Research & Development & Innovation

A joining of multiple telecommunications standards organisations, 3GPP creates global specifications and standards for the five existing generations of mobile wireless technologies. The organisation continuously works on parallel releases, with common timelines for opening and freezing them, meeting four times a year to outline and develop new releases:
  • global standards organization for mobile communication
  • a new generation every 10 years
  • a new release every 15 to 24 months
  • ~20  WG&TSG each meeting 4 to 8 times a year
  • contribution driven
  • consensus based

5G NSA-SA migration

The introduction of 5G SA is expected to facilitate simplification of architectures, improve security and reduce costs. Operators are increasingly experimenting with and deploying 5G standalone (SA) networks. With a totally new, cloud-based, virtualised, microservices-based core infrastructure, anticipated benefits of 5G SA technologies include faster connection times (lower latency), support for massive numbers of devices, programmable systems enabling faster and more-agile creation of services and network slices, with improved support for management of service-level agreements within those slices. The technology is expected to enable customisation and open up new service and revenue opportunities tailored to enterprise, industrial and government customers. 
Realizing the full benefit of 5G requires the implementation of 5G Standalone architecture. That introduces new 5G technologies end-to-end, supporting new 5GC-based services and removing the dependency on the older LTE. MNOs started the commercial introduction of 5G by deploying New Radio (NR) technology in their networks based on the so-called Non-Standalone architecture (NSA) which requires that LTE acts as the master technology, while also using the existing 4G core network (EPC).The next step in network migration is the introduction of the 5G core network (5GC) allowing for a 5G Standalone architecture (SA) to realize the full set of 5G advantages. An Option 2 architecture, where customers’ devices are supported exclusively by NR carriers, is proposed to support this migration. However, the simple introduction of Option 2 risks the new 5G SA customers in many networks experiencing data rate performance levels below the level of existing NSA customers or even the legacy LTE customers. To avoid this situation, SA network architecture should be complemented by the development of the Option 4 as an extension to Option 2 in order that 5G can quickly realizes its full potential in all networks.In addition to Option 4, 3GPP defines two additional architecture options known as Option 5 and Option 7 that enable the use of LTE as well as NR with the 5GC. 
  • Option 5 is applicable in areas without any NR coverage but with LTE coverage.
  • Option 7 is applicable in areas with limited NR coverage using higher frequency bands still requiring LTE as an anchor layer.
  • Option 4 is applicable in areas where NR provides wide coverage and is overlaid by LTE, enabling use of NR as the master technology.
Option 4, Option 5 and Option 7 are complementary to Standalone Option 2 in supporting a faster adoption of the 5GC in a wider variety of deployments. Option 4 is the only one that makes NR the master technology and as NR spreads more widely, more and more network areas become suitable for Option 4 deployments. Option 4 removes the inherent NSA evolution constraint whereby NR is dependent on LTE to provide the anchor connection. Option 4 therefore encourages wide area NR as well as 5GC deployment, enabling a flexible long-term migration towards Option 2. For these reasons it is proposed to focus initially on Option 4 as the more strategic long-term option that addresses the main impediments to SA Option 2 deployment. Two technology alternatives have been proposed to mitigate the reduction in user data rates arising from use of SA without Option 4, Dynamic Spectrum Sharing (DSS) on all carriers and On-demand NSA fallback. Without Option 4, the 5GC with NR SA developments are at risk of near-term deployment due to the lack of sub 3 GHz NR spectrum for most operators, making NR SA uncompetitive compared to NSA in data rate performance in many locations. Therefore, Option 4 should be considered as a standalone complement to facilitate a near-term move towards NR SA deployments.In practice, Option 4 is a comparatively small extension to Option 2. Option 4 should not add any hardware complexity in either the terminal or the base station. Implementation costs mainly relate to software development and testing. Option 4 is already standardized within the Release 15 core specifications. However the commercialization of Option 4 requires 3GPP to finalize work in two areas: test specifications and bands combinations.

4G/5G Core network architecture comparison

The transition from 4G to 5G is not merely an enhancement in speed or bandwidth; it's a revolutionary leap that brings a fundamental transformation in network architecture.  
  • 4G is primarily based on the LTE (Long-Term Evolution) standard. The main components in the 4G architecture are: User Equipment (UE) the mobile device used by the end user, Evolved NodeB (eNB) tThe base station that connects the UE to the core network, Evolved Packet Core (EPC) the core network in the 4G architecture, which has several entities like SGW, PGW, MME, etc.
  • 5G is designed to offer faster speeds, lower latency, and support for many devices. The 5G architecture is based on the Service-Based Architecture (SBA) and introduces the concept of Network Functions (NF). Some of the key components include: User Equipment (UE) similar to 4G but enhanced to support 5G speeds and features, gNodeB (gNB) the equivalent of eNB in 5G,  5G Core (5GC) the core network in 5G with several network functions like AMF, SMF, UDM, etc.
 5G Core (5GC). 5G has introduced a new core architecture known as the 5G Core (5GC). It's a radical departure from the EPC and is designed to be more modular, scalable, and flexible. Key components of 5GC include:
  • AMF (Access and Mobility Management Function): Manages user access to resources and mobility.
  • SMF (Session Management Function): Handles session management tasks like session establishment, modification, and release.
  • UDM (Unified Data Management): The new version of HSS, it manages user data.
  • AUSF (Authentication Server Function): Assists in user authentication.
  • NRF (Network Repository Function): Maintains information about network functions, aiding in service discovery.
  • NEF (Network Exposure Function): Exposes network capabilities and resources to third-party applications.
  • PCF (Policy Control Function): The evolved version of PCRF, dictates policy rules.
  • UDR (Unified Data Repository): A storage function for structured data.
  • UPF( User Place Function ): The UPF is a key component in the 5G core network architecture, responsible for packet routing and forwarding, packet inspection, and QoS handling for user data.
  • NSSF (Network Slice Selection Function): The NSSF is responsible for selecting the appropriate network slice instance based on the UE (User Equipment) and the service requirements
  • AF (Application Function) : The AF interacts with the core network, primarily for policy and charging purposes. It represents external applications that need to communicate with 5G core components.
Key differences:
  • Service-Based Architecture: Unlike 4G's more rigid design, 5G adopts a Service-Based Architecture (SBA), allowing different network functions to access services from other functions.
  • Separation of Control and User Planes: 5G further decouples the control and user planes, enabling scalability and flexibility in network deployment and management.
  • Stateless NFs: Network functions in 5GC are designed to be stateless, meaning they don't store session state. This improves reliability and fault tolerance.
  • Network Slicing: 5G introduces network slicing, allowing operators to create multiple virtual networks with varied performance characteristics on a single physical infrastructure.

Protocols difference:
  • 4G Core Protocols. The 4G core, often referred to as the Evolved Packet Core (EPC), primarily uses the following protocols: S1AP (S1 Application Protocol): Utilized between the eNodeB (base station in LTE) and the MME (Mobility Management Entity). It's responsible for handling signaling related to mobility and session management. Diameter: This is the primary signaling protocol in EPC, utilized for authentication, authorization, and accounting. It is used among various entities such as MME, HSS (Home Subscriber Server), and PCRF (Policy and Charging Rules Function). GTP (GPRS Tunneling Protocol): Divided into GTP-C (control plane) and GTP-U (user plane), GTP handles user data transport between the eNodeB, SGW (Serving Gateway), and PGW (PDN Gateway). GTP-C manages signaling for session establishment, while GTP-U is responsible for user data encapsulation and transport.
  • 5G Core Protocols. In the 5G core (5GC), the Service-Based Architecture (SBA) introduces new protocols and refines existing ones: NGAP (NG Application Protocol): Used between the gNodeB (base station in 5G NR) and the AMF (Access and Mobility Management Function). It's the 5G counterpart of S1AP, dealing with signaling related to mobility and session management. HTTP/2: In the SBA of 5GC, many functions communicate using HTTP/2-based services, allowing for a more flexible and scalable network. PFCP (Packet Forwarding Control Protocol): Replaces GTP for control signaling related to user plane function. Used between the SMF (Session Management Function) and UPF (User Plane Function). GTP-U: While GTP-C is largely phased out in 5GC, GTP-U still remains as the primary protocol for user data encapsulation and transport between the gNodeB and UPF.
Differences between 4G and 5G Core Protocols:
  • Service-Based vs. Interface-Based: 5GC adopts a Service-Based Architecture where network functions provide services to other functions over a common bus. This contrasts with the more rigid, interface-based approach of the EPC.
  • Introduction of HTTP/2: 5GC uses HTTP/2 for many internal signaling and communication tasks, contrasting with the more traditional telecom protocols used in 4G.
  • Protocol Refinements: While GTP continues to be used in both 4G and 5G for user data, its control plane counterpart, GTP-C, is largely replaced by PFCP in 5G.
  • Unified Data Management: 5GC introduces protocols and mechanisms to handle data in a more unified manner, especially with functions like UDM (Unified Data Management).
  • Network Slicing: 5G introduces protocols to support network slicing, allowing for multiple logical networks on the same physical infrastructure, catering to different service requirements.
  • Decoupling of Control and User Planes: 5GC further separates control and user planes, enabling more flexible deployments and scalability. This is reflected in the protocols and their interactions.
Reduced Capability in 5G NRRedCap is a new device class that is designed to provide a cost-effective way to connect devices that do not require the full capabilities of 5G.5G RedCap devices are designed for mid-tier IoT use cases, such as:
  • Industrial automation: Remote monitoring and control of machines and equipment
  • Smart city applications: Traffic management, environmental monitoring, and smart lighting
  • Wearable devices: Fitness trackers, smartwatches, and augmented reality glasses
  • Connected vehicles: Telematics, infotainment, and autonomous driving
These use cases do not require the same high data rates as eMBB applications, such as video streaming and gaming. RedCap devices can provide a good balance of performance and cost for these use cases.In addition to the lower data rates, RedCap devices also have a number of other limitations compared to eMBB devices.These include:
  • Reduced number of antennas: RedCap devices typically have a single transmit antenna and a single receive antenna, while eMBB devices can have multiple antennas. This can limit the performance of RedCap devices in areas with poor signal reception.
  • Lower transmit power: RedCap devices transmit at a lower power than eMBB devices. This can limit their range and coverage.
  • Higher latency: RedCap devices have a higher latency than eMBB devices. This is because they use a different radio access technology (RAT) and have to go through more processing steps.
Despite these limitations, RedCap devices offer a number of advantages over other IoT technologies, such as NB-IoT and LTE-M. These advantages include:
  • Higher data rates: RedCap devices can provide significantly higher data rates than NB-IoT and LTE-M. This can be important for applications that require real-time data transfer, such as industrial automation and connected vehicles.
  • Improved spectral efficiency: RedCap devices are more spectrally efficient than NB-IoT and LTE-M. This means that they can support more devices in a given area.
  • Support for 5G features: RedCap devices support a number of 5G features, such as network slicing and 5G LAN. This can make them more versatile and future-proof than NB-IoT and LTE-M devices.
Overall, 5G RedCap is a promising new technology for mid-tier IoT use cases. It offers a good balance of performance, cost, features, and it is well-positioned to meet the needs of a wide range of applications.RedCap devices support a maximum downlink throughput of 150 Mbps & a maximum uplink throughput of 50 Mbps. They also support a maximum bandwidth of 20 MHz. This is significantly lower than the capabilities of 5G eMBB devices, but it is still sufficient for many use cases.RedCap devices are also designed to be more power efficient than eMBB devices. This is important for devices that are battery-powered, such as wearables & sensors.
5G Network slice managementNetwork slicing is seen as one of the key features for 5G, allowingvertical industries to take advantage of 5G networks and services. The 5G system is expected to be able to provide optimized support for a variety of different communication services, different traffic loads and different end user communities. For example, the communication services using network slicing may include V2X services, 5G seamless eMBB service with FMC and massive IoT connections. A network slice is a logical network that provides specific network capabilities and network characteristics, supporting various service properties for network slice customers.Management and orchestration of 5G networks and network slicing is a featurethat includes the following work items: management concept and architec-ture, provisioning, network resource model, fault supervision, assurance andperformance management, trace management and virtualization managementaspects.
3GPP SA5 specifies network slicing charging per UE and per tenant based on the charging requirements from NGMN, GSMA (NG.116), 3GPP SA1 and SA2 (TS 23.501 and TS 23.502) for different business models (e.g. B2B Services, NSaaS). 
5G architecture support for XR and media servicesTo provide immersive experience for users, XR and Media services (XRM) are usually characterized by high data rate and low latency. 3GPP SA2 has started the XRM Work Item in Rel-18 to identify 5GS functionality and capability enhancements. including multi-modality transmission, 5G system (5GS) information exposure, PDU set-based QoS handling, uplink-downlink transmission coordination, Packet Delay Variation monitoring and reporting, and Power Saving enhancements.
3GPP general timing modelIt shows that the standardization of mobile networks is also considering a wider approach to timing issues rather than simply setting extremely strict constraints on the wireless segment. Considering the timing of the application itself, as well as the higher communication layers, can lead to better resource efficiency, and considering different metrics can increase the flexibility of the system with respect to the needs of different applications. 
Edge computingThe concept of edge computing is not confined to 5G, and the most general concept of edge computing is referred to as Multi-access Edge Computing (MEC), which is access agnostic. MEC is defined to be a system which provides an IT service environment and cloud-computing capabilities at the edge of an access network which contains one or more type of access technology, and in close proximity to its users. In practice, the concept of edge is not limited to the edge of the access network and can vary from the end device (e.g., sensors, smartphones, etc.) used to the local data centres of service providers. At 3GPP, the edge computing work revolves around the definition of MEC with focus on 5G, where the IT service environment and cloud computing capabilities are provided at the edge of the Next Generation Radio Access Network (NG-RAN) and the edge computing capabilities are natively supported in the 5G Core (5GC). Edge computing, simply put, is where core network and cloud computing capabilities are moved to the "edge" of the network closer to the customers, reducing the physical distance for communications. 
  • The most prominent benefit of edge computing is the reduced latency. Taking an illustrative example below for a 5G network, one observes that it is possible to reduce the latency significantly (factor of 2 to 10 depending on assumptions) with edge computing.
  • It is also possible to optimise data flows and leverage local processing/storage with edge computing, which may reduce the operational costs and complexities associated with moving data to/from cloud servers.
  • Furthermore, edge computing can improve security and privacy protection, as the data is kept at the edge (or even in the customer premise). Whilst inadequate security measures can make devices and the network nodes in edge vulnerable to hacking, the data possessed by the edge devices and network nodes tend to be limited compared to the public cloud servers that store massive data. This hopefully will reduce incentives to hack edge networks.
  • Edge computing can also enhance reliability and resiliency of the system. The possibility of having local processing/storage can enable Information Technology (IT) applications to function continuously even in the midst of intermittent connectivity. For example, when local servers are able to process most of the demand and synchronise central cloud servers occasionally, the IT applications do not have to rely on connectivity and central servers to function. Furthermore, edge computing systems may enhance resiliency of the overall systems. As functionalities are distributed over different edges, a failure of an edge does not lead to failure of other edges. On the other hand, if central server fails, the centralised computing system fails overall.  
  • The benefits above translate into business potentials and new use cases in various sectors. 
Non-Public Networks (NPN)NPN, also referred to as a private network, is a network that is intended for the non-public (or private) use. Whilst the concept has been implemented in previous generations, 5G specifications support various configurations of NPN. The 5G system needs to cater to a variety of use cases, spanning a range of requirements including the ones that were not traditionally served in the past generations. For example, some use cases require the data used to stay within the boundaries of the organization, to protect privacy and resolve security concerns. In such cases, some (if not all) of the network entities need to reside locally in the organization's premise and/or be owned by the organization. Indeed, the non-traditional (i.e., vertical) use cases and organizations that have not used mobile connectivity for its private use tend to come with these new sets of requirements, requiring NPN to be deployed.For these use cases, the NPN may provide three benefits in terms of performance. 
  • First, NPN brings optimized coverage for the owner as the network is deployed for the owner's purpose and not for the public in general. 
  • Second, the local presence of some network entities reduce physical distance and number of network hops required for the use case, reducing the latency of NPN. 
  • Finally, the local presence reduces uncertainty of network operation, as the disruption (e.g., outage, fault) of the corresponding entities can be managed/resolved inside. Note that the benefits may not be significant (or non-existent) depending on the deployment architecture.
  • NPN also enables more control from the owner's side in general. The security of the network can be enhanced by restricting unauthorized users and the privacy of the data treated can be enhanced. In addition, traffic can be prioritized depending on the types of users in the organization as per its requirements. Similarly, management of congestion and interference can be optimized according to the requirement of the organization.
The definition of non-public network is a network that is intended for non-public use. It can be deployed in a variety of configurations where both virtual and physical network elements can be utilized. It includes the concept of private network, which is an isolated network deployment that does not interact with a public network. While there are many possible configurations of NPNs, 3GPP defines two major categories of NPNs: Standalone Non-Public Network (SNPN) and Public network integrated NPN (PNI-NPN).
NPN provides a 5G network solution specifically used for industry users, including stand-alone non-public network (SNPN) and public network integrated NPN (PNI-NPN). In the Rel-17 eNPN work item, SNPN is enhanced to support more NPN application scenarios through cooperation between different networks or entities. For example, access authorization is provided for the UE without local certificate or subscriptions, and IMS as well as video, imaging and audio for professional applications (VIAPA ) scenarios are supported. The Rel-18 eNPN_Ph2 work item further studies how to support mobility between SNPNs, how to access SNPN via non-3GPP, how to enable UE to discover, select, access the local hosting NPN and the localized services via the hosting NPN with proper authorization, and whether to enable NPN to support 5G proximity services.
The term Non-Public Network was introduced in Rel-16. In Rel-15, the terms private network and isolated deployment were used. The starting point for the security of NPNs is that the same level of security for 5G deployments in Public Land Mobile Networks (PLMN) of course also applies to Non-Public Networks. The difference between NPN and PLMN from a security standardization point of view is only based on different use case requirements for security features. 3GPP SA3 has included security features for Non-Public Networks from Rel-15 and extended them in Releases 16, 17 and 18. 
MFA Alliance for private networks offers guidance to businesses looking to harness the potential of private networks. From demystifying the implementation process to collaborating with industry giants, MFA plays a pivotal role in driving the global adoption of private networks.MFA is making significant strides to facilitate the adoption of private networks within various industry verticals. MFA’s approach is centered around the development of the MFA’s Uni5G™ technology blueprints for 5G, a pioneering tool designed to bridge the gap between industry requirements and the vast array of optional features offered by the 3GPP specification. At the core of MFA’s mission lies the continuous refinement and expansion of these blueprints, ensuring that they remain in step with 3GPP standards and the ever-evolving needs of vertical industries. 
3GPP 5G NTN StandardsThe Mobile Satellite Services (MSS) market has historically been a niche market due primarily to the fact that MSS is based on proprietary technologies. However, 3GPP is working with the satellite industry on a global standardized solution, called 5G Non-Terrestrial Networks (NTN). 5G NTN will enable seamless roaming between terrestrial and satellite networks, using largely standard cellular devices, i.e., eliminating the need for proprietary terminals and fragmented satellite constellations. This could dramatically increase the addressable market for mobile satellite services.In 2022, 3GPP introduced two parallel workstreams in its Release 17 specifications addressing 5G satellite-based mobile broadband and low-complexity IoT use cases:
  • NR-NTN (New Radio NTN) – adapts the 5G NR framework for satellite communications, providing direct mobile broadband services as well as voice using standard apps. This will enable 5G phones operating on dedicated 5G NTN frequencies and existing sub-7GHz terrestrial frequencies to link directly with Release-17 compatible satellites. Release 17 also includes enhancements for satellite backhaul and the inclusion of 80MHz MSS uplink spectrum in L-band (1-2GHz) plus a similar amount of downlink spectrum in S-band (2-4GHz).
  • IoT-NTN – provides satellite support for low-complexity eMTC and NB-IoT devices, which expands the coverage for key use cases such as worldwide asset tracking (for example, air freight, shipping containers and other assets outside cellular coverage). IoT-NTN is designed for low data rate applications such as the transmission of sensor data and text messages.
Release 17 established the NR-NTN and IoT-NTN standards while the upcoming 5G Advanced Release 18 will introduce new capabilities, coverage/mobility enhancements and support for expanded spectrum bands.Only a few NTN-based satellites have been launched to date. A noteworthy example is Spanish LEO operator Sateliot, the first company to deploy satellites complying with 3GPP’s Release 17 IoT-NTN standard. Sateliot currently has two satellites in orbit and recently carried out a successful roaming test between its satellite network and Telefonica’s 5G terrestrial network using an IoT device with a standard SIM card. Sateliot plans to start commercial activities in 2024. Ultimately, the company hopes to launch a total of 250 nanosatellites, which will enable it to offer global 5G IoT-NTN services.No satellite operator presently supports 3GPP’s Release 17 NR-NTN standard for voice and data.
Artificial Intelligence in 3GPP 5G-AdvancedThe Artificial Intelligence/Machine Learning (AI/ML) techniques and relevant applications are being increasingly adopted by the wider industries and proved to be successful. These are now being applied to the telecommunication industry including mobile networks.Clearly the adoption of AI/ML technology is opening a new era for creating more business value in terms of improved system performance, higher efficiency, enhanced end use experience as well as creating new business models and use cases for 5G and future generation mobile networks, AI/ML capabilities are used in various domains in 5GS, including management and orchestration (e.g., MDA), 5GC (e.g., NWDAF), and NG-RAN (e.g., RAN intelligence).
3GPP Release 18 marks the start of 5G-Advanced, which represents a major evolution of 5G system and includes comprehensive work in the area of artificial intelligence (AI)/machine learning (ML). AI/ML can be used to manage complex networks intelligently, address system optimization problems, and improve user experience in 5G toward the sixth-generation (6G) cellular networks. In this article, we provide an overview of the AI/ML work in 3GPP Release 18, spanning across multiple 3GPP groups from services and system aspects (SA) to radio access network (RAN).
  • Before Release 18, 3GPP has carried out some initial work to embrace AI/ML techniques and data analytics in 5G system design. In Release 15, 3GPP introduced network data analytics function (NWDAF) in 5G core (5GC) network.
  • In the RAN domain, an initial study on AI-enabled RAN was completed in 3GPP Release 17. The study identified high-level principles and provided a functional framework for AI-enabled RAN intelligence. The study also investigated several use cases and solutions for AI/ML in RAN, including network energy saving, load balancing, and mobility optimization.
  • 3GPP has also introduced management data analytics function (MDAF) as part of OAM. MDAF, as a key enabler for network automation and intelligence, can process data related to network status and service events to provide analytics reports. The input data can come from different types of NFs (e.g., NWDAF) or entities (e.g., next generation node B (gNB)) in the network. MDAF can be deployed at different levels, including at domain level (e.g., RAN, core network) to provide domain-specific analytics and in a centralized manner to offer end-to-end or cross-domain analytics service.
The 3GPP work on incorporating AI/ML techniques and data analytics into 5G system design conducted before Release 18 lays a solid foundation for further embracing AI/ML in 5G-Advanced evolution. Indeed, 3GPP Release 18 includes a diverse set of study and work items in the area of AI/ML, spanning across multiple 3GPP working groups, as illustrated in figures: AI/ML Model Transfer, AI/ML for Media.
Management, orchestration and charging5G New Radio (NR) specifications and by putting in place the 5G Phase 1 System Architecture have brought about the need for new management standards, as 5G adds to the ever-growing size and complexity of telecom systems. 3GPP Release 15, which includes building up a new service-oriented management architecture and all the necessary functionalities for management and charging for 5G networks.Management and orchestration of 5G networks is a feature that includes the following work items: management concept and architecture, provisioning, network resource model, fault supervision, assurance and performance management, trace management and virtualization management aspects. Commercial deployment of the 5G System will not be possible without capabilities for Operators to be able to monetize the various set of features and services. One key evolution of the charging architecture is the adoption of a service based interface integrated into the overall 5G system service based architecture, enabling deployments of charging functions in virtualized environment and use of new software techniques. The new charging function (CHF) and service based interface Nchf are introduced in the 5G system architecture. 
Network softwarizationThe evolution to 5G and future mobile telecommunication networks is characterized by a significant surge in demands in terms of performance, flexibility, portability, and energy efficiency across all network functions. Softwarized network architecture integrates the principles of Software-Defined Networking (SDN), Network Function Virtualization (NFV), and cloud computing to mobile communication networks. The softwarized network architecture is designed to provide a suitable platform for novel network concepts that can meet the requirements of both evolving and future mobile networks.The developments of SDN, NFV, and CC will pave the path for various other services to be migrated to virtualized platforms. For instance, cloud factories will combine the resources in multiple physical locations to a virtualized platform and coordinate their operation to facilitate manufacturing through cloud-based technologies. It is evident that this type of a virtualized network requires end-to-end orchestration of network and services. Furthermore, service network automation, together with network slicing and multiaccess edge computing, leads mobile networks to be fully softwarized 5G and B5G networks. Network softwarization facilitates self-organizing, self-configurable, and self-programmable flexible network infrastructure to facilitate emerging heterogeneous applications and use-cases. In line with these developments, 6G is expected to facilitate intelligent network softwarization by harnessing the capabilities of AI techniques such as machine learning and deep learning. The path from traditional nonsoftwarized networks toward intelligent softwarized networks is illustrated in figure below.
Open issues and future directionsWireless communication is a fast-growing field that attracts much research attention, from both industry and academia. As we progress to the next decade, the need for wireless connectivity will continue to grow, and the big challenge arises when the expected wireless system support new applications, such as metaverse, vehicle communication, and at space-air-ground domain. 
  • Joint communication and sensing.  Future wireless system not only desires for highly reliable and faster connectivity, but also the ability to sense surrounding dynamics with ubiquitous wireless signals. Such sensing capabilities include range and velocity estimation, object detection, collision avoidance, and localization. In the past, communication and sensing have been designed separately, where either communication or sensing is the main target design, the other as the by-product. However, integrating both functionalities within one system has clear advantages in providing better power and spectral efficiency, as well as reducing hardware and signaling cost through efficient resource coordination. Joint communication and sensing has attracted extensive attention from the perspective of jointly considering and unifying two operations, especially in emerging applications like vehicle-to-everything (V2X), where simultaneous information exchange and radar-like parameter estimation are critical. Yet, achieving integration gain from Joint communication and sensing system faces numerous challenges. Among them, Joint communication and sensing physical layer design such as waveform optimization, collaborative resource allocation, and beamforming are of paramount importance.
  • Space-air-ground communication.  One limitation for 5G wireless system is that it provides network access mainly for terrestrial (ground) communication. Emerging applications in the space, such as satellite, and much different. For example, UAV communication expects LoS scenarios (signal propagation is in favor condition) but power and trajectory optimization are more important design factors. Recently, the space-air-ground integrated network (SAGIN) has been proposed. SAGIN provides seamless coverage for larger areas, including sea, space, ground, and air. SAGIN has to consider various factors from each segment, current design focuses on protocol optimization, resource allocation, performance analysis, mobility management, and inter-segment operation. Furthermore, the network design and system integration in SAGIN are of great significance.
  • Semantic communication.  Communication should be used not only for exchanging data bits, but also for semantic exchange. In fact, many scenarios involve semantic information. For example, transmitting natural languages.Current solution needs to convert language into bits via upper layer operations, then simply send those bits through medium. Essentially, semantic communication will utilize advanced ML techniques to perform semantic encoding at transmitter and correspondingly semantic decoding at the receiver end, avoiding the source encoding/ decoding directly. The challenge is that current system will undergo significant modifications, for example, the need for new metric for semantic entropy, semantic channel, and noise factor. Nevertheless, semantic communication is regarded as an important component in 5G beyond.
  • Data-driven communication system design.  The exponential growth of data traffic and recent advancement on ML fuel the data-driven communication system design. In particular, collected data can help innovative designs on modulation, coding, scheduling, architecture, resource management, and even end-to-end. ML is one of the most powerful tools and can effectively explore massive data and make accurate predictions and plannings. For example, future communication design may not start from problem formulation and then be solved with traditional convex (or non-convex) optimization. Rather, the model design can learn from data and reach something we have yet seen. Another big challenge that data-driven approach can address is the network scalability, which is of critical importance in massive IoT era. 
Sensory 6G and digital immersionSensory 6G describes digital services that humans can experience with an expanded range of physical senses. When 6G networks are combined with optimized devices, applications, and content, people should interact with virtual and physical environments in new ways. People may experience a tactile response, where the human skin can sense conditions like temperature, texture, or friction. Equally, people could experience kinesthetic effects like position, speed, and force.Delivering the appropriate quality of experience (QoE) requires the interplay of three layers: an infrastructure layer considering network and rendering capabilities; a consumption layer considering how to access experiences in terms of format and device; and a human layer considering mental, physical and perceptual factors.Mental factors matter because digital sensory interaction can have good or bad neurological impacts. Physical factors include the span of human senses involved and required body placement to  experience the service. Perceptual factors include degrees of immersion in virtual contexts andwhether the provided environment is functionally realistic. All together, these aspects should become core to any quality of experience framework for 6G. 
Massive twinningThe development of a digital twin (DT) may fundamentally benefit from data about events occurring in the real world at a certain time and place.Future RAN aims to bridge the gap between the physical and digital worlds, making it possible to create a digital counterpart of anything in the actual world. The ability to create a trustworthy, efficient, and effective digital duplicate of any physical object will open up hitherto inconceivableavenues of exploration, such as agriculture, smart cities, manufacturing. 
  • To improve the quality of life in a city, a comprehensive smart city must give a wealth of data, such as a digital reproduction of the city with real-time data on control utilities, pollution maps, traffic. An interactive 4D map is a key component of an immersive smart city, since it can be used to better organise and design essential infrastructure and services, like transportation, waste management, plumbing, electrical wiring, and more. Sensors installed to track, monitor, and update status of these services must be localised for this kind of dynamic map to be possible. Moreover, units may be stationed by the side of the road to aid in communication, while simultaneously operating as traffic monitors in various sections of the city.
Shaping waves in complex media with tunable metasurfaces: From acoustics to electromagnetismThe main idea is to replace the transmitting antennas by a smart modification of the wireless environment by physically shaping the propagation medium to achieve optimal focusing and channel diversity. Optimization of these metasurfaces results from the generalization of the time reversal mirror concept to the one of the convolution products of different time-reversal mirrors associated with each transmitter and receiver. These approaches are in term of the number of spatio-temporal degrees of freedoms of a wavefield. These metasurfaces are called reconfigurable intelligent surfaces (RIS) both in the narrowband regime for electromagnetic communications and in the wideband regime for sound communications. When dealing with audible sound waves, the problem is much more complex that for narrowband RIS, as the reconfigurable surface needs to cover a wide range of frequencies spanning several decades. This research opens the field of ultrawideband RIS communications systems.Digitally-controlled reconfigurable metasurfaces Future communication networks require real-time control of how radio waves radiate, reflect and scatter through the propagation medium. In recent years, metasurfaces, engineered bidimensional structures that are naturally prone to reconfigurability, have been proposed as a key component for the next communication systems as they could be employed both to design beam-scanning antennas and to implement reconfigurable metasurfaces in general.
  • The first types of reconfigurable antennas relied on phase array architectures, based on the use of power amplifiers and phase shifter, which require a considerable amount of power to steer the antenna beam. 
  • A new type of reconfigurability, suitable for greener, more power efficient and environment friendly devices, is based on the use of reconfigurable elements such as varactors, PIN diodes, transistors, switches, to name a few. These elements require a considerably lower amount of energy to perform their operation, when compared to a phased array configuration and can be integrated directly onto the metasurfaces. Their use can lead to power efficient beam-scanning antennas, as well as to quasi-passive Reconfigurable Intelligent Surfaces.
Despite the promising effects of this new type of electronic reconfigurability, the design of the devices that make use of this technology can be highly complex, if the number of reconfigurable elements is high and/or if the device needs to provide multiple functionalities. Therefore, designing such complex structures if often prohibitive when using commercial electromagnetic software.