Software-Defined Networking (SDN) and Network Function Virtualization (NFV) for a Hyperconnected World

Data networks are a fundamental infrastructure of modern societies, supporting industries, transport, communication, government services, finance, etc. Traditionally, since their inception, data networks have been built around tightly coupled, hardware-dependent functions. This design has resulted in rigid networks that are slow to adapt to changes in communication technologies and services. As networks grew in scale and heterogeneity, and service demand increased, these limitations became increasingly significant, motivating the development of more flexible and programmable networking paradigms.

Software-Defined Networking (SDN) has been proposed to address the rigidity of traditional data network architectures by separating the network control logic from the underlying forwarding infrastructure. This separation allows the network’s functionality and purpose to be managed through software, thus, abstracting the complex mechanisms of the underlying hardware. In parallel, Network Function Virtualization (NFV) has been proposed to replace network functions traditionally implemented in dedicated hardware appliances by softwarised functions deployed on general-purpose computing platforms.

Together, SDN and NFV have helped to make data networks more flexible and adaptive to new requirements and changes. However, the evolution of data networks towards 5G, 6G, O-RAN, Edge computing, and Internet of Things is introducing new complexity and heterogeneity challenges. In addition to accommodating diverse technologies and services, future networks are expected to adapt to application requirements, security threats, and resource constraints. As such, the role of SDN and NFV has evolved from abstracting complex hardware configurations and service requirements to providing the foundation for intelligent network operations. This evolution is particularly important in the context of emerging systems like 5G and 6G where resilience, efficiency, and autonomy are becoming essential.

Against this background, this Special Issue was conceived to explore how SDN and NFV can support the evolution of data network architectures towards a hyperconnected Internet. The contributions in this issue address this question from different yet complementary perspectives. These perspectives include intelligent and adaptive network control, security and resilience, emerging low-latency services, and sustainable network operation. Collectively, these contributions illustrate how SDN and NFV could be extended beyond their initial role to support increasingly complex, heterogeneous, and dynamic data networks.

The timeliness of this special issue is underscored by the emergence of Open Radio Access Networks (O-RAN), AI-native networking and the increasing pursuit of autonomous network operations. O-RAN extends openness and programmability into radio access networks, creating new opportunities for software driven control, automation and innovation. At the same time, future 6G systems are expected to embed Artificial Intelligence (AI) into network management. These developments place greater demands on flexibility, observability and orchestration across heterogenous environments. SDN and NFV provide key enabling mechanisms for meeting these demands through abstraction, programmability and dynamic service deployments.

The first group of contributions addresses the use of SDN to enable more intelligent forms of network management. The first article in this group investigates how fair spectrum sharing under heterogeneous quality-of-service requirements could be achieved in SDN-enabled O-RAN systems using a combination of programmable control and game theory-based resource allocation schemes [1]. The second article addresses improving latency, load balancing, and resilience in hierarchical cloud–edge networks by proposing joint controller placement and node assignment using SDN [2]. The third article investigates the use of AI in software-defined wireless networks to adapt these networks to changing operating conditions and application requirements [3].

The second group of contributions addresses security and resilience in programmable networks. Mitigation of denial-of-service attacks in vehicular networks is investigated in one of the papers in this group through location-aware SDN [4]. The security and resilience of smart city environments is also investigated through the combination of SDN and machine learning [5]. At the SDN data plane, a review of existing research highlights the need for stronger security mechanisms and identifies programmable data planes, heterogeneous-network security, and trust-based approaches as important future directions [6].

Beyond network control and security, SDN and NFV are also examined as enablers of emerging networked services. Both concepts are proposed as enablers of tactile Internet and immersive communications by supporting ultra-low latency, high reliability, and dynamic adaptation required by applications such as remote surgery, industrial automation, and extended reality [7].

Sustainability and resource efficiency constitute another important theme considered in this issue. The application of SDN, NFV, along with Cloud-RAN is investigated in the context of cellular networks to improve resource efficiency and flexibility [8]. This perspective extends the role of programmability beyond performance and adaptability, highlighting its potential to support more energy-efficient network operation.

All the contributions presented in this issue showcase the evolution of SDN and NFV beyond their initial intended programmability and virtualization roles. Across the different application domains considered in this special issue, it is clear that these technologies are increasingly adopted to provide support for more efficient and intelligent management of networks and their resources. The central takeaway from the special issue is that the evolution of networking is moving from programmable networks toward autonomous networks, and SDN and NFV are emerging as the foundational technologies enabling this transformation.

However, such evolution still faces several challenges before truly autonomous data networks can be achieved. Reliability and trustworthiness of the intelligent control mechanisms used in data networks remain challenges to address. Another challenge is the need to strengthen the security of programmable network components. The efficient use of resources, especially energy in data networks also remains a challenge alongside improving traditional performance objectives such as throughput and latency.

Several emerging research directions are also expected to shape the future evolution of programmable and virtualised networks. These include digital twin for real-time monitoring, optimisation and predictive management, explainable AI to improve the transparency and trustworthiness of network control decisions, and federated learning approaches that enable distributed intelligence, while preserving data privacy and security mechanisms for protecting AI-driven SDN controllers and orchestration frameworks.

We hope that the contributions presented in this Special Issue will stimulate further research toward more intelligent, adaptive, secure, and efficient network architectures.