Network Sharing Fundamentals.

(Updated & revisited thoughts on the benefits of network sharing)

Dr. Kim Kyllesbech Larsen

.
Why sharing a part or all of the mobile network?

Between 40% to 50% of sites are low or no profitable (“the ugly tail”)

Frees up cash to be spend in areas that matters (“Save for Growth”).

Effective Opex & Capex measure increasing operational efficiencies.

Increased network quality for a lot less than standalone (“Best network”).

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals 2

Network sharing boils down to 4 major considerations.

Who to share with? (your equal, your better or your worse or all)

What to share? (sites, passive, active, frequencies)

Where to share? (rural, sub-urban, urban, all, etc.)

How to share? (“the legal stuff”)

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals 3

Mobile Emerging markets – growth slowing down.
Mob 3G
20 % %
10
%
% 25
%
76
68
%
MEA Example
% 112
%
60
60
%
% Revenue 2009 – 2012 + 48% pa.
88
55
% 111 Revenue 2012 – 2017 9% pa.
%
%
Opex 2012 – 2017 + 12% pa.
134
%

& 3G has to get started

- Top-line pressure (voice & sms).
- Opex pressure.

Emerging Market Growth on expense of profitability

Long-term development troublesome.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 4

Tellabs global study1 …The End-of-Profit.
The end to the mobile-only business model as we know it?

Mobile carrier study

“End-to-Profit” (Tellabs study). Causes:
- Modernization pressure.
- Exponential data growth.
- Capacity Crunch.
- Need for more spectrum.
- LTE introduction.

- Revenue slow down.

- Decline of legacy business.
Mounting cash - Increased competition as market
pressure saturate.
- Increased cost.
Business
model
breakdown What can be learned?

• Risk of End-of-Profit next 5 years?

• Un-managed mobile data demand.
• Short-term price-plans wo long-term view.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 5

 New business!?

Europe 2020 … A soft landing? QoS, LTE, IoT, Media, FM

C, …

Defend philosophy!
Stop / Slow Revenue Decline
2020
2008 +15%
CAGR 13% CAGR 0.8% 1.9% + New Revenue?
Max 30+%
Total Revenue

The Hunt for $30+Bn − Usage Cost

2002 Efficiency game
Optimize: Defend / −Slow Market Invest SAC & SRC
Europe Mobile Revenue Ebitda Decline
− Personnel Cost

Just prior to crisis − Technology Cost (ca. 15% – 20%)

CAGR 1.9% 0.8%
− Other Cost

-15% = EBITDA (WEU ca. 37% 1)

2005
2010 Increased cash pressure − Network Depreciation
Europe Mobile Opex New technology /
Modernize
− Spectrum Amortization
38%
37% − Spectrum invest (0.8 – 0.05 € per MHz-Pop)
38%
31% − Capex (new rollout < +10+% of Revenue)
35%
Red color represent Technology driven cost
Europe Mobile Ebitda
1 BoA ML Global Wireless Matrix 1Q11, margin data for 4Q 2010.

Dr. Kim Kyllesbech Larsen , Network Sharing Fundamentals. 6

Mobile broadband journey … be prepared.
The lessons learned from mature markets.

Voice revenue decline (faster than data revenue uptake)

Messaging revenue decline (particular for Smartphones & OTT)

Cash and margin pressure from new technology introduction.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 7

Transform
or
Perish

Dr. Kim Kyllesbech Larsen, , Network Sharing Fundamentals.. 8

So why should you share your network?

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 9

Deutsche Telekom sharing examples (1 of 4).

Going Dutch – converging to “The rule of Three”.

2001 - 2004: 3G sharing – Capex avoidance strategy.
T-Mobile US – Cingular – The GSM Factory.
 Initial discussion with Orange NL started in mid-2001. 2001 – 2004: Regional GSM Sharing JV.
 JV operational from mid 2002 to 2004.
-
T-Mobile– Orange NL merger… 2008 – 2009.
Site & ancillary sharing. Geographical GSM RAN sharing agreement.
-
2008
Common plan & buildAcquisition.
organization.
 T-Mobile US (via JV) responsible for NYC Metro areas.
- No common procurement.
 JV closed down in YE 2004.
TMUK – H3G 3G RAN sharing – more for less.
Population ca. 22+M
2007:
 Price of Orange Joint
NL was ca.venture
€1.3B or design,
ca. €600 plan & build-co MBNL
per customer. #Base Ltd.
Stations ca. 2,300 (at time of breakup).
- Staff resistance (them vs us)
-  One
Different strategic single network by 2010 with
objectives.  Cingular (via JV) responsible for California/Nevada areas.
- TMNL decides no need
- Ca.for5,000
ancillary TMUK
radio
sharing.
fewer addsand
nodes 3,000 – 5,000 3G Node-Bs that would otherwiseca.
 Population not40+M T-Mobile– Orange UK Network JV.
have(TMUS
been had 1.7M subs @ breakup in CA/NV)
-
financially/economical
More economical to share own infrastructure than common. feasible. 2009: EENetwork (ad)Venture
#Base Stations ca. 5,000 (at–time
TheofBIGGEST
breakup). Network in UK!
- Ca.3,300 (ca. 50%) fewer site locations.
 Common consolidation
 network
Oct 2007 T-Mobile acquire Orange; 3G plan & buildstarted.
organization (MBNL Ltd).
 Venture discontinued in 2004 with Cingular – AT&T Wireless merger.
 Securing future competitive growth.
  TMUK EBITDA net of £50m  1
(ca. single
Nov 2008 all Orange customers were migrated to T-Mobile’s radio network “run-rate”
Positive 4% networkpays
by (net)
2014ish
 TMUS avoidance).
with
$2.3B for 30%-40% denser +grid
California/Nevada addthan standalone.
spectrum optionality.
 leveraging on higher spectral efficiency by consolidating.
– - Starting
 TMUS point a network
forced of 14,000
to spin-off
H3G benefits from faster and much more efficient deployment . sites,
10MHz today
in NYC the end-game
Metro is 18,500.
markets (very painful!).
 On track to deliver synergies in excess of €1+B by 2013 (in time
 Total
 Positive annual Capex benefit of £79m
 &
9,000
by 2012
money).
Nationwide
site
(18% roaming
locations
“run-rate” willagreement.
be terminated (33% reduction)
avoidance).
Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom
– H3G capital benefits far in excess of£0.5B
13 (estimated saving & avoidance).
Leveraging higher spectral efficiency by consolidation.
 Substantial site lease cost savings and cash prevention
 Large and readilyexpected.
achievable synergies in both Network & IT. Kim Kyllesbech Larsen, Technology - T-Mobile. 15

– From 2011 and onwards.

 Significant synergies with NPV in excess of £3.5 bn.
- Opex run-rate synergies ca. 35% (on relevant cost!)
Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 14
- Capex “run-rate” synergies up-to 25%.
 Integration
Kim Kyllesbech Larsen,& termination
Technology cost
Economics – Deutsche Telecom of up-to
16 £1.2 bn.
 EE has the BIGGEST mobile network(s) in UK which will remain so even
after consolidation and integration has been finalized.

Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 17

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 10

Deutsche Telekom sharing examples (2 of 4).
Going Dutch – converging to “The rule of Three”.
2001 - 2004: 3G sharing – Capex avoidance strategy.

 Initial discussion with Orange NL started in mid-2001.

 JV operational from mid 2002 to 2004.
- Site & ancillary sharing.
- Common plan & build organization.
- No common procurement.
 JV closed down in YE 2004.
- Staff resistance (them vs us)
- Different strategic objectives.
- TMNL decides no need for ancillary sharing.
- More economical to share own infrastructure than common.
T-Mobile– Orange NL merger… 2008 – 2009.
 Oct 2007 T-Mobile acquire Orange; network consolidation started.
2008 Acquisition.
 Nov 2008 all Orange customers were migrated to T-Mobile’s radio network

 Price of Orange NL was ca. €1.3B or ca. €600 per customer.

Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 13
 One single network by 2010 with
- Ca. 5,000 fewer radio nodes and
- Ca.3,300 (ca. 50%) fewer site locations.
 Securing future competitive growth.
 leveraging on higher spectral efficiency by consolidating.
 On track to deliver synergies in excess of €1+B by 2013 (in time & money).

Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 14

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 11

Deutsche Telekom sharing examples (3 of 4).

T-Mobile US – Cingular – The GSM Factory.

2001 – 2004: Regional GSM Sharing JV.

 Geographical GSM RAN sharing agreement.

 T-Mobile US (via JV) responsible for NYC Metro areas.

 Population ca. 22+M
 #Base Stations ca. 2,300 (at time of breakup).

 Cingular (via JV) responsible for California/Nevada areas.

 Population ca. 40+M (TMUS had 1.7M subs @ breakup in CA/NV)
 #Base Stations ca. 5,000 (at time of breakup).

 Venture discontinued in 2004 with Cingular – AT&T Wireless merger.

 TMUS pays (net) $2.3B for California/Nevada + add spectrum optionality.
 TMUS forced to spin-off 10MHz in NYC Metro markets (very painful!).
 Nationwide roaming agreement.

Kim Kyllesbech Larsen, Technology - T-Mobile. 15

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 12

Deutsche Telekom sharing examples (4 of 4).
TMUK – H3G 3G RAN sharing – more for less.
2007: Joint venture design, plan & build-co MBNL Ltd.

 TMUK adds 3,000 – 5,000 3G Node-Bs that would otherwise not have been
financially/economical feasible.
 Common 3G plan & build organization (MBNL Ltd).
 Positive TMUK EBITDA net of £50m (ca. 4% “run-rate” avoidance).
– H3G benefits from faster and much more efficient deployment .
 Positive annual Capex benefit of £79m by 2012 (18% “run-rate” avoidance).
– H3G capital benefits far in excess of £0.5B (estimated saving & avoidance).
 Substantial site lease cost savings and cash prevention expected. T-Mobile– Orange UK Network JV.
– From 2011 and onwards. 2009: EE Network (ad)Venture – The BIGGEST Network in UK!

 1 single network by 2014ish with 30%-40% denser grid than standalone.

- Starting point a network of 14,000 sites, today the end-game is 18,500.
Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 16  Total 9,000 site locations will be terminated (33% reduction)
 Leveraging higher spectral efficiency by consolidation.
 Large and readily achievable synergies in both Network & IT.
 Significant synergies with NPV in excess of £3.5 bn.
- Opex run-rate synergies ca. 35% (on relevant cost!)
- Capex “run-rate” synergies up-to 25%.
 Integration & termination cost of up-to £1.2 bn.
 EE has the BIGGEST mobile network(s) in UK which will remain so even
after consolidation and integration has been finalized.

Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 17

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 13

Vodafone sharing examples.
Vodafone – Orange sharing … 2007. Vodafone – Telefonica sharing … 2008.
Network sharing agreement in Spain and Romania. Network sharing agreements for Germany, Ireland and the UK
with detailed discussions ongoing in the Czech Republic.

 Rural Area / Geographical sharing.

 Passive RAN network site sharing.
 National roaming like sharing concept with joint field services.
 Traffic managed independently of each other.
 Rural areas with population of less than 25,000 pops.
 Customers expected to benefit from improved coverage.
 Ca. 1,500 Node-Bs where shared by YE2007 with max 5,000 by YE2009.
 Benefits in the order of ”hundreds of million” £ for both over next 10 years.
 Venture frozen in 2009 as VF announced sharing deal with Telefonica.
 Today (May 2012) they share 4,000 site locations.

Vodafone – Telefonica sharing … 2012.

Getting a lot more for less …. Capex & Opex avoidance.

(i.e., VF-Europe Opex in 2008 was £16.4 bn and TF-Europe Passive sharing including backhaul.
2008 Opex was in the order of £13 bn). (i.e., VF-Europe Opex in 2008 was £16.4 bn and TF-Europe 2008 Opex was in the order of £13 bn).
Kim Kyllesbech Larsen, Technology - T-Mobile. 18 Kim Kyllesbech Larsen, Technology - T-Mobile. 19

 Common Build JV, planning & design separately.

 1 single network by 2015 with doubling the site count to standalone.
- Each has ca. 10,300 sites today with shared end-game of 18,500.
 Total of ca. 2,000+ site locations will be terminated (10% reduction).
 Geographical (50%-50%) sharing (i.e., London halved).
 No Frequency sharing.
 Individual supplier relationships (missing out on procurement scale?).
 Massive Opex and Capex avoidance.
 This is NOT an Opex reduction game but rather matching EE super-grid.

Kim Kyllesbech Larsen, Technology Economics – Deutsche Telecom 20

Dr. Kim Kyllesbech Larsen 14

The Good and The Ugly.
Recipe for successful merger (or network sharing) and is matching technology
landscape and strategic outlook. Matching spectrum position and network grid are
much more valuable (short-term) for synergies than complementary spectrum.
AT&T – Cingular merger Sprint - Nextel

Dominated by Nextel

Matching technology landscape and strategic outlook. Mismatch in technology landscape & strategic outlook.

Good complementary spectrum (high grid match). Complementary spectrum but relative low grid match.

Fairly symmetric & matching business structures and Very different business structures and models.
models.

Dr. Kim Kyllesbech Larsen 15

Expectation management – the full sharing potential.

Expect up-to 35% saving on Tech Opex

Total Opex
100%

Up-to 5% on Total Corporate Opex

Cluster Opex
40% Termination cost 1.5 – 3+ × of Opex savings

Technology Opex
Integration Capex synergetic with BaU Capex
NT 14%

RAN 7% Instant Cell split potential Enhanced Capacity

RAN
saving
Spectral efficiency gains (>10%+)
Illustration

Dr. Kim Kyllesbech Larsen 16

Stages of sharing benefits.
The best sharing strategy depends on the business cycle and
technology age.
< 5 years 5+ years > 5+ years

LTE GSM – UMTS

UMTS UMTS - GSM
(LTE piggybacking)

Rollout Phase Steady State Modernization

UK: 3G T-Mobile – 3 UK UK. T-Mobile UK – Orange JV (EE Ltd). Poland: PTC – Orange incl. LTE

Passive sharing: Site Lease & Civil Works, Illustration

Mast/Tower sharing, Ancillary & Rack sharing, and Backhaul Sharing.
Active sharing: e.g., Frequencies, TRXs, PAs, Baseband, CPU, ports, ….

 High Capex prevention.  Little Capex benefits.  High Capex prevention.

 Opex prevention.  Opex savings.  Opex savings.
 Cash optimized startup.  Significant write-off.  Minor write-off.
 Best network.  High re-structuring cost.  Re-structuring cost.
 Extended coverage.  Instant cell split.
 Better network.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 17

 Economics of RAN sharing benefits.
Sharing stages Capex Synergy OPEX Synergy Restructure Cost Write-off
Passive sharing
• Site build
Opex prevention
Rollout Phase • Mast Restructuring Relative low
• Site lease
• Rack / Ancillary cost can be low if exposure if little
Bulk (>80%) of sites • Non-telco services
Active sharing little legacy legacy
and nodes to be • Telco services
• MW/Fiber infrastructure is infrastructure is
• Energy
deployed. • Electronics present. present.
• Resources
• Spectrum
• Resources

Opex saving if Termination As most of the

Passive sharing absolute number •Site lease. network has
Steady State of site locations •Site restoration. been deployed at
80% of coverage and are reduced. •Service this stage the
sites deployed. Mainly Low Capex level Primarily Opex Contracts. write-off
Active sharing prevention in •Personnel cost exposure can be
capacity additions and Other significant even if
case of site
coverage maintenance. number •JV overhead equipment can
expansion. •Legal, etc.. be re-used.

Opex saving if Restructure cost If decision for

Passive sharing
Modernization/ absolute number can be network sharing
is taken in the
of site locations significant.
Obsolescence are reduced. Although contract
renewal /
Medium Capex level obsolescence
Active element / node Active sharing Primarily Opex termination can phase write-off
replacement, prevention in be less costly exposure can be
technology migration. case of site due to longer relative light both
number operational for equipment
Site consolidation. Substantial Capex period.
expansion. and site-build.

= Low = High
Illustration Synergy potential
Cost exposure = Low = High

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 18

The ugly tail … the low profitability areas.
Should drive sharing in low-traffic areas
Illustration

Urban Sub-urban Rural-like areas

Cumulated Revenue (Traffic)

0%
Low profitability
sites
20%
50% revenue ≈ 10% sites
50% sites takes
40% less than 10% revenue

60%
Top 30% sites ≈ 80% revenue.

80%

100%
0% 50% 100%

Sites

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 19

Network Sharing can provide better economics and
market timing …
Frequency Site Radio Backhaul Backbone Core BSS
(MHz) (acq. + build) (electronics) (transport) (transport) (switch & control) (bill & care)

plmn 1

plmn 2
MNO 1
Core BSS
plmn 1 + plmn 2
(optional) BTS &
NODE-B
MNO 2
eNodeB Core
BSS

Capex Efficiency Partly

40%-60% < 35% up-to 50% up-to 50% Less likely
prevention enabler possible

Opex Efficiency scale scale Partly

< 35% ca. 35% Less likely
prevention enabler discount discount possible

Regulatory
HIGH LOW LOWER LOWER LOWER HIGH HIGH
complexity

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 20

Anatomy of network sharing.
RAN sharing guaranties competitive differentiation, operator
independency and vast consumer quality improvements.

• Sharing: Costly Radio Access Network infrastructure will be shared,

• Not shared: All core network and service infrastructures that provides respective
customers with differentiated services, applications, handsets, rate plans, etc.
• Result: A network with greater capacity (i.e., instant cell split) and improved coverage.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 21

Network sharing.
A mean to close the mobile broadband coverage gap (in CEE).
Network Sharing Strategies Safe for Service by Sharing.
Rural areas1 Operator A SHARED Operator B

Substantial improved
BILL PRICE BRAND SALES BILL PRICE BRAND SALES
coverage, capacity boost and
SERVICES SERVICES quality of services to
CORE CORE
the consumer at a Quality Level
2G, 3G & LTE RAN incl. BACKHAUL SHARE
GSM UMTS LTE NOT

Urban areas1 Idealized Illustration economical viable in standalone.

BILL PRICE BRAND SALES BILL PRICE BRAND SALES

SERVICES SERVICES
Benefits.
CORE CORE • Opex avoidance & savings.
• Substantial Capex avoidance.
2G, 3G & LTE RAN incl. BACKHAUL SHARE • Shared Modernization.
GSM
900 & 1800
UMTS
900 & 2100
LTE SHARING
800, 2100 & 2600 MHz
GSM
900 & 1800
• Shared LTE deployment.
• A Much better network.
Note: frequency bands not to scale!

1 Note sharing spectrum between two (or more) MNOs might not be regulatory allowed.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 22

Network sharing flavors …
Site sharing (*) RAN Sharing (*) National Roaming (*)
Capacity limited Coverage limited Rural

HSS
HLR HSS HSS
HLR HSS HSS
HLR HSS

Core Core Core Core Core Core

BSC BSC BSC BSC BSC

RNC RNC RNC RNC RNC

Shared site and passives Shared Radio, aggregation Wholesale arrangement,

Independent BTS, NB, eNB. & frequencies (optional). geographical partnership.

 Passive sharing.  Active sharing (MOCN1)  Geographic sharing.

 shared transport (possible).  Shared transport.  One frequency sufficient.
 Independent frequencies.  Frequencies sharing.  Wholesale/cost-sharing.,

1 Multi-Operator Core Network supporting RAN Sharing, (*) For LTE there is no BSC/RNC, core networks connected directly to the eNode-B.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 23

Common Frequency Sharing…
Solution for low-demand, rural areas and symmetric demand scenarios.
Frequency pooling (*)  3GPP Release 8, 2009 (earliest) onwards with the following sharing
concepts:
 Gateway Core Network (GWCN) shared core network (CN) (multiple CNs
... connected to a common core, connected to the shared RAN).
HLR HLR
 MOCN: Multi-Operator Core Network where only the RAN is shared (i.e.,
NO common CN).
 Introduction of Iu Flex allowing multiple CNs connecting to shared RAN.
Core
Core ...  Multiple core networks connected to a common radio access
network (RAN) sharing a single frequency or a pool of frequencies.
Shared IP
backhaul
 Service requirements & capabilities not limited by the sharing
requirements (i.e., resides in core network or service creation
LTE platforms above the core network).
eNode-B
 Requires user equipment support (i.e., R8 or later).
 Non-supporting user equipment will ignore the broadcast system
Shared Freq., Radio & aggregation.
information related to sharing functionality.
 Fairly complex coordination issues on resource allocation among
 1 operator share its sharing parties, making this concept more interesting for low-traffic
spectrum with others. rural areas (where demand is no issue) or highly asymmetric
 Multiple operators pool their traffic situations.
spectrum assets together 1 MORAN = Multi-Operator Radio Access Network sharing of all active electronics with exception of
and share total spectrum. frequencies. 2 MOCN = Multi-Operator Core Network = two core networks connected to 1 frequency. (*) For
3G network core networks connect to the RNC that then connects to the Node-B.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 24

Network components mapped to network layers.
The deeper into the network infrastructure is shared the more the sharing
concept will appear as a merger or NetCo concept.

Network Merger – Netco concept

Network Sharing

Radio Access CS & PS Signaling VAS IT & CS

Network Network Network Network Network

 Spectrum /  Classical MSC/VLR  Classical HLR  SMSC  Billing system

Frequencies  R4 MSC Server &  NG HLR  MMSC  Rating
 GSM BTS Gateway  IN platform  VMS  Mediation
 GSM BSC  Multiplexing  Interconnect  WAP  CRM
 GSM TRX  GGSN & SGSN  NMS & operations  Portals  SAP/Finance
 3G Node-B (packet core).  Etc.  3rd party content systems.
 3G RNC  Evolved Packet Core.  NMS & operations  Business Intelligence.
 3G Carrier & Channel  IP networks (routers,  Etc…  Call center systems
elements FW, etc..) (call routing, ..)
 e-Node-B (LTE RAN)  Backbone transport  OSS
 Backhaul (MW & LL)  Interconnect  IT Operations.
 Routers, switches and  NMS & operations.  Etc.
multiplexing  Etc.
 SW Licenses &
features.
 NMS & operations.
 Etc.
Note: above categorization is guiding but not fully un-ambiguous.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 25

Why one should NOT commence Network Sharing.
2 out of 3 NS deals considered are put on ice again!

Complex Governance

Technology mismatch

Divest / Spin-off / merger

very complex

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 26

 Rural Tower & RAN Sharing scenario (illustration)
Doing much more for a lot less refocus cash on areas that matters

Towers JV (IPO optionality) Single RAN Sharing (NetCo).

0 Node-B 0 Node-B
500 Towers 500 Towers
MNO1 500 old BTS
xx% geo coverage
500 old BTS
yy% geo coverage
Today
High Capex Synergies (>50%)
MNO2 by Joint procurement!
Consolidation Harmonization
Tower 650 SRAN
Dismantling of surplus No Re-use (new)
JV Tower locations
Shared Single
-350 RAN deployment
650 Towers 650 SRAN
35% Consolidated Rural & Sub-urban Shared
+1.5–2 Years
Coverage focus ensure
configurationally &
1,000 Towers 1,000 SRAN
Real Opex saving! Consolidated
operational simplicity
(LTE option) End-game
No frequency sharing!

High Capex & Opex More More Shared Shared

avoidance, high Towers RAN backhaul backbone
ROCE

NodeB RNC MNO2

 Opex equivalent to 500 Towers (1+1 “=“ ½). BTS BSC CN
 Better quality & more capacity (1 + 1 > 2 effect). MNO1
 Favorable cash impact compared to standalone Much better & efficient network! CN
End-game: 1,000 Towers
Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 27
 3-party rural network sharing (illustration only)
Doing much more for a lot less optimizing cash and margin.

Towers (Rural – Sub-urban Areas). Single RAN (SRAN) Sharing.

600 Towers 350 Towers 440 Towers 600 SRAN 350 SRAN 440 2G & 3G
2&3G 2&3G Legacy nodes 2012
A
Consolidation Harmonization
B Dismantling of 800 SRAN 800 SRAN LTE
2&3G Nodes upgrades (new)
surplus Tower
Shared Single RAN
C locations
deployment
Shared -590 800 Towers
Consolidated
800 SRAN
2&3G + LTE 2014
~40% 1,000 SRAN
1,000 Towers
Consolidated Multi-mode End-game

• 800 Towers shared by 3. Less Shared

Fewer Shared
• 800 whilst effective paying for 267 RAN backbone
Towers Backhaul
(1+1+1 “=“ 1/3). eNodeB A CN
• Improved network with 33% to
NodeB RNC
130% increase in sites.
• Much improved TCO and ROI. BTS BSC B CN
• Low LTE entry cost and future
modernization cost. Optimized target
C CN
End-game: 1,000 Towers
Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 28
National Roaming boils down to 3 major
considerations.

GUEST HOST
Under-utilized network.
No Coverage / No Network.
(“plenty” of capacity)

Wholesale Tariff better economics Wholesale revenue at no or

than Network Cost & Invest. very little additional Cost and Invest.

Long rollout lead-time Wholesale income more attractive

National Roaming a timing-bridge. than risk of competitor network access.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 29

 Profit & Loss … what to expect
Personnel savings by High usage cost jvf.
resource sharing. Wholesale agreement

Standalone Network Sharing National Roaming

Revenue Revenue Revenue
− Usage Cost − Personnel
Usage Cost savings by − Usage Cost
−
redundancies.
Market Invest SAC & SRC − Market Invest SAC & SRC − Market Invest SAC & SRC

− Personnel Cost − Personnel Cost − Personnel Cost

− Technology Cost − Technology Cost − Technology Cost

− Other Cost − Other Cost − Other Cost

= EBITDA = EBITDA = EBITDA

− Network Depreciation Network DepreciationHigher Opex savings & Network Depreciation

Opex savings &− −
prevention prevention
− Spectrum Amortization − Spectrum Amortization − Spectrum Amortization

− Spectrum invest
− Spectrum investLow depreciation charges
− Spectrum invest
(typically high write-off).
− Capex − Capex − Capex

Red color represent Technology driven cost Red color represent Technology driven cost Red color represent Technology driven cost
Capex prevention Less depreciation Capex prevention.
typically re-prioritized. (& some write-off)

It is far from obvious that National Roaming should be more economical than
Network Sharing … Structurally it is more complex to get right.
Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 30
National roaming … another way of sharing.
Why National Roaming?
National Roaming (*)
• 1 party has coverage, the other not.
• per Technology (i.e., GSM-only, UMTS-only, …). Typical Rural
• Different regional spectrum positions.
• Contiguous well-defined network (e.g., 50 to 100s of radio
nodes). HSS
HLR HSS
• Often geographical splits
• Can also apply to infrastructure network sharing.
• Time to market. Core Core
• Customer experience can be controlled independently by
Roamer.
BSC
• Capacity is not an issue for the hosting MNO. BSC
RNC RNC
• Attractive economics compared to building network:
• Provides very similar benefits to infrastructure sharing.
• Roaming MNO gets Capex & Opex avoidance, but will have cost
associated with traffic on Hosting network.
OpCo1 Host to OpCo 2 OpCo2 Host to OpCo1
• Hosting MNO gets wholesale revenue typically in low-traffic areas with
low or no profitability (i.e., increased utilization & efficiency). Wholesale arrangement,
• Regulatory encouragement (or enforcement). geographical partnership.
• Relationships tends to be of temporary nature.
 Geographic sharing.
• 2G & 3G National Roaming are standardized with working
 One frequency sufficient.
technical solutions used in several countries between MNO and
 Wholesale/cost-sharing.,
MVNO.
(*) For LTE there is no BSC/RNC, core networks connected directly to the eNode-B possibly via IP aggregation & switching.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 31

National roaming … many case stories around the world!
National Roaming (*)
Some examples;
• T-Mobile US (New York) & Cingular (California) – Terminated. Typical Rural
• T-Mobile US 2G roaming on AT&Ts network - Active.
• T-Mobile US 3G roaming on AT&Ts network – Not operational (too complex).
• FCC (US Regulator) issued a ruling (2011) requiring MNOs to sign mobile data HSS
HLR HSS
national roaming1 agreements with anyone who asks (at reasonable terms &
conditions … last not been specified by FCC).
• T-Mobile Austria on Hutch 3G network in rural areas – Active. Core
Core
• Hutch on T-Mobile Austria’s GSM network – Active (decreasing)
• H3G UK on Orange GSM – Active (decreasing).
• O2 Germany, 2G national roaming on Deutsche Telekom GSM network outside their BSC BSC
RNC
own 2G coverage (particular rural and sub-urban areas) – Terminated. RNC
• T-Mobile UK and Orange UK mutual national roaming on each other’s 2G networks
extending the coverage for both customer bases – Active.
• Free Mobile (Iliad) in France has a national roaming agreement with Orange. This
agreement covers both 2G and 3G – Active. OpCo1 Host to OpCo 2 OpCo2 Host to OpCo1

• India is likewise (in)famous for many 2G (“3G”) national roaming deals between the Wholesale arrangement,
as many mobile MNOs – Active (3G still a regulatory issue). geographical partnership.

I have found2 no examples where an MNO  Geographic sharing.

decommissioned its network for national roaming.  One frequency sufficient.
 Wholesale/cost-sharing.,
1 Note both AT&T and VERIZON was very much against this FCC ruling as the correctly pointed out that it is very difficult to control & plan for mobile data
traffic and that they were already spectrum constrained and therefore do not have excess capacity. 2 Though I have been part of discussions entertaining
such an idea. (*) For LTE there is no BSC/RNC, core networks
connected directly to the eNode-B possibly via IP
aggregation & switching.
National roaming … can be a flawed business logic!
Why maybe not?
• Firstly, complexity is not so much in the technical area but very much contractual and ensuring
sufficient risk mitigation against operational disruption.
• Regulatory & competition authority issues
• If MNOs setup mutual agreements, Regulator might enforce those agreements onto other interested
parties (i.e., see 2011 FCC Ruling).
• Further Regulator might decide non-compliance with spectrum utilization or conditions of the
roaming party (i.e., its spectrum is no longer in use).
• Decommissioning of existing infrastructure and investments (i.e., write-offs).
• MNOs could financial compensate each other if decom infrastructure would be taken into use by
Hosting MNO.
• Operational risk of relying 100% on the other partners network.
• Compared with network sharing that provides for co-ownership & co-control of network.
• Can be mitigated to some extend in contract and by choosing symmetric areas (i.e., ensuring
symmetric threat levels)
• Can carry very substantial operational risks.
• Change of ownership.
• Bankruptcy.
• QoS & Customer Experience guaranty for 3G mobile data usage very difficult & carries great mutual
risks.
• Change of mind / contractual (even illegal) break-up or non-compliance.
• Tends to be very complex commercial negotiations, resulting contracts, processes and
procedures.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 33

Other business models …LTE as a Service.

Emerging business models – LTE network factory

Cash
optimized
startup via
virtualization
& OTT
based
services.
Option:
Small cell centric
startup and
Capacity as a
Service.

Provides. Enablers.
Attractive (startup) cost economics. Profitability & cash crunch.
Relative low Capex – cash optimized. Incumbent spectrum crunch.
Increased spectral efficiency & utilization. MVNO / tier-2&3 MNO appetite.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 34

Other business models … ultra-efficient transformation.

Emerging business models – piggybacking on Virtualization & Cloud

3rd party, media companies,
MNO/MVNO CDN & SDNs.

3rd parties 3rd parties

(supplier)
delivers BSS /
delivers core
OSS cloud network
services to functionality
SmartCo (off- (i.e., HSS,
the-shelf) PCRF, etc..)

Provides. Enablers.
Data-only QoS transparent network. Regulatory support.
Network services to MNO & MVNO. Spare Spectrum (i.e., typical Startup).
Dedicated OTT network services. MNO & MVNO appetite.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 35

Technology cost and synergy potential.
Synergy potential Share of Managed services Network sharing
Mobile ONLY Technology Opex €€ (€) €€€ (€)
NT FTE Ca. 10% Typical 20% Min. 20% - 35%
HC reduction
Typically Capex >35% but depends on
NT Services Ca. 15% commitment network reduction.
Good savings
potential, though risk >35% but depends on
Rental & Leasing Ca. 25% - 30%
for future sharing network reduction.
optionality
Ca. 5% - 10% Opex – Capex More Opex – Capex
Transmission (can be a lot higher if majority
leased transport) trade-off trade-off

10% - 20% Minor opportunities

IT FTE 5%
HC reduction <10% due to scale.
Opex – Capex Minor opportunities
IT Services 25%
trade-offs <10% due to scale.

Other 10% - 15% Minimum 10% pa At least 35%

Illustration
Note: Above numbers serve as illustrations only. Different operations may have different Technology Opex distributions..

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 36

Key messages.
What we need to be passionate about!

First things first

Utilize technology to achieve the best operational performance
RRH, SDR RAN, Single-RAN, FTTS, Virtualization, Cloud, …

don’t over-focus on financial savings!

Network sharing provides cost reduction & increased quality.
& increased complexity
& upfront cash needs
& don’t forget!
Sharing models for mobile applies to fixed broadband as well.
Maybe Even more so!

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 37

Key elements for successful network sharing

CEOs agree with & endorse Network Sharing.

Sharing Partners have similar perceived benefits (win-win feel).

Focus on creating a better network for less.

Both parties share a similar end-goal and similar strategic outlook.

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 38

Last but not least.

Network sharing is a very long term engagement (“for Life”!)

Do consider that break-up can happen … and be prepared! (“legal stuff”)

Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals. 39

The key value proposition of a mobile network
is ....

Freedom

& Mobility