Network Slice Orchestration

Slicing only scales if creating and running slices is automated, and that automation is what slice orchestration provides. It covers the whole lifecycle: taking a service requirement, translating it into a slice design, instantiating the NSIs and NSSIs across RAN, transport, and core, activating them, then monitoring, scaling, healing, and eventually tearing them down — coordinating all three domains so the end-to-end slice actually meets its SLA.

The hard part is the cross-domain stitching. Each domain (RAN, transport, core) has its own manager and its own NSSIs; the orchestrator has to compose those into a coherent end-to-end NSI and keep its KPIs within target as load shifts. In 3GPP's management model this maps onto functions like the slice management function (NSMF) over per-domain subnet managers (NSSMFs), and it leans heavily on closed-loop automation and increasingly on intent-based and AI-assisted optimisation. Without it, slices are hand-built artefacts; with it, they're products you can spin up on request.