Tuesday, December 6, 2022

5G Networks and Beyond: Hype versus Reality?

The various generations (Gs) of digital cellular networks have roughly followed a 10-year development cycle. While each previous G was mainly associated with one killer application, (i.e., 2G (voice), 3G (mobile browsing) and 4G (mobile broadband)), 5G was designed from the outset to support various requirements, including enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC) and massive machine-type communications (mMTC).

5G Networks: Open Platforms for Vertical Industries

To simultaneously support a wide range of heterogeneous requirements, 5G has been made flexible at several levels, including the radio interface (e.g., numerology and flexible slot length), architecture (e.g., network function disaggregation and functional splits), deployment options (i.e., private, public or hybrid), spectrum bands (i.e., low, mid, and high bands) and spectrum licensing regimes (e.g., local licenses evolving to dynamic spectrum access). This has effectively made 5G an open platform that can be tailored to virtually support the needs of any use case across various vertical industries (e.g., Automotive, Ground/Marine/Aerial/Space Transport, Factories-of-the-Future, Healthcare and Public Safety). In this context, network slicing has emerged as a new paradigm to create a set of customised logical network instances on top of the same physical infrastructure. With slicing, the network becomes become like a pizza base, on top of which slices with different toppings can be created to satisfy different users.

Reality of 5G Networks

Despite these large potentials, some stakeholders are suspecting there has been lots of speculation about the potential of 5G to support vertical industries, particularly in the current situation where most of live 5G networks are still relying on their 4G counterparts in non-standalone (NSA) deployments that cannot offer the most advanced 5G features (e.g., short latency, support for massive numbers of devices and fully programmability). In my opinion, there may be lack of understanding of the capabilities of commercially available 5G networks, but there is no doubt that 5G is surpassing its preceding technologies. At the same time, I believe that, for each vertical industry, the value propositions of 5G should be rigorously assessed, considering all stakeholders, identifying the pain points of existing solutions, and highlighting the contribution of 5G in overcoming them.

Beyond The Hype: Will 6G Overcome the Fundamental Limits of 5G?

Beyond 5G and its on-going enhancements (i.e., aka 5G Advanced), there has been lots of effort to shape/influence the vision of future 6G networks. While I agree with some of the identified trends (e.g., integration of non-terrestrial networks (NTN) and more active role of Artificial Intelligence/Machine Learning (AI/ML)), I believe it is too early to set concrete values for the target 6G key performance indicator (KPIs) (e.g., 1000x increase in achievable bitrates), or even to decide about the core 6G technologies. For instance, there have been recently lots of hype about specific technologies (e.g., THz and reconfigurable intelligence surfaces (RIS)) with some folks considering them as the next big thing, while it is well understood that most of these will be useful only in very particular scenarios.

At this point in time, I believe we should not focus on trying to push the envelope of what 5G can achieve, but rather on solving the fundamental limits of 5G and its preceding generations. The first key point is that, moving forward, we should stop focusing on the achievable performance in the most favourable conditions, but rather on the minimum guaranteed performance. In other words, 6G should strive to connect the unconnected/under-connected (e.g., via NTN integration) and contribute to closing the digital divide. The second key area is sustainability (e.g., energy consumption, electromagnetic fields (EMF) radiation, and social impact), which should be one of 6G design requirements as opposed to being an after-thought as it was the case in previous Gs. The third area of improvement is the business viability of the developed technological solutions. Instead of first building fancy technologies and then looking for their suitable usage, the needs for innovation should emanate from actual use cases that are yet to be identified. To achieve that, vertical industries should be onboarded early in the standards (i.e., third-generation partnership project (3GPP)) development process to let them shape 5G from the beginning and have more influence on the specifications roadmap. This can build upon the contribution of some industries to the 5G standards, including the 5G Automotive Association (5GAA) and 5G Alliance for Connected Industries and Automation (5G-ACIA) for automotive and industry 4.0, respectively.

Longer Term Vision

On the long run, 6G networks and beyond should take the flexibility of 5G and its enhancements to the next level. Further adoption of openness at all fronts (e.g., disaggregation of components through open interfaces, open white-box hardware, and open-source software) will be key. Major improvements in network slicing will be needed to facilitate end-to-end orchestration across all possible technological and administrative domains. All the practical aspects (e.g., isolation, scalability, and security) associated with the design, deployment, and operation of network slices should be dealt with efficiently. With the increased complexity of the network and its growing dependency on software, future networks are expected to be automatically managed through an agile methodology relying on development and operations (DevOps) tools and continuous integration/continuous delivery/deployment (CI/CD) pipelines, where incremental changes made by developers can be compiled and packaged into feature artifacts. This will effectively eliminate the 10-year development cycle of telecom’s Gs and will thus significantly speed up innovation. However, the recent progress made on the emerging Open RAN paradigm indicates that such approach would require a mindset change across the telecom ecosystem, where all stakeholders should work together to reap the full benefits of future agile networks. This would obviously take some time before becoming mainstream.