- Fixed wireless access (FWA) will be one of the first 5G use cases.
- There are opportunities for operators to deliver broadband services to homes and small and medium-sized enterprises economically using FWA.
- The paradigms for fixed and mobile broadband are different, in both subscription offerings and dimensioning.
- When dimensioning an FWA network, several characteristics must be considered to uphold user experience during busy hour.
- FWA inherits dimensioning properties from mobile broadband; the data rate which users experience will vary with the position of the household relative to the base station site.
- An operator considering FWA for a particular area should assess factors such as density of households, plans to build out 5G coverage and service demand.
- With FWA deployed on top of an existing LTE or 5G mobile network, there are a number of technical synergies to be captured.
Around half of all households in the world – over 1 billion – do not have a fixed broadband connection. Given the current speed and capacity of cellular networks with LTE and its evolution to 5G, there are opportunities for operators to deliver broadband services to homes and small and medium-sized enterprises economically using FWA.
Devices with form factors suitable for FWA customer premises equipment (CPE), and without the stringent requirements on size, weight and power consumption that come with smartphones, will be among the first to reach the market.
For FWA to be a viable alternative to fixed broadband, including xDSL, cable and fiber-optic access technologies, it must be able to be dimensioned with comparable capacity and performance. While 5G will make this possible, there is also a range of markets to be addressed with LTE technology on the way to 5G.
In order to explore these opportunities, some questions need to be answered: can a mobile network handle FWA traffic and support the “unlimited data” model common to fixed broadband? And can it accommodate the anticipated growth in mobile broadband and FWA, in terms of both subscriptions and traffic per subscription over time?
Fixed-style subscriptions, mobile-style dimensioning
The paradigms for fixed and mobile broadband are different, in both subscription offerings and dimensioning.
Fixed broadband tariffs are commonly structured around a maximum data rate and include unlimited data traffic. The user traffic is often managed so that it does not exceed the maximum data rate for the subscription.
The mobile broadband subscription paradigm is dominated by monthly data traffic limits (“buckets”), and additional monetization is achieved through top-ups and upgrades to larger data buckets. For mobile broadband, the network normally transmits the maximum rate that the system resources and mobile device can handle.
FWA inherits the subscription paradigms of fixed broadband rather than those of mobile broadband; households will pay for FWA based on data rate and should not be concerned about traffic generated.
With FWA, the last hop is wireless, so all the characteristics of a wireless network apply. Unlike fiber, but similarly to xDSL loop length, connection quality will vary across households. And, unlike fixed broadband overall, the last hop is point-to-multipoint radio and therefore shared, which means that speeds will degrade with increasing cell load. All these characteristics must be considered when dimensioning an FWA network.
Focus on busy hour
FWA inherits dimensioning properties from mobile broadband, as the data rate which users experience varies with the position of the household relative to the radio base station, and generally decreases with the load of the shared radio resource.
First, households are situated at varying distances from a radio base station site, which results in different experienced data rates.
Second, the data rate differs during busy hour (when large numbers of users are active), compared to late at night or early in the morning (when there are fewer active users in the cell).
The figure below illustrates this combined behavior, showing how downlink data rates vary between households and decrease with load for users in different locations.
- The best-performing 5 percent of households (see the green line in the graph below) are in good geographic positions, with limited degradation even at high cell load.
- The average households (see blue line) have significant degradation at high cell load.
- The lowest-performing 5 percent of households (see purple line) are in difficult positions, with performance decreasing to critical levels during periods of high cell load.
The network should be dimensioned to maintain a given user experience under high load periods. This can be defined as the minimum data rate that a household could experience during busy hour.
It can, for example, be set to allow a standard-definition (SD) television stream, or multiple streams of SD or high-definition (HD) quality, depending on the addressed FWA segment. The minimum data throughput rate could be from 2Mbps up to 30Mbps, depending on the addressed segment. As the cell load increases and reaches busy hour, the lowest-performing 5 percent of households will finally reach the minimum rate, and the cell load at this point is defined as the capacity per site (as illustrated in the figure below). The minimum rate is only experienced by the lowest-performing 5 percent of households, during busy hour – and is sufficient to deliver all of the services described in the subscription agreement.
Analyzing FWA opportunities
FWA opportunities can look very different around the world, based on subscriber needs and expectations and availability of alternatives. An operator considering an FWA launch in a particular area should assess factors including the following:
- density of households in the target area
- availability and uptake of fixed broadband alternatives (xDSL, cable and fiber-optic access)
- service demand, including telephony, web access and television/video
- government subsidies
- LTE radio access network and infrastructure, including synergies between mobile broadband and FWA
- plans and/or activities to build out 5G coverage for other use cases
- availability of additional spectrum needed
The assessment should include market segmentation and targeting, creating a map of feasible coverage areas with subscription offerings constructed and priced around the resulting target segments.
As an operator’s network evolves from LTE through various stages of LTE-Advanced to 5G New Radio (NR), the cost of a delivered gigabyte declines and the amount of capacity increases – allowing the operator to address more opportunities.
FWA and mobile broadband
All the FWA opportunities outlined are included on the basis that they are dimensioned on top of an existing LTE or 5G mobile network built out primarily to serve mobile broadband subscribers. As well as the advantages of serving an additional customer segment, there are a number of technical synergies to be captured beyond the administrative efficiencies of common billing and support systems. These include network trunking effects, spectrum sharing gains, and combining businesses with different or offset busy hours, all of which improve capacity utilization.
A clear path to expansion
With the performance and capacity gains from enhanced mobile broadband and the evolution to 5G, FWA will be a feasible opportunity for communications service providers to deploy in many places. Previous experience from FWA and fixed broadband has shown that an “unlimited” traffic paradigm does not result in infinite demand and network congestion, but is manageable with a combination of performance-based service offerings and average consumption patterns.
Furthermore, analysis of a demanding example described here, the American suburb case, illustrates a clear path to capacity expansion by following a procedure of “utilize, add and densify”. First, network assets already in place should be fully utilized, including radio sites, spare capacity in deployed spectrum and associated radio, baseband and transport equipment. Next, spectrum and radio network capabilities should be added, such as higher-order modulation, advanced antenna systems and beamforming, increased sectorization and 5G NR access as needed. Finally, densify with the addition of macro and small cells when necessary.
Case study: An American suburb
The case depicts a suburban area in North America with LTE mobile broadband coverage, along with an initial 5G build-out using mid-range spectrum. On average, the area has 1,000 households per square kilometer and is covered with radio base stations that have an inter-site distance of 2,000 meters. Some fixed broadband offerings are available in the area, but there is little or no access to fiber to the home (FTTH). FWA is deployed on the sites to compete directly with fixed broadband by meeting the need for higher-rate offerings and capacity. Services to be offered include IPTV (allowing two 4K UHD video streams, or a combination of multiple SDTV and HDTV streams), internet access and IP telephony (also known as “triple play”).
As this case is demanding in terms of both capacity and performance, we include a description of its dimensioning and spectrum utilization based on simulations.
To reliably support the demands described above, a service provider must offer fiber-like speeds of 100–1,000+Mbps with a minimum data rate of 30Mbps and capacity to serve busy-hour usage per household of 3GB/h. Assuming 10 percent of total traffic occurs during busy hour, this corresponds to average monthly traffic of 900GB per household.
In order to deliver traffic with the expected performance, coverage can be densified so that, on average, a base station covers 550 households. With a projected 30 percent service uptake rate, the initial deployment will need to be dimensioned to serve 165 households per site. Households in the lowest-performing 5 percent should experience a downlink data rate of at least 30Mbps during busy hour.
The initial deployment consists of:
- undeployed band 10MHz FDD in sub-3GHz bands for LTE
- 40MHz TDD LTE in new 3.5GHz band (radios with 8 Tx/Rx including Multi-user MIMO)
- 400MHz TDD NR in millimeter wave (mmWave) band (Massive MIMO radio)
- selective densification on utility/light poles (5 poles per macro site)
- deployment of all 4 bands above 1GHz on the macro sites
- outdoor CPE focus to maximize performance and indoor CPE as complement
Step 1 is calculated to have throughput capacity of 1,330Mbps per site, allowing average busy-hour traffic of 3.6GB/h per household, which provides some headroom over the projected actual busy-hour consumption of 3GB/h.
The lightest blue region on the graph above represents Step 1. The “X” indicates the starting requirements: 30 percent of households within the coverage area and average busy-hour consumption of 3GB/h.
Dimensioning can evolve to handle increased capacity requirements – either additional connected households or higher usage per household. The following is an outline of how additional capacity can be provided.
Add 60MHz of TDD spectrum (new band), such as 3.7–4.2GHz, and increase bandwidth from 400 to 600MHzin the mmWave band
The mid-blue region of the figure indicates the capacity achieved in Step 2. The network could, for example, serve 30 percent of households with average busy-hour consumption of around 8GB/h. Alternatively, the higher capacity could be used to serve a higher percentage of households with unchanged average busy-hour consumption (or any combination between those points along the curve). Radio units supporting 5G NR are needed to handle the new sub-6GHz band.
Deploy an additional 40MHz of TDD spectrum, such as 3.7–4.2GHz, and increase bandwidth from 600 to 800MHz in the mmWave band.
The dark blue region in the graph shows what can be achieved in the third step. This configuration would enable the operator to serve 30 percent of households with an average busy-hour consumption of above 10GB/h, or alternatively serve 100 percent of the households with over 3.5GB/h or any combination of consumption and uptake along the curve in between.
Beyond this, additional ways to increase capacity are to exploit MU-MIMO in the mmWave band and/or to densify further with more sites.