Keywords

1 Introduction

In the post-pandemic era, the transformation of human experience that links the physical, digital, and biological worlds is likely to bring about novel possibilities as well as serious challenges. In this context it is often claimed that advanced features and new communication architectures provided by the sixth generation (6G) network technology will help realizing the vision of a “connected globe” and support growth, sustainability, and the idea of complete digital access. However, in order for this optimistic vision to be realized an ambitious 6G strategy is needed [1]. Such strategy should allow realization of scenarios in which 6G technology dramatically boosts and enhances human potential and capacities, going beyond digitalization of the 2020s [2]. Here, high quality digital (virtual) worlds that are disconnected from geographical locations may attract people, delivering new experiences. Bearing this in mind, it also becomes clear that processing needs to be performed closer to the data sources (often smart devices), in an effort to minimise latency, save bandwidth, improve security, guarantee privacy and increase autonomy leveraging the IoT-edge-cloud continuum of computation. For instance, when next generation of devices with user-friendly interfaces, and new sensing capabilities will materialize, humans commanding their “avatars”, will be able to activate different aspects of the physical world. Here, the interconnection will be provided/realized through a portfolio of biological and mechanical sensing devices accessible through futuristic human-computer interfaces. The 6G techniques bring also promise for the next generation of industrial applications, as they are to offer substantial performance boost and new dimensions of interactions between physical, biological and digital environments. Recent articles [3,4,5,6,7], discuss the relevant 6G vision(s) and technologies. However, one has to keep in mind that deep-dive analysis must wait till about 2030, when the start of actual adoption of 6G is to materialize. This assumption is based on the fact that current duration of each network generation is around ten years. In this context, outline presented here should be seen as an attempt to outline potential of 6G infrastructures for future Edge-Cloud deployments more than hard-facts based prediction.

In this context, the birds-eye view allows one to state that while there will be an increase in demand for mobile broadband from both consumers and businesses, the actual adoption of 6G-based ultra-reliable and low latency connections will primarily be driven by specialized and local use cases in conjunction with private networks. Moreover, such “local deployments” will frequently involve support for various forms of artificial intelligence. Here, automobiles, industrial machinery, appliances, watches, and clothing are just a few examples of objects that will use 6G networks to automatically adapt behaviour, environment, and business operations, in order to learn and organise themselves to meet demands of (domain) specific use cases. Another important design consideration for 6G is going to be energy efficiency, because the amount of energy that can be found in each architectural domain will determine how well the network performs.

1.1 Historical Perspective

Since the first appearance of analogue communications networks in the 1980s, mobile communication networks have developed phenomenally. The progress has been made in subsequent generations, each with its own standards, capabilities, and methodologies. Nearly every ten years, a new generation has been introduced [8]. Figure 1 depicts the evolution of the mobile network.

Fig. 1.
figure 1

Evolution of mobile networks [8]

Full size image

(1–3)G Networks. The first generation mobile network, which had a data rate of up to 2.4 kbps and was built for voice services, was initially introduced in the 1980s. Due to the analogue signal, utilised for transmission, and the lack of a global wireless standard, there were numerous limitations, including difficult hand-offs, ineffective transmission, and no security [9]. The second-generation systems, sometimes referred to as 2G systems, were constructed on digitally modulated techniques such Time Division Multiple Access (TDMA) and Code Division Multiple Access (CDMA). They supported enhanced audio services as well as services like Short Message Service (SMS) and had a data rate of up to 64 kbps. Throughout the 2G period, the GSM (Global System for Mobile Communication) was the dominant mobile communication system [10]. The third generation was introduced in 2000, with the intention of facilitating high-speed data transport. High-speed Internet access and data transmission rates of at least 2 Mbps were provided by the 3G network [11]. This enabled the use of services like Web browsing, TV streaming, navigational mapping, and video services that were not supported by 1G/2G networks. During this time frame, the Third Generation Partnership Project (3GPP) was created to conceptualize the technical requirements and carry on the process of developing mobile standards and technologies in order to achieve global roaming [12].

4G Networks. The 4G network infrastructure, which has been launched in the late 2000s, was an all-IP network capable of transmitting data at speeds of up to 1 Gbps in the download, and 500 Mbps in the upload. In order to meet the requirements of applications like video chat, High Definition TV, and Digital Video Broadcasting, it boosted spectral efficiency and minimised latency. Additionally, 4G enabled terminal mobility to offer wireless services, whenever and wherever they were needed, through autonomous roaming across wireless network boundaries. Long Term Evolution-Advanced (LTE-A) and Wireless Interoperability for Microwave Access are examples of the 4G technology (WiMAX) [13].

5G Networks. The development of hardware facilities, initial basic testing, and the standardization process for the fifth generation mobile communication network are virtually complete, and it will soon be widely used commercially. With 5G, it is expected that ground-breaking improvements in data speeds, latency, network dependability, energy efficiency, and widespread connection will materialize [14, 15]. Here, 5G networks should significantly boosts data rates by using the millimeter-wave band for the first time as well as the new microwave band’s spectrum (3.3–4.2 GHz) (up to 10 Gbps). Beam Division Multiple Access (BDMA) and Filter Bank Multi-Carrier (FBMC) are two examples of advanced access technologies used in 5G. Moreover, massive MIMO for expansion projects, Software Defined Networks (SDN) for effectiveness, Device-to-Device (D2D) for transmission rate, Information Centric Networking (ICN) for traffic decrease, and Network Slicing for rapid deployment of different services are a few of the cutting-edge innovations that 5G combines to improve the performance [16]. Enhanced mobile broadband (eMBB), Ultra-reliable and low latency communications (URLLC), and mMTC (Massive Machine-Type Communications) were the three main 5G utilisation scenarios that have been suggested and are still considered to be valid (and waiting to be practically validated).

2 Architecture of 6G Networks – Current Vision

As 5G networking moved into the commercial deployment stage, research institutions started to focus on theoretical foundations and early-stage laboratory-based experiments with 6G networking. In 6G network, it is postulated that the peak data rates of 1 Tbps and exceptionally low latency measured in microseconds can be delivered. This means that 6G is expected to seriously enhance the performance of data transmission. Specifically, it is to offer up to a thousand times greater bandwidth than 5G technology, because of terahertz frequency transmission and spatial multiplexing. Moreover, by combining satellite network connectivity with underwater communications, 6G aims to provide ubiquitous connectivity and global coverage [17]. To illustrate the progress of communication capabilities of 4G, 5G and expected 6G networks, Table 1 summarizes their main characteristics.

Table 1. Performance evaluation of 4G, 5G, and 6G [18]
Full size table

In addition to a safe and automated coordination design, the 6G architecture comprises building blocks that cross key architectural areas of a telecommunications network, beginning at the physical layer and going all the way up to the service layer (see, [19]). As shown in Fig. 2, the main building components for the 6G architecture have already been defined. The Nokia Bell Labs 6G architectural breakdown consists of four key interconnected components that offer an open and distributed reference architecture. The “Net-Cloud” component of the 6G architectural cloud shift includes notions such an information circulation run-time architecture, accessible, flexible, and agnostic hardware-accelerated, and more. It effectively acts as the architecture’s infrastructure platform. The functional part of the architecture is covered by the “functions” component, which also covers RAN-CORE converge, cell free and mesh connection, link building, and AI. The development of specialized networks and related functional characteristics is a key transformational aspect of the 6G-based ecosystems. Devices that will join 6G networks are becoming ever smarter in collecting, processing and transmitting data, posing new challenges, thus it will be key for 6G implementations to achieve increasingly demanding levels of reconfigurability, self-* and automation, in order to scale efficiently, manage resources, and optimise operation while handling multi-vertical traffic with distinct demands. Consequently, architecture elements that affect performance, such as adjustable offload, extreme chopping and subnetting, are presented as parts of the “specialized” core component. For a business to succeed, the “automation” component of the change to the 6G architecture, which will offer an open service allowing for environment engagement, domain resource monetizing, in addition to cognitively closed system and mechanisation, is essential.

Fig. 2.
figure 2

Architecture of 6G network [19]

Full size image

2.1 B5G Networks

Before moving to the visions of advances that are to be brought by 6G networking let us focus, for a moment, on the time when movement towards 6G will materialize. Here, as a metaphor one can consider 5G networks realized “on top of 4G”. This is known as “beyond 5G” (B5G) and just a few of the B5G applications include robotic communications, e-health, human-robot interaction, multimedia, e-learning, public security, e-governance, logistics and transportation manufacturing, and transportation technologies. In order to enable Internet of Things, wireless communication technologies are essential. They refer to environmentally responsible, energy-efficient, sustainable, and intelligent technologies [20]. Future commercial and non-terrestrial wireless communications networks will need multi-beam antennas as essential components. These antennas’ many beams will make it possible for different terrestrial, airborne, and space-based network nodes to dynamically connect with one another. When the operating frequency for sixth-generation (6G) and beyond 5G (B5G) systems increases to the high millimetre wave (mmWave) and terahertz (THz) bands, relatively few techniques are anticipated to dominate in the design of high gain multi-beam antennas [21].

3 Selected Application Areas of 6G Networking

Let us now “move forward” and consider where 6G networking could be useful. Here, only few areas of particular interest to our research have been selected. Let us start from the reflection concerning changes of the leading paradigm for development and deployment of large distributed ecosystems.

3.1 Towards Edge-Cloud Continuum Computing

The beginning of 21st century has been characterized by a general trend that can be summarized as: “Business solutions should migrate to the cloud”. It is only recently, when the Internet of Things (IoT) ecosystems started to materialize, when it became obvious that cloud-centric solutions cannot support fast-growing sizes of IoT deployments. The unprecedented data explosion and the evolving capabilities of virtual infrastructures, set the scene for developing a new paradigm for data and compute resource management. Here, there are multiple aspects of the IoT ecosystems that do not fit cloud-centric thinking. Among them are (and these are just few examples): (a) amount of data generated by the sensors is so large, that networking infrastructure between the sensors and the Cloud may not be able to efficiently transfer them, (b) between the sensors and the cloud, within the ecosystem, multiple computing nodes with reasonable capabilities exist and the cloud-centric model does not support their utilization, (c) large IoT ecosystems may require extremely time-constraint decision loops, which cannot be realized in cloud-centric deployments, (d) privacy and security of travelling data is put at stake.

In response to these issues, a novel approach to performing computational tasks, based on the concept of the edge-cloud continuum has been proposed. It advocates consuming and processing data along the whole “architectural spectrum” – from the edge devices up to the cloud. In this approach, computing and/or storage “nodes” can be located anywhere across the network (ecosystem), offering a “computational fabric” spanning (any fragment of) the path from (constrained) devices to the powerful cloud systems. Devices will be able to unload their activities and offer (possibly pre-processed) data to the nodes constituting the distributed infrastructure of the “fabric.” The decentralized computing model, as compared to the centralised one, provided by the cloud alone, is a key advantage of the edge-cloud continuum approach. Additionally, as data and tasks may be processed and evaluated close to their “point of origin”, latency is reduced. Furthermore, the edge-cloud continuum approach aims at increasing the network capacity and the frequency spectrum use. The highly distributed node placement improves computation and storage capacities “in places where they are needed”, which may also have a beneficial effect on system reliability [22]. However, the challenge of seamlessly integrating various edge technologies into a homogeneous “continuum” remains open, since current centralised deployments store and process long-term data, relying on the cloud, but lack capabilities needed to handle key open challenges, such as (i) cloud-centricity by default in most legacy in-place systems, (ii) latency, (iii) cost, (iii) network congestion, (iv) heterogeneous smart devices - posing heterogeneous access concerns- and (v) the lack of a security and privacy reference implementation considering CI/CD and DevOps procedures.

Here, it is relatively easy to realize that support for the edge-cloud continuum based computational model has a potential to become a major application of the massive data handling capabilities that are to be provided by the 6G technology. Given the high number of devices exchanging data and (concurrently) running large number of applications, the computing and storage services provided by the nodes placed within the edge-cloud continuum will be crucial; and thus becomes a potential 6G deployment-driving use case. Additionally, 6G infrastructure will offer a number of useful features like geo-distributed data, for example. The 6G communications will enable interconnectivity of the nodes and allow for achieving an extremely low latency and efficient bandwidth usage [23]. Here, it is worthy realizing that large scale edge-cloud continuum ecosystems will not require 6G infrastructure “everywhere”. Rather, it will be possible to obtain expected performance characteristics by deploying 6G infrastructures in selected “regions” (parts of) the ecosystem. Worth noting, the decision on whether or not to apply 6G features to a specific vertical (or continuum deployment) will rely on the collaboration of such networks with the governance implemented by the next generation meta-operating system that will orchestrate the IoT-edge-cloud continuum (some examples are appearing as European-funded initiatives in the meta-OS program of the EC, where the project aerOS promises to be a reference [24].

The ever increasing number of edge devices in use requires the provision of a stable communications platform that can exchange large amounts of data with low latency [25, 26]. Through a 6G wireless communication network, the edge-cloud continuum can achieve its full potential and enable a variety of physical and virtual nodes to exchange information in a very efficient way. Effective methods to increase the spectrum and energy efficiency include simultaneous wireless information and power transfer (SWIPT) and cooperative spectrum sharing.

With significantly increased data rates and spectral efficiency, 6G wireless communication and networks will continue to shift to higher frequency and larger bandwidth. Given the diversity and density of the edge-cloud continuum deployments, it may be necessary to expand the 6G wireless network to support modern random access (RA) for applications distributed across the “continuum”. In such a distributed data and compute scenario, the so-called network compute fabric, the network should host computing intertwined with communication for the highest level of efficiency, to support heterogeneous systems, ranging from simple terminals to performance-sensitive robots and augmented reality nodes. This can be done by designing smart protocols and using signal processing and communication technologies. A good possibility is provided by contemporary RA techniques such massive multiple-input multiple-output (MIMO), OFDMA, nonorthogonal multiple access (NOMA), sparse signal processing, or innovative orthogonal design methodologies. Grant-free transmission for distributed architecture should be planned for a successful implementation of 6G-based computational continuum, where applications frequently take part in self-organizing decision-making.

Nevertheless, it should be also realized that the implementation of the paradigm of “computing continuum”, based on capabilities of 6G networking, will not be without its problems. In particular, difficulties involving resource allocation, work offloading, energy efficiency, delay minimization, justice, and security will emerge [27]. They will be further magnified as the deployments will become part of large-scale edge-cloud ecosystems. Here, 6G “patches” will have to be managed not only taking into account their “internal needs”, but also to “understand the context” of their role within the ecosystem. For instance, while internally (within the 6G patch) some energy saving measures may seem appropriate, they may be prohibited by the needs of a workflow that is being realized to deliver user-defined service. Hence, 6G in edge-cloud continuum may need be reactive to the dynamically changing, context of the “outside world”.

3.2 Artificial Intelligence in Edge-Cloud Ecosystems

It is easy to observe that mobile applications, driven by artificial intelligence, are being developed as a result of the widespread use of smart mobile devices. In multiple application areas, such as computer vision, natural language processing, and autonomous driving, AI has made remarkable successes. AI activities require a lot of computing and are typically trained, developed, and used in data centers with specialized servers. However, with growing power of mobile devices, significant number of intelligent applications are anticipated to be implemented at the edges of cellular connections [28]. AI at the edge of the network promises to be beneficial not just at functional but also at business level, allowing the realisation of federated/distributed AI scenarios and adjusting to the capabilities of the continuum applying techniques such as frugal AI. To support “intelligent applications” at different edge handheld devices with extremely heterogeneous communication, compute, hardware, and energy resources, the 6G wireless network may deliver the needed infrastructure [29]. Here, the 6G network is expected to provide adaptable platforms to enable ubiquitous mobile AI services. In addition, it will provide a thorough method for optimizing telecommunication, computing, and storage infrastructure to accommodate modern AI functions across end-devices, networking boundaries, and cloud data centres [30].

Nevertheless, it is easy to see that the extreme heterogeneity of devices “roaming within the 6G network infrastructure” will, again bring the need for edge-cloud continuum thinking. While some devices may be able to get involved directly into AI/ML activities, such as mode training or use, others may “need help” (provided by devices that are somewhere within the continuum). It is at this point where novel approaches (such as the one brought by project aerOS [24]) will come into place, materializing federated AI in a dynamic and smart way (adapting the sharing of models/data depending on the needs and network and node capabilities). This being the case, it is easy to realize that the 6G networking by itself, regardless how fast it becomes will not be enough to facilitate deployment of intelligence across the continuum.

3.3 Distributed Ledger Technologies and 6G Networking

Globally, multiple business and research groups have adopted distributed ledger technology (DLT; [31]). In what follows we will, however, use the term blockchain to represent all DLT solutions. One of key advantages of blockchain is decentralisation, which eliminates the need for middlemen and “trusted-centralised” third parties; another is transparency with anonymity; a third is provenance and non-repudiation of the transactions made; a fourth is immutability and tamper-proofing of the distributed ledger’s content; a fifth is the removal of single points of failure (improving adaptability and resistance to attacks like DDoS).

Given that blockchain technology has been designated as one of the key enabling technologies for 6G mobile networks, it is essential to look into the various benefits, opportunities, and difficulties related to its application [32, 33].

The 6G vision includes a wide range of applications that can be made possible or enhanced by using blockchain. The idea behind using blockchain, to provide or enhance these applications in 6G, comes from the possibilities stemming from their key characteristics, namely decentralisation, transparency, immutability, availability, and security [34].

Moreover, the 6G infrastructure itself offers a wealth of application potential for utilising blockchain to improve performance or enable new services/use-cases. Particularly, Service Level Agreement (SLA) Management, Authentication, Authorization and Accounting (AAA), Decentralized Network Management Structures, Network Service Pricing, Billing and Charging, and Spectrum Sharing. Here, it is worthy noting that majority of these use cases naturally belong to the IoT deployments. Hence, they are also going to find their way into edge-cloud continuum deployments.

3.4 Industry 4.0 and Beyond

The use of the expected 6G capabilities will be significantly influenced by the industrial applications in 6G. Industrial environments adapt themselves particularly well to the problems and important characteristics of blockchain technology. Holographic communication, for instance, require decentralised systems that are also trustworthy in order to support industrial use-cases like remote maintenance or widespread networking of industrial manufacturing equipment [35]. From another perspective, 6G and the realization of an actual IoT-edge-cloud orchestrated continuum will allow the creation of highly flexible, sustainable (green) modular digital production lines and manufacturing of a new product in a low-volume production (high customisation). The previous will unleash implementation of smart rapid response features in connection with self-optimisation, re-configurations ramp-up, adaptation of the production line and the operations, driving manufacturing closer to Industry 5.0 principles [36].

3.5 Healthcare Applications

The number of chronic patients will rise in the future, and the medical resources needed to give them the care they will be in short supply. These issues will have to be addressed by healthcare systems [37, 38]. Adoption of ICT is essential for the upcoming health services. The ability to measure some health indicators in real-time and assess the patient’s overall condition, along with the automatic sharing of the collected clinical data among all members of the care team, is vital for the personalised management of the patient’s health. This is especially true for patients who have multiple chronic diseases. 6G can alter the future wireless healthcare system which will allow the implementation of new healthcare services and essentially change the current telecare paradigm. The specification of 6G future communication application scenarios and the associated enabling communications technologies is the main are of the ongoing work and the 6G vision, which also includes significant drivers, research requirements, obstacles, and critical research problems [39].

To address enduring challenges in 5G networks, smart healthcare in 6G will need to develop. Here, it is worthy noting that future healthcare systems may be based on blockchain technologies. Here, the privacy concern is the next in line for these technical issues. Furthermore, the immutability feature offered by blockchains makes it possible to maintain the integrity of healthcare data. Blockchains particularly can provide user-controlled privacy and safe data storage without the need for a centrally trusted third party. Here, as noted above, there is a direct applicability of 6G networks and blockchains.

3.6 Environment Protection and Monitoring

Environmental sensing applications that are decentralised and cooperative may need blockchains to store trusted data, and they may be possible by 6G on a global scale. These capabilities can be put to use in applications like smart cities, transportation, and environmental preservation for the green economy.

In summary, it can be clearly seen that there is a direct connection between the (1) vision and theory underlying future development of 6G based networks, (2) multiple application areas and tools that need them, and (3) edge-cloud continuum that will be used to realize practical use cases, delivering actual (and needed) services to the users.

4 Security of 6G Networks

Many innovations and developments in terms of architecture, applications, technologies, regulations, and standardization are being researched to be included in the 6G network concept. Similar to generic 6G vision, which adds intelligence to 5G networks’ cloud infrastructure and software, 6G security vision closely integrates AI to produce security automation (Fig. 3). Here, one has to realize that the “adversaries” also work non-stop to produce new kinds of security risks. For instance, identifying zero-day attacks is never easy, but stopping them from spreading is the most practical defense. Therefore, integrating intelligent and adaptable security methods will be more crucial than ever for predicting, detecting, reducing, and preventing security threats as well as restricting the spread of such vulnerabilities in the 6G networks. Ensuring trust and privacy in the relevant areas and among participants is also important. Particularly, security, privacy and trust are themes that are closely related to one another, with security relating to the protection of the data itself, privacy assuring the hiding of the identities associated with that data, and trust bringing meta level reflection influencing interactions between (semi-)intelligent entities. While privacy and security are mutually exclusive, the reverse is not true: in order to protect privacy, there should always be security measures in place to protect data. At the same time, trust goes beyond privacy and security. Here, it should be noted that an entity can be secure and may preserve privacy. However, in the context of large-scale IoT ecosystem it may not be trusted that it will deliver required service within the contracted time (see, also [40]).

Fig. 3.
figure 3

Security vision of 6G network [32]

Full size image

To be able to assess the value of 6G deployments (including security-related aspects), Key Performance Indicators (KPIs) and Key Value Indicators (KVIs) are needed to properly account for the aspects of the impact that are outside the reach of deterministic performance measures, and will help determine the limits of 6G [32]. It is anticipated that 6G systems will include unique components such embedded devices, local computer and storage, integrated sensing, and artificial intelligence [41]. These aspects will necessitate development of new KPIs and KVIs, such as, for instance, slightly more accurate, computation round-trip period, and AI model time complexity, as well as improvements to existing KPIs. This being the case, the UN Sustainable Development Goals’ principles of sustainability, security, inclusivity, and reliability may be used to measure the value of future 6G-related technologies [42]. Therefore, it is anticipated that the new features, brought by 6G network-based systems (e.g. edge-cloud ecosystems), will significantly affect also how security KPIs are developed and assessed. A set of KPI where proposed in [43] were: level of protection, time to respond, coverage, autonomicity level, AI robustness, Secure AI model convergence time, security function chain round-trip-time, deployment cost for security functions. Moreover, several factors, including PLS, network information security, and security-related to AI/ML, should be taken into account when characterizing security [44].

5 Concluding Remarks

The successful launch of 6G architectural research marks the beginning of the path towards the 6G connected worlds. At the same time, commercial 5G installations are currently underway, or will start soon in the majority of markets worldwide. The architectural conceptualization of 6G is still ongoing and is expected to continue for at least another eight years. Beyond that, the future holds enormous potential for a fundamentally new human experience due to real-time connectivity and synchronisation across the physical, digital, and biological worlds. Physical can be broadly referred to as physical and analysis focus from the perspective of a network, digital is all about upcoming software architecture and sophisticated automated representation, and biological includes unique human-machine interfaces and biosensors. To accomplish the adaptability, simplicity, dependability, security, efficiency, and automation required to fulfil the multiplicity of future applications, they will all be closely interconnected. Analysis performed on the basis of current knowledge indicates that the implementation of 5G and, in the future, 6G technology on a massive scale will result in a revolution, with a large number of leapfrogging changes in the area of engineering fundamentals of entire information systems. The currently dominant microservices architecture, requiring highly interconnected functional components [45], will be an additional impetus forcing faster mass deployment of 6G networks.

At the same time, cloud-centric solutions will start to be replaced by edge-cloud deployments. The so-called computing continuum is an active field of research that aims at governing nodes spanning across IoT to edge to cloud, forming a fabric that will be closely tied to the performance (and further capabilities of 6G) networks improving features like latency and availability. This process will be particularly important for large-scale, highly heterogeneous ecosystems. Here, to be able to deliver services to the users, 6G based networks will be needed. However, 6G networking, even if it will achieve all of its currently predicted properties will not be enough. Since, for a long time into the future, 5G and 6G will be available locally, while edge-cloud deployments will have to encompass highly heterogeneous networking, additional software layer will be needed. Such software layer will treat 5G/6G network patches as parts of “global ecosystem” and provide “context information” needed to realize user-requested workflows. Development of this software layer (meta-operating system governing the orchestration) is the goal of the aerOS project. We will report on our progress in subsequent publications.