SAMENA response to CITC consultation on spectrum management
1. 1
SAMENA Response to CITC on a “National Spectrum Strategy
2025” (5/1440)
April 2019
Introduction
SAMENA Telecommunications Council welcomes the opportunity to respond to this
important consultation in the Kingdom of Saudi Arabia. As outlined below SAMENA
Telecommunications Council is an industry association that represents the interests of our
members in the region. Our members include the current mobile operators in KSA, namely
STC, Mobily, and Zain.
In consultation with them SAMENA has prepared this response to CITC’s preparatory
questions on developing a national spectrum strategy for 2025. SAMENA and its members
believe that spectrum is a key resource in public policy objectives such as the Digital
Society, Smart Cites, etc. as well as helping to boost economic growth. Many citizens now
seem to prefer accessing many government services electronically, with mobile devices
being a preference for the younger generation.
The growth in demand for spectrum for mobile services, amongst other competing demands,
places increasing pressure on this scarce resource needed by the private and public sectors.
This makes such a review on a future spectrum strategy timely and important.
SAMENA believes that it is important that KSA/CITC maintains (and if possible, increases)
its reputation as a global though leader in spectrum. This is to help bring about the beneficial
harmonisation of spectrum use to:
1. reduce harmful cross-border radio interference; as well as
2. maximising the benefits to citizens in KSA (and beyond) of economies of scale.
The upcoming WRC-19 in Egypt potentially offers a significant opportunity for CITC to
increase its influence, and actively engage as an international thought leader in spectrum
issues.
SAMENA believes that an active role is required by KSA/CITC in October at the WRC-19 to
ensure the best opportunity for mobile to support the 2030 Vision and its objectives.
For questions regarding this paper please contact:
Mr Roberto Ercole CEng, Director Public Policy, SAMENA,
roberto@samenacouncil.org
#304, Alfa Building, Knowledge Village, P.O. Box: 502544, Dubai, UAE
2. 2
About SAMENA
SAMENA Telecommunications Council1
is tri-regional not-for-profit industry association
spanning more than 25 countries, including Afghanistan, Algeria, Bahrain, Bangladesh,
Egypt, Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Nepal, Oman, Pakistan,
Palestine, Saudi Arabia, Sri Lanka, Sudan, Syria, Tunisia, Turkey, United Arab Emirates,
and Yemen. It represents the interests of more than 85 telecom operators and service
providers in the fixed and mobile space, and stakeholders from the wider digital ecosystem,
including technology-, equipment- and software manufacturers, internet companies,
consulting companies, academia and regulatory authorities.
It is SAMENA’s mission to serve as a sector-development partner to governments and
industry for the joint creation of a flourishing and sustainable ICT sector to enable
sustainable digital transformation. Our key objectives are to enable sustainable growth,
incentivize investments and broaden value creation through effecting adoption of new
regulatory approaches in the areas of Digital Services, Data Regulation, Spectrum
Management, and Industry Fees & Taxation.
Answers to Specific Questions:
Question 1: What are the current or expected challenges related to using spectrum today?
The main challenge in spectrum management is to ensure that the goals and objectives (as
set out in the ITU 2015 Handbook on National Spectrum Management) of a national
regulator are met as efficiently as possible. That is making radio spectrum available for
government and non-government users to stimulate social development and economic
progress. It is therefore important that spectrum be viewed as a scare resource for the
benefit of the economy in general and citizens and government.
It is the growing demand for spectrum that makes national spectrum management much
more challenging than it has been in the past. In the past it was possible to control
interference by assigning relatively large blocks of spectrum as “guard bands” that could be
held effectively “fallow”. It was also possible to hold large pieces of spectrum in reserve to
cope with new demands or services. However, the pressure today on the demand for
spectrum makes that very difficult if a national regulator is to make the best use of a scare
resource like radio spectrum. It therefore requires a more strategic approach as to how to
use spectrum, than was perhaps required in the past (when there was ample spectrum for
demand).
Question 2- Improving Spectrum Management
SAMENA believes that in the short term commercial mobile operators would benefit from the
ability to share the spectrum licenced to each other with their permission.
It is likely that at times, in some areas, all mobile opcos would benefit from such an
arrangement and can control interference problems by coordination amongst themselves.
This is especially true as most mobile base stations and handsets tune across all the
frequency ranges assigned for mobile in KSA. This means such an arrangement can be
1 https://www.samenacouncil.org/index
3. 3
very ad hoc and time limited as required. This has the potential to solve short term capacity
bottlenecks in an efficient manner.
Whilst other more exotic schemes (such as white space and dynamic spectrum access) may
be suggested, in practice they have not proven to be commercially successful and would
seem to have little potential in the short to medium term.
Question 3 - In what ways can KSA ensure spectrum management safeguards for government
use whilst allowing commercial to prosper?
SAMENA believes that new and innovative approaches will be needed to try and “square this
circle”. In the past when spectrum demand was relatively low it was possible to ensure that
all interested parties could be reserved sufficiently large blocks so that any government
department or agency could run its own individual radio network. In the future it may make
sense to consider:
1. Government agencies use a single network shared between them rather than
having separate networks; and or
2. That government agencies might use commercial networks using enhanced and
guaranteed QOS, such as via network slicing.
The advantage of these approaches is that it reduces costs by avoiding duplication. It is likely
that option 2 (using commercial networks will deliver the biggest savings and hence most
beneficial). It is also more spectrally efficient to have a single larger network than many small
ones. This is due to factors such as trunking gain and being able to get access to better located
sites (fewer networks chasing limited mast sites). This is especially true as we move towards
broadband radio networks. Whereas before finding 12.5 kHz channels for push to talk voice
was relatively easy for utilities and government users, this is no longer the case.
We are seeing a move towards networks that need 10 MHz plus to take advantage of new
multiple access schemes such as OFDMA. This coupled with the ever-higher data rates
required by users makes the multiple small independent network approach harder and harder
to implement. This inefficiency reduces when one considers the large networks operated by
the current commercial operators.
Increasingly government users are seeking to benefit from economies of scale and wish to
use commercial mobile standards (such as LTE) as far as possible. This means that the
spectrum ranges available become even narrower as these government users need access
to commercially harmonised bands to benefit from these economies of scale. The danger
becomes not only of increased costs but having to coordinate lots of independent smaller
spectrum networks in the same geographic area. This may lead to spectrum being effectively
lost due to “guard band” requirements.
SAMENA believes that the long-term solution to this is to try and encourage government
users to share a single managed network (run by a government agency or a commercial
contractor) or to use a commercial 5G network. It is likely to be cheaper for a commercial 5G
network (using network slicing perhaps) to upgrade their network QOS, than for a
government user building a new network from scratch.
4. 4
Question 4 – the role of KSA generally, and CITC in particular in representing local
stakeholder interests internationally
SAMENA believes that it vital that CITC continue its role in international meetings and
regional groups regarding spectrum. Because of the size of the economy and thought
leadership in many areas related to spectrum. From a mobile commercial viewpoint there
are two key factors that require CITC to play an influential role in spectrum matters, namely:
1. Harmonise bands and use to gain from economies of scale for mobile devices and
equipment: and
2. To prevent/control/reduce harmful cross border interference.
SAMENA believes that CITC has and will continue to play a key role in the above issues. One
example to date has been the 2.6 GHz band that has been made available as all TDD (3GPP
band 41). This is recognised by many as a better option for flexibility and 5G deployment. As
SAMENA understands it, CITC has been successful in persuading a number of other
neighbouring administrations to change from adopting the primarily FDD (band 7) plan, used
in Europe for 4G.
The issue is that the two band plans give rise to difficult cross-border coordination problems
(for more detail see reference2
). It appears that Bahrain has now changed its proposal to use
primarily FDD to follow the KSA band plan. This will potentially save much time and effort in
coordinating base stations for 5G services to the benefit of the economies and consumers of
both KSA and Bahrain. However, other administrations such as Jordan and Kuwait still use
Band 7, and others are considering it. In SAMENA’s view all bands (700/800/900 L-Band/
1800/2100/2300/2600/C band) would benefit from a framework of cross border coordination
agreements.
Whilst this is only one example, there are other issues to do with cross-border interference
that the leadership and influence of KSA/CITC are important for helping to resolve. For
example, the core 5G capacity band (3.4-3.8 GHz) is not fully recognised and protected under
the ITU Radio Regulations in this region. This means that 5G deployments in 3.6 to 3.8 GHz3
may be much more complex (and hence expensive) because of the international regulatory
rules. This also gives rise to the potential that networks once deployed will need to be (in the
worst case) switched off in border areas to protect neighbouring countries.
SAMENA understands that CITC has used its influence to try and resolve this at the Arab
Spectrum Management Group (ASMG) level. However, not all KSA neighbours are in the
ASMG, which may lead to future difficulties. SAMENA would ask CITC to continue its efforts
in this area to seek a fair and balanced resolution to this, if possible, via the ITU Radio
Regulations.
2 www.slideshare.net/RobertoErcole/45g-crossborder-coordination-in-bahrain-e-field-trigger-
levels
3 Mobile services have secondary protection in 3.6-3.8 internationally in the ITU Radio
Regulations.
5. 5
Question 5- Factors affecting spectrum demand for commercial mobile
SAMENA welcomes the greater emphasis being placed on mobile services and the drive
towards 5G. According to our data KSA has moved from less than 300 MHz available for
commercial mobile in early 2017 to just over 1000 MHz today (after the recent C band
awards), which places KSA first in the region for mobile spectrum.
*** Source: Policy Tracker ** including 800 and 2.6 bands under consultation
*assumes 3.6-3.8 available.
Figure 1
This increase in mobile spectrum has not only helped put KSA in prime position to deploy
cost effective 5G services (with capacity and coverage bands available), it has helped drive
an increase in data speeds available to consumers. According to data presented at the
recent ITU workshop in Riyadh (December 18), average 4G download speeds increased
from 9.3 to 26.3 Mb/s from 2017 to 2018.
SAMENA and the mobile industry welcomes this strategic focus on making spectrum
available for commercial mobile systems but considers that mobile data use will continue to
grow. The current 1000 MHz for mobile does not match the demands calculated at ITU
level. The ITU (M.2290 from December 13) suggested that the total spectrum requirements
for terrestrial IMT would be from around 1400 MHz to 2000 MHz by the year 2020.
However, this was based on assuming traffic would grow from the 2010 level by 80 times by
the year 2020.
According to Cisco, global mobile traffic was 237 Petabytes per month in 20104
. Also,
according to Cisco, the global mobile traffic was 19 Exabytes per month in 20185
, which is
80 times more than the 2010 figure. This means that the mobile data level projected as high
in M.2290 for the year 2020, has already been reached. Cisco predicts that by 2020 the
4 https://blogs.cisco.com/ar/cisco-vni-mobile-data-traffic-forecast-2010-%E2%80%93-2015
5 https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-
vni/white-paper-c11-738429.html
1,010.0
720.0 700.0
625.0
528.8 480.0
358.4 330.0 320.0 310.0
220.0 210.0 189.0
120.0
0.0
200.0
400.0
600.0
800.0
1,000.0
1,200.0
Commercial Mobile Spectrum (MHz)
6. 6
mobile data usage per month will be 41 Exabytes, which would be 170 times the 2010 data
usage level.
Of course, it is not possible to do a simple one to one mapping to suggest that 2000 MHz
might be needed by 2020 in KSA. The numbers used in the ITU study where not for specific
countries, and technology has improved since 2014. Also, IMT-2020 is supposed to be 3x
more spectrum efficient than IMT-advanced.
In fact (according to Cisco) in 2017 the monthly mobile traffic figure was 59 Petabytes in
KSA6
. Which equates to around 2 GB per month per user in KSA, that places KSA slightly
below the OECD average (see below) but on a par with Spain and Holland. One would
expect the demand for mobile data to increase in KSA towards Korean and the Nordics data
usage over time.
Figure 2
It is not possible for SAMENA to give any exact figures on the future demand for mobile
spectrum in KSA, but based on the above, it is likely that more spectrum will be required at
some point after the year 2020 perhaps. The availability of 3.8 to 4.2 GHz may be part of
the solution.
This could also be affected by a number of other factors. One important one might be the
amount of spectrum for low power radio LANs such as Wi-Fi. The US and EU7
are
considering more spectrum for such services in 5925 – 7125MHz and 5925 – 6425 MHz
respectively. Wi-Fi is used by many operators for traffic offloading and may carry around
50% of mobile operators traffic in some cases.
6 https://www.cisco.com/c/dam/m/en_us/solutions/service-provider/vni-forecast-
highlights/pdf/Saudi_Arabia_Consumer_Highlights.pdf
7 http://ec.europa.eu/newsroom/dae/document.cfm?doc_id=50343
15.45
5.57
2.94
0
2
4
6
8
10
12
14
16
18
GB per month in 2017, per mobile broadband subscription - Source: OECD
7. 7
Question 6 – overall expectation on current use in terms of coverage
SAMENA will limit its comments to the use of coverage requirements and obligations set on
commercial mobile services.
Coverage Requirements
SAMENA believes that mobile broadband coverage is important and that every effort should
be made to try and increase the geographic coverage and data rates/quality of service
available to consumers in KSA. Some believed that this could be done by relying on market
forces and competition alone. Those views have changed, and it has been recognised that
competition is not the complete answer.
UK Ofcom in its December 18 consultation on 5G spectrum bands has stated that “targeted
public policy interventions may be needed to deliver services to the hardest-to-reach areas”8
.
It noted that commercial investment driven by competition had reached its upper bounds at
around 84% of landmass (para 4.21).
The Ofcom paper undertakes a cost benefit analysis on how such obligations should be set.
Their analysis suggested a geographic coverage obligation of 90% UK landmass would
deliver the social benefits at least equal to the underlying costs (para 4.54). The costs were
substantial and noted that for the UK the cost of moving from 82% to 92% of landmass was
£305m to £535m (figure 4.3). Given the UK is a much more densely populated country (and
smaller) than KSA (263 vs 15 people per km2
), the costs are likely to be higher in KSA – as
more cells will be required (these extra cells will likely generate little income).
Whilst ensuring that user experienced data rates and the level of coverage are valid public
policy objectives, and indeed SAMENA operator members invest heavily in trying to boost
this, there is an associated cost. It is important that as with all public policy objectives there
should be a consideration of the costs and benefits associated with improving coverage and
data rates.
According to TowerXchange9
there are nearly 36,000 towers used in KSA by mobile opcos.
This represents a substantial cost in terms of initial investment as well as ongoing operating
costs. The 5G investment being made by the operators is multi-billion $ to upgrade existing
sites as well as for site densification.
The way a coverage and user experience data rate (or a proxy for this such as field strength
or average data rates etc) is set can have a major impact on network costs. This can be a
particular problem in very large countries such as KSA with a low population density. This
can lead to many more sites being built than can be justified in any cost benefit analysis.
The issue holds true not just in terms of landmass covered, but also required data rates (also
if this should be supported indoors or not). This is because the data rate required impacts
on the required link budget towards the cell edge, which can reduce the size of cells quite
significantly.
According to information supplied by PolicyTracker10
85% of the population in KSA lives in
just over 25% of the landmass (around 500,000 km2
). It is thus possible to cover most of the
population with a combination of capacity bands (above 1 GHz) and coverage bands (below
8 www.ofcom.org.uk/__data/assets/pdf_file/0019/130726/Award-of-the-700-MHz-and-3.6-3.8-
GHz-spectrum-bands.pdf
9 www.towerxchange.com/wp-content/uploads/2019/01/TX_MENAProspectus_2019.pdf
10 www.policytracker.com/
8. 8
1 GHz). However, for the remaining 15% of the population it is necessary to rely on lower
frequency coverage bands such as 700 and 800 MHz.
To go from 85% to 95% population coverage, increases the landmass coverage from 25% to
61%. This is a major cost in a country the size of KSA. This means operators need to build
enough cells to cover and extra 36% landmass or nearly 800,000 km2
. That could equate to
around an extra 5,000 sites per opco, assuming an average 7 km radius for rural cells. If the
way the QOS were set reduced the cell size to say 5 km that would mean around 10,000
sites would now be required to meet this new QOS.
The relationship between population and landmass for this remaining 15% of the population
in rural areas is not linear. That is, it requires ever more cells to reach the next 1% of
population as one tries to cover more and more of this rural 15%.
It should also be considered that the amount of spectrum is more limited in these coverage
bands below 1 GHz (comparing downlinks of 30 MHz and 30 MHz in 700 and 800 MHz)
compared to C band with 100 MHz per opco (TDD). There are also different spectrum
efficiencies (see below) to do with urban and rural deployment scenarios. This may mean
the data rates may be more limited in some circumstances (in terms of cell peak or average
rates) in these rural areas relying on sub-1GHz bands.
The spectrum efficiency of rural cells will be less than that of dense urban sites (due to less
inter-cell interference and other factors). For example, ITU-R M.2410 (table 1) shows a
downlink spectrum efficiency11
of 0.12 bits/s/Hz for rural enhanced mobile broadband
(eMBB), compared with 0.225 bits/s/Hz for dense urban (a reduction of around 50%). To
equate this to user experienced data rates (download) suggested using 100 MHz in C band
(in dense urban areas) would give 22.5 Mb/s. For 700 and 800 MHz rural eMBB (20 MHz
downlink per opco) would be 2.4 Mb/s. This is not to say this is what would happen in
practice, and users would experience higher average data rates. But it shows how sensitive
the results are to how one sets the metrics and measures of such obligations. SAMENA
believes careful consideration should be given to the metrics used and the way they are
measured, preferably after an industry dialogue.
A paper has been produced by PolicyTracker (with the author) that looks at how coverage
and quality of service obligations can be evaluated looking at a four-stage process12
that
goes into more details. It also looks at the connection between cell sizes and proposed
QOS levels.
SAMENA believes that as it is important that any coverage and quality of service obligations
be set after a cost benefit analysis has been undertaken. This is to ensure that the social
benefits of increasing coverage/quality of service match the increased costs to the networks.
11 5th percentile user spectrum efficiency that relates to user experienced data rate.
12 www.policytracker.com/blog/coverage-obligations/