Depending on who you speak to, 5G is either humankind’s greatest imminent blessing or its greatest imminent curse. Still in its infancy, and not yet commercially standardized, this technology has already been the most polarizing advancement we have ever seen in communication.
Consumers worldwide are captivated by promises of super-fast download speeds, split-second responsiveness and next-level mobile phone communication, but are divided on the possible sacrifices of privacy and security.
Detractors continue to issue condemnations of 5G cellular’s possible health risks. Supporters continue to shake their heads in disbelief. Governments jostle for geopolitical supremacy; 5G is seen as both a proxy for power, and also a critical infrastructure for the furtherance of national economic interests.
Why has 5G been so controversial and divisive? We never saw anything approximating this level of contention and animosity in the rollouts of 1G, 2G, 3G and 4G. What is so special about the 5th generation of wireless connectivity?
Quite simply, the stakes are higher than ever before.
Though the possibilities that may be realized through 5G are still largely theoretical, even conservative predictions make it clear that this technology will have an unprecedented impact on the way humans live.
Everyone wants a piece of that pie. But it’s not yet clear exactly how they’ll claim it.
Many concerns around 5G are related to this discrepancy in strategy, and the fear that public and private interests are rushing into the rollout of something they do not yet fully understand.
The reality is a little less gung-ho. Outside China, widespread commitment to the establishment of 5G networks has been varied, and in some areas even slow to pick up. This is largely the result of financial tension: the upside of gaining a competitive advantage in this space may be significant, but the capital risks are just as great.
Because 5G is not the same as the Gs that came before. Implementing this technology does not simply require an upgrade of the current network, it demands a new type of network altogether.
Initially, 5G will need to integrate the established operations of 4G, especially LTE, but in order to scale their networks quickly and access 5G’s full potential, operators will need to redefine their network architecture, operations, and services.
5G virtualization is crucial and inevitable.
Without virtualization, 5G will be unable to meet its connectivity requirements. The network will not be able to adapt quickly enough to keep up with the rampant technological changes in ancillary domains. Telcos will not profit from their investments.
A critical evolutionary response
Evolution is not linear. Its course is determined by the emergence of a new need–or the redundancy of an old one–followed by a creative response. Necessity, they say, is the mother of invention. It is this way in the natural world, and the technological.
5G is a natural next step in the evolution of wireless networks because we are moving into a new era of human interaction. We are in the early stages of the fourth industrial revolution (4IR), a redefinition of the way we live and work through the assimilation of the physical and the digital.
Harnessing the power of artificial intelligence (AI) to make decisions based on rapid analysis of vast amounts of information, our technology will soon begin to manage and constantly improve our industrial and commercial systems.
The social face of this dramatic shift is Society 5.0–a vision of the extraordinary advances in human society that will be precipitated by the integration of technologies like big data and AI. By merging cyber, physical, and biological we will see leaps of progress in areas like healthcare, mobility, infrastructure, agriculture and energy.
Two sides of the same evolutionary leap, 4IR and Society 5.0 will rely on the continuous transfer of huge amounts of data generated within a massive internet of things (mIoT).
Video traffic on wireless networks is expected to grow from 56 exabytes globally in 2017 to 240 exabytes globally in 2022. This is not just a lot of YouTube. Public surveillance systems, traffic control systems, industrial management systems–these will all rely on continuous streams of video, a data-dense medium of communication.
As augmented reality (AR) and virtual reality (VR) move towards ubiquity, data volumes will continue to grow at unprecedented rates. Current network structures are simply not up to the task. 5G is not a nice-to-have, it will be essential if we are to realize the potential of our current data development trajectory.
5G is also our best hope of supporting the mIoT, a planetary network of devices, sensors and processors that is expected to grow beyond 75 billion in the next five years.
As part of the performance standards set out in IMT-2020, an international 5G specifications campaign spearheaded by the International Telecommunications Union (ITU), the minimum requirement for 5G’s connection density is 1 million devices per km2. 4G supports about 5% of that density.
Enabling 5G is imperative if we are to facilitate the emergence of Society 5.0, but this technology will not achieve its full capacity on static physical infrastructure. We will need virtualization.
Only by transcending tangible hardware to become cloud-based and software-managed will 5G be able to liberate its whole array of latent benefits: high speed, low latency, lower operational cost, greater energy efficiency, improved scalability, and increased agility.
5G Virtualization, NFV, SDN, and a network slice for every snowflake
Network Virtualization (NV) releases the network from its anchor in hardware and runs a virtual network on top of the physical network. The result is a more dynamic system that can be controlled from a central plane, removing the need for humans to manually configure pieces of hardware.
5G network virtualization will permit the division of hardware resources into functions that can be controlled by software: network functions virtualization (NFV). In network management, NFV seeks to directly optimize network services. The associated network management approach, software-defined networking (SDN), establishes a centralized view of the network by detaching the control and forwarding planes.
As a result of NFV, network resources can be configured and allocated to service the needs of specific customers or service categories, without needing physical adjustments or dedicated infrastructures.
Such a restructuring will pave the way for much-vaunted 5G capacities like network slicing. This architecture introduces the possibility of multiple virtual networks on top of shared physical infrastructure.
Each network slice can be dedicated to specific functions, clients or use cases, delivering elevated service within each segment, and a higher-performing network overall. Network slicing will be the key ingredient in 5G’s ability to support and deliver value from the three ITU-specified generic services with vastly heterogeneous requirements:
- enhanced Mobile Broadband (eMBB)
- Ultra-reliable and Low-latency communications (URLLC)
- massive Machine Type Communications (mMTC)
Within these three areas, we will see the emergence of high-speed mobile applications, driverless cars, and mIoT. But, importantly, these use cases will all have different network requirements such as speed, latency, stability, and security. Network slicing makes it possible to satisfy these needs in a dispersed, yet coordinated and tailored way.
SDN and NFV are used to customize the network offering, supported by automation, service provisioning and orchestration. But the uncoupling of hardware and software does not only facilitate greater network efficacy and efficiency.
It inherently lends itself to a more democratic approach to wireless innovation, promising improved services, better network economics, and shorter times to market for new network vendors.
Seen through the retrospective lens of a 5G-enabled world, the traditional structure of cellular networks borders on archaic. The arrangement favors a few vendors who control the architectural and infrastructural advancement of the ecosystem. The software in these networks is predominantly proprietary and vendor-controlled.
Such a hierarchical and closed structure leads to a handful of players wielding a disproportional influence in network growth. But, more critically perhaps, it stymies innovation, limits agility and slows evolution.
By effectively separating the hardware and software components of the network, and introducing new capacities like virtualization and cloud computing, 5G invites open source development of its various features and assets.
This opportunity is already being embraced by telcos who have aligned themselves to ONAP (the Open Networking Automation Platform), with the intention of improving their operations and business support systems.
The real gains from open source should ultimately be felt by the end user. The limitations created by having a small pool of vendors from whom equipment and networking services can be bought has long been a complaint among operators. By gaining more control over their own infrastructure, operators should be able to optimize the services they provide to their customers.
The spirit of open source is being championed by initiatives like OpenRAN and OpenAirInterfaceTM Software Alliance (OSA), who are focused on improving the quality and efficiency of network development through open source’s democratized, disaggregated, software-based approach.
Such platforms may currently be testbeds, not production-ready environments, but the path towards 5G open source stack is clear.
Though it will have some impact on the radio access network (RAN), open source’s most lasting effect will be related to virtualization, especially in “softwared” network domains and services, such as network slicing, automation, and mobile edge computing.
New opportunities, new threats
5G’s promised benefits are widely discussed, but, understandably, so are its risks. 5G brings into play numerous cybersecurity concerns that do not exist in earlier network generations.
The surge in connectivity broadens the network and provides a larger potential attack surface. And by facilitating a massive internet of things, 5G invites billions more devices to be connected to networks. Most of these devices lack baked-in cybersecurity, or are managed by applications that are themselves poorly secured.
5G networks are also inherently more vulnerable due to their achitecture. Previous network iterations were built on a hardware-based hub-and-spoke model that centralized processes and allowed cyber hygiene to be performed at hardware choke points.
The distributed software-driven routing that defines 5G, however, removes the need for such choke points. This improves speed and efficiency, but it eliminates a very useful security mechanism.
Virtualization has a multi-faceted role to play in 5G.
Firstly, it is part of the problem. Because of their open, flexible, programmable nature, SDN and NFV open up a new range of security threats. A cyber attack that targets the SDN controller, for example, could bring down the entire system.
When network functions are moved into software they become instantly more open to attack–software is by its nature hackable. Furthermore, because virtualized processes will increasingly be managed by AI applications, the wider network becomes more susceptible to damage as a result of those AI operators being hijacked.
However, virtualization should also be part of the cure. According to a recent AT&T Cybersecurity Insights Report: Security at the Speed of 5G, “enterprises will need to take advantage of virtualization to make the network nimbler and more responsive, with the ability to provide just-in-time services.”
5G requires end-to-end security. But, in a virtualized network, virtualized security can be deployed rapidly to almost any network location and automatically respond when new attacks are discovered. Automation is a critical component of this strategy and is made possible through virtualization.
5G’s high speed and low latency, accentuated by virtualized services, will make effective cybersecurity (at least partly) reliant on AI or machine learning for timely threat detection and response.
It is imperative that we get this right. Failure to do so will have possibly devastating results.
One of 5G’s defining features is the potential for the convergence of the cyber and physical realms. Despite its many benefits, this union also spells a new type of danger. Cyber attacks that successfully infiltrate the virtualized world of 5G networks could have very real results in the physical world, where 5G will be used to drive autonomous vehicles or permit remote brain surgery or operate military drones.
As soon as these threats become cyber-kinetic in nature, the danger to human life is elevated.
5G will dramatically alter the way we live, work and play, but its full scope of opportunity is arguably so great that we don’t see it all yet. Regardless of what those diverse use cases eventually turn out to be, we will not get there without reengineering the architecture of the network.
Virtualization is a profound step towards the liberation of 5G’s genuine capacity. By uncoupling from hardware and moving network functions and management into software and the cloud, the network becomes more responsive, agile and open.
Unfortunately, however, this also means more open to attack. As we test and experience more of this quantum leap technology we may be surprised by how we can use it, but we will be equally surprised by how it can be used against us.
The choice is ours. 5G truly could be the network that changes the world, but only if we apply the same amount of innovation to its security as we are applying to its development.
Marin Ivezic is a Cybersecurity & Privacy Partner in PwC Canada focused on risks of emerging technologies.