. To begin with, varied IT environments may have a set of traditional and novel technologies. This can complicate the deployment and functioning of SHNs. Therefore, specialised expertise and investment is needed for a tailored approach to network design and configuration to ensure that SHNs are compatible with existing networks.
A key challenge with SHNs is that, amidst the rising number of apps being deployed by organisations, gathering and maintaining data at a granular level becomes increasingly difficult. This data may become fragmented when multiple vendors are involved, leading to inaccuracies. Therefore, human intervention may be required once in a while to check if SHNs are aligned with business goals. Further, network engineers may need to club “captured intelligence” with “local observation” to fix more complicated problems.
Another challenge related to vendors is the issue of “lock in”, as the connectors between the management software and the network still revolve around specific vendors. This is a major roadblock to standardising application programming interfaces to manage the core network.
At the same time, it is crucial that SHNs advance with the evolution of network challenges. IoT has enabled the interconnectivity of devices, which come with their own set of vulnerabilities.
For instance, IoT devices are susceptible to cyberattacks due to weak passwords, unsafe data transfer and storage, outdated software and improper privacy safeguards. The breach of sensitive customer/employee information can jeopardise an organisation’s brand image. Achieving the right balance between protecting sensitive information while the network carries out other autonomous functions can be tricky but is crucial to preventing data leaks and maintaining privacy.
The road ahead
In an era of organisational IT complexity, the need for hyperconnectivity, the surge in data traffic and the mass adoption of 5G will make SHNs indispensable.
A survey projects the upward trajectory of the SHN market from the current $0.75 billion to $13.7 billion in 2032, registering a CAGR of 33.7 per cent. It expects the US and Canada to be the main drivers of this growth because they are home to vibrant research and development ecosystems and top network healing firms.
Moreover, the need to diminish network costs is expected to spur investments in AI by CSPs. According to a study, performance optimisation and network security will constitute more than 50 per cent of global operator expenditures on AI by 2028. Additionally, worldwide operator spending on AI for network orchestration will generate $20 billion by 2028. This is a remarkable 240 per cent rise from $6 billion currently.
Lastly, with self-healing capabilities expected to undergo further advancements by the time 6G is rolled out in 2030, SHNs will play a critical role in supporting 6G.
6G’s features are likely to be an extension of 5G, particularly in terms of AI-related capabilities. Sustained advancements in AI and ML algorithms will enhance the precision and efficacy of SHNs, leading to improved data insights and adaptability. This will result in the evolution of more sophisticated models, greater automation and the resolution of more complex issues.
