Mobile Access
Mobile access covers radio access architecture from baseband and massive MIMO through Cloud RAN and transport-facing interfaces. The exam tests how mobile services map to network slices, QoS flows, and centralized/distributed RAN functions, and how Nokia AirScale and Cloud RAN designs affect capacity and latency.
Radio and Cloud RAN architecture
Nokia AirScale baseband and ReefShark-based radios support 5G NR and massive MIMO. Cloud RAN splits baseband functions between a centralized unit and a distributed unit; the lower-layer split places real-time scheduling near the radio, while higher-layer functions can be centralized. Fronthaul capacity, latency, and synchronization become service architecture constraints. 5G QoS is flow-based, not bearer-based as in LTE; each PDU session can carry multiple QoS flows with different 5QI values.
Rules to apply
- A PDU session is the connectivity anchor; QoS flows inside it carry different treatment.
- Network slicing combines S-NSSAI, QoS policy, and transport isolation; a slice is not just a VLAN.
- NSA anchoring uses an LTE anchor for control-plane signaling; SA removes that dependency.
- Cloud RAN does not remove fronthaul; it changes where functions run and tightens latency and timing requirements.
- Peak cell throughput is a radio capability, not a committed per-user rate.
Worked example
A 100 MHz TDD carrier has 75 percent of resources allocated to downlink and an average cell spectral efficiency of 10 b/s/Hz. Average downlink capacity is roughly 100 MHz × 10 b/s/Hz × 0.75 = 750 Mbps. If 20 users each demand 50 Mbps, demand is 1 Gbps. The cell must add a carrier, use more MIMO layers, or offload to small cells. The reasoning: dimensioning uses average spectral efficiency and scheduling, not the peak modulation rate.
Traps that cost marks
Confusing QoS flows with PDU sessions; treating Cloud RAN as fronthaul-free; assuming peak MIMO throughput is a committed service rate; and ignoring synchronization and latency in split selection.