In MOCN, operators share everything at the radio — the site, the gNB hardware and one spectrum carrier — and the cell broadcasts every operator's PLMN in SIB1. What stays separate is each operator's core: its own AMF, SMF/UPF (user plane) and UDM/AUSF (subscriber DB). A UE selects its home PLMN from SIB1 and the shared gNB routes it over NG to that operator's core.

MOCN isn't one feature — it's a stack. Identity + broadcast, then the resource split, then admission & QoS per operator, then load balancing, overload protection, slice/service steering and AMF selection. Every code below is a real Samsung FGR from the RAN Feature Description; this is the full toolbox you configure and trial.

MOCN is built in a strict order — identity first, then user plane, then the resource split, then mobility, then commit. Do it out of order and SIB1 or the per-PLMN routing breaks.

Put it all together on a concrete cell: one n78, 100 MHz TDD carrier (30 kHz SCS → 273 PRBs), shared 60/40 between two operators. Operator A takes the larger share; both run an eMBB slice to their own core. These are example PLMN codes — the engineering is exactly what you enter in LSM.

Set each operator's PLMN, TAC and PRB share plus the channel bandwidth — the LSM config, the SIB1 broadcast and the per-PLMN PRB allocation regenerate live. The whole page, made hands-on.

FGR-RS1801 gives each PLMN a controllable slice of the cell's PRBs via the per-PLMN distribution ratio (plmn-N-ue-distribution-ratio, 0–100 %). Drag the sliders to allocate the shared carrier between three operators — the split re-normalises live, exactly as the scheduler apportions PRBs.

Before a UE ever sends data, MOCN can bias where it camps — nudging idle UEs of a PLMN toward a less-loaded frequency via cell-reselection. It runs in manual mode (fixed ratios you set) or auto mode (the gNB weights by measured load), re-evaluated every few seconds and only once load crosses a trigger. This is the cheapest load control — no signalling, no HO.

Once UEs are connected, MLB moves load off a hot cell with a load-based handover. The gNB learns neighbour load through resource-status reporting over F1/Xn; when the serving cell crosses its load threshold it hands selected UEs to a lighter neighbour. Evaluate by resource (PRB) or user (UE count).

IFLB works through reselection, so these thresholds decide which cells an idle UE even considers. Tighten them and camping stays local; loosen them and load spreads wider.

Redirection moves a UE at connection release (or fallback) to a target frequency/RAT; SPID/RFSP-based steering lets you route by subscriber profile — useful to give an operator's premium SPID a preferred layer.

Distribution ratios share the average; admission & overload protect each operator at the edge. Per-PLMN hard caps stop one operator monopolising context, CRP (Call Retention Priority) decides who survives preemption, and overload protection sheds new load gracefully instead of dropping calls.

This is the heart of the day. Default = the exact factory value from the Samsung SVR25A Parameter Description (5G CU, 2026); Recommended = the golden-parameter starting point for a shared MOCN cell. Defaults are correct as shipped; recommended values are engineering guidance you always trial per Day 6 (baseline → single change → 48 h soak → verify → roll back).

FGR-MC1015 makes the measurement / report configuration per PLMN, so each operator's UEs hand over and reselect using that operator's own thresholds, neighbour list and frequency priorities. Every value below is exact from the Samsung Parameter Description (SVR22B NR). RSRP thresholds use the 3GPP 0–127 index (RSRP −156…−31 dBm, so 51 ≈ −105 dBm).

The whole point of MOCN operations is that every KPI is per-PLMN. Samsung's per-PLMN counter family 5583 breaks throughput, loss and drop down by CELL · PLMN · 5QI · S-NSSAI — so you can prove each operator's SLA independently.

Give each operator enough to meet its SLA, but keep the carrier statistically multiplexed so idle share is reused — a hard split (MORAN-like) wastes spectrum, a too-soft split breaks SLAs. The optimum is per-PLMN CRP floors + a distribution ratio tuned to each operator's real load.

Two operators, one carrier. Set the contracted ratios (plmn-N-ue-distribution-ratio), each operator's offered load and the cell capacity — the studio applies the RS1801 rule (portions enforce under contention; unused share lends), grades each SLA, and runs the blame router: whose problem is a bad operator KPI on a shared cell?