Unleashing 5G & 6G with
AI-empowered Automation

qualitative management of emerging and future networks and services.

The ZSM framework (depicted in Figure 1 on page one, and specified in ETSI ZSM 002) is versatile and built on service-based principles offering scalability, modularity, extensibility, and flexibility. It supports the transfer of autonomy from the operator to the network using intent-based interfaces. The framework provides capabilities to integrate AI-based functionalities and enable closed-loop automation.

The framework supports the separation of management and automation into areas of concern, or management domains (where the scope is delineated,for example, by an administrative or technological boundary, such as Radio Access) and end-to-end cross-domain service management domains. Each management domain is responsible for the fulfillment and assurance processes within its scope. The separation of concerns allows the abstraction of the complexity of the management domains.

The ZSM framework supports open interfaces as well as model-driven service and resource abstraction. The management services, which are exposed by the management domains, are described and specified. The architecture allows operational data to be kept separately from the management applications, enabling rapid and efficient access to current, real-time management data within and across the management domains to support the automation processes.

The framework is designed to enable adaptive, closed-loop automation—providing a feedback loop (depicted in Figure 2 on page two) between data monitoring, data analytics, decision-making and adaptive actions that aims to reach and preserve a set of objectives without external intervention.

ETSI ZSM 009-1 on closed-loop automation enablers specifies “Governance” services that allow for the creation, execution, and lifecycle management of a closed loop as well as the configuration of related policies and rules to steer the behavior. The “Governance” services also support interaction between closed loops and external entities, such as human operators, allowing them to supervise the operation and performance of closed loops. As more automation and closed loops are deployed and start to operate safely and efficiently, human trust will increase and the requirement for a level of supervision and visibility will diminish.

As depicted in Figure 3, closed-loop operation can be implemented at the management-domain level. Closed-loop operation can also occur at the end-to-end