NSI lifecycle, GSMA NG.116 templates, enterprise slice onboarding, day-2 operations, slice SLA management, monetisation
1. What Is Slicing Operations — The Simple Version
Network slicing is where 5G monetisation happens. An operator can sell consumer broadband on one slice and a guaranteed-latency industrial automation service on another — using the same physical infrastructure. But delivering this commercially requires more than configuring S-NSSAIs in NF software. It requires an operational model: how do you onboard an enterprise customer to a slice, how do you monitor their SLA in real time, how do you respond when the RAN slice degrades during busy hour?
This article covers the end-to-end slicing operations model — from GSMA NG.116 standard templates to the day-2 operational playbook.
| 3GPP Reference |
| 3GPP TS 28.530 — Management and orchestration of networks and network slicing |
| 3GPP TS 28.541 — 5G NRM (Network Resource Model) for slicing |
| GSMA NG.116 — Generic Network Slice Template (standard commercial slice parameters) |
| 3GPP TS 28.550 — Performance assurance for network slices |
2. GSMA NG.116 Slice Templates
GSMA NG.116 defines standard slice templates that operators can use as a starting point for enterprise SLAs. Using standard templates reduces negotiation time and enables automated slice provisioning:
| Template | SST | Key SLA Parameters | Target Industry |
| eMBB Enterprise | 1 + custom SD | DL/UL throughput guaranteed; moderate latency; high reliability | Office connectivity, retail, hospitality |
| URLLC Industrial | 2 + custom SD | Latency < 5ms E2E; 99.9999% availability; PRB reservation | Factory automation, robotics, AGV |
| mIoT Massive | 3 + custom SD | 10k+ devices/km²; low power; relaxed latency | Smart meters, logistics, agriculture |
| V2X | 4 + custom SD | < 3ms V2X latency; near-zero packet loss; geographic continuity | Connected vehicles, smart road infrastructure |
| Emergency/MCPTT | SST 3 (3GPP) | Priority access; guaranteed capacity even under congestion; pre-emption | Emergency services, government |
| Private 5G (custom) | 1 or 2 + SD | Dedicated UPF, data sovereignty, custom SLA per enterprise | Oil & gas, ports, utilities, defence |
Table 1 — GSMA NG.116 slice templates (simplified). Real deployments customise these templates per enterprise requirement, but standard templates enable faster sales cycle and automated provisioning.
3. NSI Lifecycle — From Onboarding to Termination
A Network Slice Instance (NSI) lifecycle has four phases. Most operators have the Day-0 and Day-1 phases working. Day-2 (operational management) and Day-N (scaling and termination) are where most gaps exist in first-generation SA deployments.
| Phase | What Happens | Key Systems Involved | Common Gap |
| Day-0: Design | Translate enterprise requirement into GSMA NG.116-based template. Define S-NSSAI, DNN, QoS profile, UPF placement, SLA parameters. | OSS/BSS, GSMA NG.116 template library | No standard template: every enterprise is a custom design, slow sales cycle |
| Day-1: Provision | Configure NFs: UDM subscription, NSSF TA policy, SMF UPF pool mapping, PCF PCC rules, gNB PRB reservation, transport QoS. | NSMF (Network Slice Management Function), NF configuration APIs | Manual configuration across 6+ NFs: error-prone, slow. Automation gap. |
| Day-2: Monitor & Assure | Continuous SLA monitoring vs per-slice KPIs. Alert on breach. Corrective actions: scale UPF, increase PRB, reroute traffic. | Prometheus/Grafana per-slice dashboards, NSMF autoscaling | No per-slice KPI dashboard: SLA breach discovered from enterprise complaint, not proactive monitoring |
| Day-N: Scale & Modify | Modify SLA parameters (increase AMBR, add TA coverage). Scale NSI (add UPF capacity). Terminate NSI at contract end. | NSMF, CI/CD pipeline for NF config changes | No NSMF automation: manual changes require ticket to core team, 2–5 day lead time |
Table 2 — NSI lifecycle phases. Most GCC first-generation SA deployments have Day-0 and Day-1 working. Day-2 monitoring and Day-N automation are the operational maturity gaps.
4. Enterprise Slice Onboarding — Practical Steps
Here is the actual sequence when onboarding an enterprise customer to a private 5G slice:
Step 1 — Requirements capture: enterprise provides use case (factory automation, office connectivity, etc.), device count, geographic coverage area (which buildings/sites), latency SLA, throughput SLA, data sovereignty requirement (on-premises UPF or central?).
Step 2 — Slice design: select GSMA NG.116 template (URLLC industrial for factory, eMBB enterprise for office). Define S-NSSAI (assign SD value from operator pool). Choose UPF placement (on-premises for URLLC and data sovereignty, edge DC for enterprise office). Design TAI coverage map.
Step 3 — NF configuration (automated via NSMF, or manual via ticket in less mature deployments): (a) UDM: add S-NSSAI to enterprise subscriber SUPI range; (b) NSSF: add TAI list for enterprise coverage area to new S-NSSAI policy; (c) SMF: add UPF pool mapping for new S-NSSAI and DNN; (d) PCF: create PCC rules for enterprise slice (5QI, AMBR, charging profile); (e) gNB: configure PRB reservation for new S-NSSAI; (f) transport: configure QoS queues for N3 traffic with enterprise DSCP marking.
Step 4 — Test: pilot device connects and attempts PDU session with enterprise S-NSSAI. Validate: correct UPF selected, correct AMBR applied, latency meets SLA, data stays within enterprise boundary (for on-premises UPF).
Step 5 — Go-live: migrate enterprise devices to new SIMs (or push S-NSSAI via OTA). Monitor per-slice KPIs (registration SR, PDU SR, latency, throughput) from day one.
5. Slice SLA Management
| SLA Parameter | How to Monitor | Alert Condition | Corrective Action |
| Latency (E2E UE→App) | Active probe from UE to app server via TWAMP or synthetic traffic | P99 > SLA threshold | Check: UPF placement, N3 latency, PRB scheduling priority, N6 path. Scale edge UPF if needed. |
| Throughput (guaranteed) | Per-slice UPF throughput counter + active throughput test | < 90% of guaranteed throughput sustained > 5 min | Scale UPF capacity; check Session-AMBR config; verify DSCP not downgraded at aggregation |
| Availability | PDU session establishment SR per S-NSSAI | < 99.9% (or per-contract threshold) | Immediate: check UPF health, SMF N4 status. Escalate to SLA breach procedure if sustained. |
| Device capacity | Active PDU sessions per S-NSSAI vs committed device count | > 85% of committed device count | Provision additional capacity (SMF session table, UPF memory) before 100% hit |
| Data sovereignty | Alert if enterprise DNN sessions appear on non-enterprise UPF | Any session | Immediate: check SMF UPF pool mapping; check for TAI gaps. This is a compliance breach. |
Table 3 — Slice SLA monitoring. Data sovereignty alert is the only zero-tolerance item — any breach requires immediate investigation regardless of SLA threshold.
6. Common Issues in the Field
| Field Note: NSSF Not Updated for New Enterprise Site — Devices Cannot Connect |
| Enterprise customer expanded factory to new building. New gNB commissioned with new TAI. |
| NSSF policy for enterprise S-NSSAI had TAI list not updated with new building TAI. |
| Factory devices in new building: NSSF returned empty Allowed NSSAI (enterprise S-NSSAI not permitted in new TAI). |
| Devices fell back to public eMBB slice. Enterprise OT traffic on shared infrastructure. |
| Discovered 2 days later. Data sovereignty concern raised by enterprise security team. |
| Prevention: automate NSSF TAI update as part of gNB commissioning workflow. |
| Field Note: PRB Reservation Not Configured — URLLC SLA Breached During Shift Change |
| Industrial automation slice (SST=2) deployed for smart factory. Core slice correct: dedicated on-prem UPF. |
| SLA: P99 latency < 3ms. Normal operations: P99 = 1.8ms. |
| 200 workers connecting smartphones at shift change on shared eMBB slice (same gNB). |
| Without PRB reservation: URLLC DRBs competed for scheduler slots. P99 latency hit 28ms. |
| SLA breach. Enterprise raised complaint. Investigated 4 hours. |
| Fix: configure minimum PRB reservation (20% bandwidth) for URLLC S-NSSAI on factory gNBs. |
7. Summary — Key Takeaways
| Topic | Key Takeaway |
| GSMA NG.116 | Use standard templates as starting point. Accelerates sales cycle. Enables automated provisioning. Customise per enterprise, not from scratch per customer. |
| Day-2 operations | The gap in most first-gen deployments. Build per-slice Grafana dashboards before go-live. SLA breach must be detected by operator, not enterprise complaint. |
| NSSF TAI coverage | Add enterprise site TAIs to NSSF policy as part of gNB commissioning — not as an afterthought. Automate this step. |
| PRB reservation | Include in enterprise URLLC SLA design, not just core slice isolation. Without PRB reservation: SLA holds at low load, fails at busy hour. |
| Data sovereignty monitoring | Alert on any enterprise DNN session assigned to non-enterprise UPF. Zero tolerance. Compliance breach. |
| NSMF automation | Manual onboarding across 6+ NFs is slow and error-prone. Build NSMF automation for Day-1 provisioning to enable faster enterprise sales cycles. |
Table 4 — Post 17 summary. Slicing is both a network architecture and an operations model. Day-2 monitoring and NSMF automation are the keys to commercial scale.
Next: Post 18 — NEF Integration & New Service Onboarding