Standardization Activities on Future Networks in ITU-T SG13

3.1 Network softwarization and IMT-2020

Network softwarization is a technical approach for designing, implementing, deploying, managing, and maintaining network equipment and/or network components by using software programming. As Figure 2 shows, a typical use case of network softwarization is to virtualize various network components, such as base station fronthaul and mobile core, and produce a network slice collecting these virtualized components. SG13 produced a basic concept of network slicing to be applied to IMT-2020, while 3GPP studied network slicing to create a set of implementable specifications.

Fig. 2. Concept of network softwarization.

The domestic discussion of network softwarization for preparing the contributions for SG13 was conducted in 5GMF (The Fifth Generation Mobile Communications Promotion Forum). Japanese experts play a leading role in the study of network softwarization including the responsibility as rapporteur for Question 21, which is mandated to address this issue. A series of recommendations, such as Y.3150 mentioned above, Y.3151 (software-defined networking parts including fronthaul), and Y.3154 (resource pooling for network slice management), have been produced.

3.2 Cloud computing

The study on cloud computing was started by FG Cloud (Focus Group on Cloud Computing) established in 2020. Historically, so-called de facto standards prevailed over de jure standards including ITU-T recommendations in the computing industry. Therefore, particular consideration should be made in terms of standardization strategy for ITU-T to have influential outcomes. SG13 has been attempting to establish a collaborative relationship with other standard bodies that are more influential than ITU-T. Even though the influence of de jure standard bodies is limited, ITU-T could play a role in specific issues such as terminology and reference architecture. SG13 established two collaborative teams with ISO/IEC JTC1^* Subcommittee 38 Working Group 3 to jointly address these issues. The outcomes of this collaboration resulted in ITU-T Recommendations Y.3500 (terminology of cloud computing) and Y.3502 (reference architecture of cloud computing).

Japanese experts focus on the system called “inter-cloud,” which combines cloud computing and a wide area network connecting different datacenters. An overview of the inter-cloud is given in ITU-T Recommendation Y.3511 (framework of inter-cloud) to which Japanese experts contributed and played a leading role.

The cloud study in SG13 is moving to big-data-related subjects running on clouds rather than the cloud itself. The current structure of big data in the industry is the silo type structure in which dominant players collect and manage their own big data. The model studied in SG13 is more horizontally federated among different players with specific roles. A set of recommendations of the Y.3600 series describe this issue.

3.3 AI/ML

AI/ML is one of the hottest topics in this industry. SG13 started AI/ML-related activities through FG ML5G, which was established to consider possible applications of AI/ML in networks. The initial target of AL/ML study was automation of network management. Similar activities are being conducted in the European Telecommunications Standards Institute (ETSI) (e.g. ZSM-ISG (Industry Specification Group (ISG) Zero Touch Network and Service Management), and SG13 liaises with these groups. FG ML5G, which was established in 2018, concluded its activities in June 2020 and submitted the deliverables to SG13. Further discussions including development of ITU-T recommendations will progress in relevant questions such as Question 20.

The basic architecture of ML studied in FG ML5G is described in ITU-T Recommendation Y.3172 (architecture framework for machine learning in future networks including IMT-2020). The architecture in Y.3172 contains two different environments, i.e., sandbox in which an ML algorithm learns using a simulator and pipeline in which the algorithm is applied in an actual management task (Fig. 3).

Fig. 3. Machine learning architecture in Y.3172.

Big data, which is being studied in SG13 in the context of cloud computing, would be applicable to AI/ML. Questions 17, 18, and 19 working on big data will join the AI/ML-related activities.

3.4 Quantum key distribution

Quantum key distribution is a system to distribute encryption keys to remote nodes in a secure manner by using quantum physics. Quantum key distribution is becoming mature for small-scale deployment including commercial services for specific applications. Question 16 has been working on networks for quantum key distribution since 2018.

It is not possible to provide quantum key distribution over a long-distance fiber network due to attenuation of optical signals on the fiber. The distance between nodes of quantum key distribution is not more than 100 km, even with the latest techniques. This imposes constraints on the design of large-scale networks such as nation-wide and global networks. The architecture being studied in SG13 introduces intermediate nodes that are placed in secure locations such as central offices of a telecom carrier (Fig. 4). This allows the provision of long-distance key distribution beyond the limitation of quantum key distribution without intermediate nodes.

Fig. 4. Conceptual structure of quantum key distribution network.

In principle, quantum key distribution provides the capability of sharing only a small amount of data, such as an encryption key, rather than a huge volume of data such as audio, video, and text. Therefore, quantum-key-distribution networks are assumed to provide enhanced security for communication applications on traditional networks. How quantum key distribution networks and traditional networks work together is an important subject of standardization, and SG13 is working on this issue.

There are other areas of quantum-related information and communication technologies such as quantum computing. ITU-T established FG QIT4N (Focus Group on Quantum Information Technology for Networks) to promote studies in these areas.

3.5 Network 2030

FG Net2030, which was proposed by a Chinese company, was established to promote study on the network vision called “Network 2030,” which will be achieved in 2030. Network 2030 supports emerging applications, such as holographic-type communications and tactile internet, which will not be available for current network technologies. In this work, a new network technology called “New IP” was proposed. Although the relationship of New IP and a set of Internet Protocol (IP)-based network technologies is not yet clear, this stimulated a significant political discussion on whether ITU-T should handle IP-related technologies. One party of the discussion expressed their concern of New IP, stating that IP-related technologies should be discussed using the multi-stakeholder approach and that the Internet Engineering Task Force (IETF) rather than ITU-T should be responsible for this study. The proponent responded that the proposed study items do not overlap with the current work items in IETF, and ITU-T can conduct the proposed study items. The proponent also proposed to change the name from “New IP” to “Future Vertical Communication Networks” to avoid misunderstanding. The discussion on this new study subject is still continuing.