The growing network of connected devices, increasing number of Wi-Fi users, and growing data capacity requirements will necessitate careful Wi-Fi spectrum planning. To satisfy these rising demands in various regions, between 500MHz and 1GHz of spectrum will likely need to be added by 2025, according to a recent study by Quotient Associates and commissioned by the Wi-Fi Alliance. As many as 1.8GHz may be required to satisfy plausible upper bound scenarios. Not only does extra spectrum need to be added, but it will be important that contiguous spectrum, with wide channels of up to 160MHz, must be constructed.
The study took into account estimated data volume demands from three sources: the Cisco VNI forecast tool, surveyed data from Ofcom UK, and a commercial source, which largely agreed in a prediction of 1.125GB per person per busy hour by 2020, with four busy hours per day. Though few predictions exist beyond 2020, it’s expected that demand growth will slow between 2020 and 2025, which is supported by reports from many mobile operators. However, an unexpected new application may cause a spike in demand. With growth rates continuing as predicted, users will consume 4.5GB per person per busy hour in 2025.
There are three main environments of Wi-Fi usage, each of which presents different demands. These are offices, assumed to be multi-storey and open plan; residences, assumed to be relatively small with two occupants; and mall locations, which are assumed to be two-storey with large central atriums.
The increasing needs of Wi-Fi spectrum also concern the devices users use to access the internet, as well as their capabilities. People are increasingly using smartphones and tablets to access the internet, while usage of laptop and desktop computers has decreased. The Quotient Associates study concludes that, globally, smartphones have become the most important way to connect to the internet. Between 2013 to 2015, smartphone internet access more than doubled while both desktop and laptop use decreased significantly.
For this reason, it is important to determine current device capabilities. Current Wi-Fi 802.11ac devices supporting anywhere from 1X1 MIMO to 3×3 MIMO at the higher end. Projecting into the future, Wi-Fi 802.11ac will be replaced by 802.11ax, and high-end devices will support as many as 4×4 MIMO for a throughput of 2.4Gbps.
The study modelled two possible scenarios, a prediction for a likely busy hour (peak usage assumed to comprise four hours per day) as well as an upper bound scenario (2x the busy hour in 2020 and 4x the busy hour in 2025). While the predicted busy hour usage necessitates only a modest increase in spectrum, it is important to consider that the upper bound scenario is entirely plausible. To prepare for the possibility of unexpected increases in data usage, spectrum should be expanded to account for the worst-case scenario.
In 2020, the upper bound data requirement calls for 960MHz of spectrum, regardless of whether 80MHz or 160MHz channels are utilised. As wider channels enable higher data rates, adding the 960MHz of spectrum in contiguous 160MHz channels presents the best available option. In 2025, the upper bound scenario calls for 1920MHz of spectrum to be added. At this time, it may be beneficial to consider channels even wider than 160MHz, though this currently falls outside Wi-Fi 802.11 standards.
The study concluded that spectrum should be expanded in the 5GHz band, rather than the 2.4GHz and 60GHz options. This decision is one of practicality: the 2.4GHz band is limited in its growth capacity, while the 60GHz band will be unavailable to many devices, either because they cannot support it or will not be in range of a 60GHz AP. Despite the greater bandwidth available in the 60GHz band because of its higher frequency, it can not be considered a good substitute for 5GHz in normal circumstances.
Though there have been some attempts to extend the 5GHz bandwidth in both the United States and Europe, there are certain challenges in this band. For example, the Earth Exploration Satellite Service (EESS) is already in usage in the 535-5470MHz band. In addition, some 5GHz channels are subject to Dynamic Frequency Selection (DFS), which may decrease their usage compared to other channels. If 5GHz presents too many challenges or is not available in particular regions, the study suggests that spectrum in the 2-10GHz range may offer an appropriate alternative.
While the increasing demands of Wi-Fi usage present certain challenges, a consideration of future demands presents a good starting point for expanding spectrum needs. If Wi-Fi usage is to continue unimpeded, and its cultural and economic benefits to continue unrestricted, the spectrum must be scaled appropriately. Such expansion will also benefit emerging unlicensed technologies Multefire and carrier aggregation such as LWA.