5G capacity dimensioning is not simply a matter of estimating traffic and adding cell sites. Network planners must also consider how radio technologies interact, how users find and connect to cells, and how shared resources are allocated under different load conditions. Analyses of heterogeneous networks (HetNets), millimeter-wave (mmWave) communications and non-orthogonal multiple access (NOMA) show that these technologies can support higher capacity, but their benefits depend on implementation choices.
For engineers, system integrators and network decision-makers, the practical lesson is to evaluate capacity as a system-level outcome. Radio performance, access procedures, interference, user distribution and site density all affect the capacity a network can deliver.
5G capacity dimensioning starts with the network architecture
Mobile network capacity can be increased by adding spectrum, improving spectral efficiency or deploying more cell sites. In dense areas, relying only on additional macro sites can become difficult because suitable locations may be limited. A HetNet offers another approach: deploy lower-power small cells within the coverage area of macro cells to add capacity in hotspots, at cell edges or in coverage gaps.
Small cells shorten the distance between a user and a base station and can offload traffic from the macro layer. However, adding nodes also changes the interference environment. In dense deployments, a user may receive signals from multiple nearby cells without one clear, dominant interferer. Interference coordination and, where applicable, coordination between neighboring cells therefore form part of the capacity-planning problem rather than an afterthought.
Network architecture also affects how resources are used. In a control- and user-plane split, the macro layer can provide broad-area control and mobility support while small cells provide high-capacity data connections. Dual connectivity can keep a device connected to the macro layer if its small-cell connection fails. These arrangements can support flexible deployments, but require coordination between layers and careful consideration of which users each layer serves.
mmWave capacity depends on effective cell discovery
Millimeter-wave spectrum can contribute to high data rates, but its use is associated with highly directional transmissions. That directionality changes a basic network operation: initial cell search. A device must discover a suitable cell and establish access before it can benefit from a high-capacity link.
Search procedures designed around less directional microwave transmissions may not translate directly to mmWave small cells. The research examined a Joint Search Method in which macro and small cells cooperate during cell discovery. In the evaluated scenarios, this approach outperformed the generic search methods considered in terms of search overhead and initial access delay. The broader planning implication is that access procedures should be included in deployment assessments: a capacity estimate that ignores the effort and delay involved in finding a cell may not represent the full user experience.
HetNet cooperation can also matter when considering access and capacity together. A macro layer can assist with coverage and mobility while small cells add localized capacity. The value of this arrangement depends on the way the layers coordinate, not just on the radio capacity available at each site.
NOMA capacity gains require interference management
NOMA allows multiple users to share radio resources non-orthogonally, unlike orthogonal multiple access (OMA), which separates users across resources. This can improve resource utilization, but users sharing a resource experience intra-cell interference. If that interference is not accounted for, it can reduce the expected capacity benefit, especially in systems where NOMA and OMA operate together.
The evaluated work identifies several implementation considerations:
- Adjust transmission parameters: Modulation and coding scheme (MCS) adjustments and additional transmission power for NOMA users were proposed to help mitigate the effects of intra-cell interference.
- Select user pairs carefully: Pairing decisions directly affect NOMA capacity. A cost-matrix approach was proposed to assess possible pairs according to the capacity gain they offer.
- Adapt pairing to network load: Maintaining a pool of pairing methods can let the network choose an approach suited to current cell conditions, balancing algorithm complexity and efficiency against capacity.
The results also indicate that the relationship between mmWave beamwidth and NOMA pairing deserves attention. Beamforming can improve channel conditions, while narrower directional coverage can reduce the likelihood of finding users suitable for pairing. Planners must consider both effects when assessing a combined mmWave and NOMA deployment.
Load balancing can improve resource use across a HetNet
Capacity is shaped by how demand is distributed between cells, not only by the total number of sites. A heavily loaded cell may have neighboring cells with available resources. The proposed Load Balancing NOMA (LB-NOMA) approach addresses this imbalance by moving OMA users at the edges of congested cells to neighboring cells, with the aim of increasing the share of NOMA users in the congested cell and making better use of available network resources.
This approach highlights why load balancing should be considered alongside radio-layer capacity. A network may have sufficient aggregate resources yet still experience local congestion if users and traffic are unevenly distributed. Strategies designed specifically for NOMA can affect how effectively those resources are used.
Practical implications for network planning
Capacity dimensioning for a 5G HetNet combining mmWave and NOMA should account for more than peak radio rates. Planning and evaluation should consider:
- Expected demand and user distribution across macro and small cells.
- Cell discovery overhead and initial access delay for directional links.
- Interference and the coexistence of NOMA and OMA users.
- User-pairing performance under different cell-load conditions.
- Load balancing between congested cells and neighbors with spare resources.
- Site counts and capacity estimates alongside the quality of service users can receive.
The underlying performance analyses and capacity calculations were based on computational simulations; their findings should therefore be read in the context of the scenarios and assumptions evaluated, rather than as universal guarantees for every deployment.
Conclusion
HetNets, mmWave and NOMA offer complementary ways to address growing mobile data demand, but each introduces planning requirements that affect the others. Small cells can add localized capacity, mmWave requires suitable directional cell-discovery procedures, and NOMA depends on managing interference and selecting users effectively. Load balancing can help address uneven demand across cells. For network planners, the practical priority is to dimension the system around these interactions, using scenario-based evaluation to connect site deployment and resource allocation with access performance and quality of service.