Intelligent edge computing for B5G networks

Beyond-5G (B5G) networks are envisioned to provide both human-centric and machine-centric services through billions of heterogeneous devices. Hence, B5G is expected to support multiple industry verticals such as transportation, food and agriculture, healthcare, energy, manufacturing, entertainment and gaming, and smart cities. Industry vertical applications (e.g., self-driving vehicles and remote surgery) require highly reliable, low-latency, and secure communications. To meet the stringent service requirements of multiple industry verticals, a new paradigm called edge computing (EC) was introduced by leading telecom operators and vendors. EC moves computing and networking facilities from the core to the network edge in order to reduce latency and enhance security. Nevertheless, intelligent and autonomous decision making is required at the network edge to efficiently utilize the limited resources available in EC nodes and optimally orchestrate network slices to provide services to industry verticals dynamically. In this chapter, first we will discuss EC architecture from a standardization point of view and the key technology enablers for EC, including network functions virtualization, software-defined networking, service function chaining, and network slicing. Then, we will discuss the role of artificial intelligence (AI) and machine learning (ML) techniques in EC-enabled B5G networks to make intelligent and autonomous decisions without violating the service-level agreements. Finally, we will discuss the challenges in applying AI and ML techniques for improving performance in EC-enabled B5G networks, and potential research directions to explore the field further.