Analysis of the ITU-R model for forecasting IMT spectrum

• ITU-R Report M.2290 presents forecasts for growth in the total
amount of mobile traffic in the World to 2020 & then models
the spectrum demand for IMT services based on the forecast
• The model is relatively sophisticated & breaks down traffic into
different service environments
The model forecasts that between 1340 and 1960 MHz of
spectrum will be needed for IMT services by 2020
• Spectrum regulators place much importance on ITU reports,
therefore these values will drive decisions to allocate spectrum
for IMT services at WRC 15.
Stakeholders have reviewed the ITU-R
‘speculator’ model
• Model uses a bottom-up prediction based on population density & traffic
use to determine traffic densities in each of the service environments
• Benchmarking the population & traffic density figures in the model
against accepted forecasts of population & mobile data growth & usage
reveal traffic densities in the model are at least 100 times higher than
any realistic benchmarks
• This is based on unrealistic use of super-high-speed data services together
with excessive population densities. For example:
The urban population density used in the model is the equivalent of putting the
population of the USA into an area the size of Paris;
The population density that the model assumes for high mobility traffic (e.g.
that in vehicles) is the equivalent of that which would be found on a 500 lane
Other comments on the model can be found at:
Review of ITU ‘speculator’ model findings
Monthly data traffic
PetaBytes per square
km; the figures from
the low density
settings in the ITU
model compared
with benchmark
values for urban area
traffic in five example
Image illustrates the traffic that the model shows in each of the service environments, compared to
generally accepted forecasts for urban areas in several different countries.
(Note: scale is logarithmic; values from the ITU model are taken from the low demand setting)
Given the unrealistic values the model uses, it’s outputs cannot be applied to any
Review of ITU ‘speculator’ model findings
• Other values are also unrealistic: e.g. spectrum efficiency
• The model does not distinguish between different types of spectrum
E.g. It assumes that mobile traffic is handled in the same spectrum
as traffic in an office: in reality they would be handled using
different network solutions in different bands such as WiFi
The model can be significantly modified to produce more
valuable results
• Other organisations have also produced alternative spectrum
demand models
• The ITU model itself, is the most comprehensive & robust
Whilst the values in the model are unrealistic, the model itself
is a good basis for IMT spectrum calculations
Based on the analysis before, we can conclude:
• The ITU model itself is a good basis for calculating demand for IMT
spectrum but requires modifications (e.g. to consider how demand is
focussed in different frequency bands);
• The population density & traffic usage values in the ITU model are
orders of magnitude too high;
• Other inputs to the ITU model (e.g. spectrum efficiency) are also
• The model cannot be applied to any country for the purposes of
assessing IMT spectrum demand in that country.
Therefore, it is recommended that:
• To be a sound basis for responsible, international decisionmaking, the inputs to the ITU model need to based on realworld values;
• Regulators should not take decisions based on the existing ITU
model results - as the inputs are unrealistic, so equally are the
• Spectrum demand should not be based on values representing
the most densely populated area(s) in the world;
• Any country should be able to apply the model to its own
situation & obtain valid results if it is to take a decision based
on that model.