Modern Telecommunication ... The Very Basic.

Modern Telecommunication
The Very Basics
Dr. Kim Kyllesbech Larsen, TechNEconomy.
Training Material 2016.

Compounded Annual Growth Rate CAGR
• Growth
Year 2014 2015 2016 2017 CAGR
Data Volume 3.45 4.83 6.28 7.53 ~ 30%
Yr by Yr growth 50% 40% 30% 20%
Wiki
Data Volume 2017
Data Volume 2014

Profit & loss (P&L).
Wiki
Total Revenue
Technology Cost (ca. 15% – 20+%)
Usage Cost−
Market Invest SAC & SRC
−
= EBITDA (APAC ca. 45%)
Personnel Cost
Other Cost
−
−
−
Network Depreciation−
Spectrum Amortization−
Capex (new rollout  +20+% of Revenue)−
Spectrum invest (0.5 – 1.0 $ per MHz-Pop)−
+ New Revenue?
Defend philosophy!
Stop / Slow Revenue Decline
New business?
QoS, LTE, IoT, Media/TV,
FMC/FMS, …
Efficiency game
Optimize: Defend / Slow
Ebitda Decline
Increased cash pressure
New technology (Fiber,
LTE, 5G,..) & Modernize
SAC: Subscriber Acquisition Cost
SRC: Subscriber Retention Cost
EBITDA: Earnings Before
Interest, Tax, Depreciation &
Amortization
Cash
Careful! Cash
calculation involves
more than what is
depicted here!
EBITDA & Margin
(EBITDA/Revenues) Key
metrics for assessing financial
health of business
Note: From Revenue we can
calculate the ARPU (Average
Revenue per User) by Revenue
divided by the Average over
Period Users.

Mega, Giga, Tera, Peta, Exa … Bytes
Name Symbol 10n Decimal
Yotta Y 1024 1 000 000 000 000 000 000 000 000
Zetta Z 1021 1 000 000 000 000 000 000 000
Exa E 1018 1 000 000 000 000 000 000
Peta P 1015 1 000 000 000 000 000
Tera
(Trillion)
T 1012 1 000 000 000 000
Giga
(Billion)
G 109 1 000 000 000
Mega
(Million)
M 106 1 000 000
kilo k 103 1 000
Source: Cisco VNI Global IP Traffic Forecast, 2014–2019
Global IP Traffic
1992: 0.001 GB per second.
↓× 30 over 5 years
1997: 0.03 GB per second.
↓× 3,333 over 3 years
2000: 100 GB per second.
↓ CAGR +44% per year
2014: 16,000 GB per second.
↓ CAGR +46% per year
2019: 50,000 GB per second.
100 MB ~ 4 minutes Youtube at HD (720p).
700 MB ~ size of standard movie on normal DVD.
1 GB ~ 5 minutes of 4K UHD TV viewing.
10 GB or 5 hours of Youtube watching per month.
25 GB ~ 1 Blue-ray movie size.
1 TB ~ 250,000+ HD songs (~ 1+ million hours of music).
1 PB ~ 13+ years of HD-TV videos.
50 PB ~ Entire written works of Mankind from the beginning.
4.5+ Billion Years (GY)
~Age of Earth
7000 Yotta atoms
In a typical human body
~ 100 T Ants
in the world
~ Number of planets in
the Universe
Less than 50 k elephants
left in the wild
~ 50 M died as a direct
consequence of WW II
Wiki

Minutes, Bytes & bits per second.
6
• Voice Usage is well defined & understood, it is a capacity & cost driver!
• Calls measured in minutes (or Erlang).
• Network Impact per time unit of voice usage is “always” the same (i.e., fixed bandwidth or resource allocation).
• Voice Usage is well defined & understood, it is a capacity & cost driver!
• Calls measured in minutes (or Erlang).
• Network Impact per time unit of voice usage is “always” the same (i.e., fixed bandwidth or resource allocation).
Voice Minutes
• Byte B is a measure of total information consumed; ∝ ∑ ; …
, rb is the supplied network bit rate in bits per second.
• Its a volumetric unit of data consumption and similar (in principle) to a Voice Minute.
• Network impact per time unit of data usage can vary enormeously (i.e., variable bandwidth or resource allocation).
• Byte B is a measure of total information consumed; ∝ ∑ ; …
, rb is the supplied network bit rate in bits per second.
• Its a volumetric unit of data consumption and similar (in principle) to a Voice Minute.
• Network impact per time unit of data usage can vary enormeously (i.e., variable bandwidth or resource allocation).
Byte (8 bits)
MB, GB, TB, …
Note ∝ ∑ ∆ =
, rb is constant. e.g., 12.2 kbps
• bits per second is a fundamental measure of the information rate.
• For data consumption the bit rate can (and often will) vary significantly between one and another instant of time.
• Networks are planned according with the expected maximum gross demanded bit rate arising from all users.
• bits per second is a fundamental measure of the information rate.
• For data consumption the bit rate can (and often will) vary significantly between one and another instant of time.
• Networks are planned according with the expected maximum gross demanded bit rate arising from all users.
bits per second
kbps, Mbps, Gbps
Byte is NOT a capacity or cost driver!!!
Bits per second is the capacity or cost driver!!!
Wiki

7
Throughput drives network expansion & cost
…Volume (Bytes) does NOT!
TRAFFIC PROFILES
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour of Day
Throughput
Mega
Bits
per second
(Mbps)
9 HOURS
4 HOURS
1
2
Traffic profile 2
same volume as 1, but
40% higher busy hour
throughput.
To keep same user
experience in the busy
hour more network
capacity needed.
Higher invest level
and network OPEX
required.
Same volumetric (Byte) demand can cause vastly different network cost
and invest levels.
Wiki

Video streaming requirements & impact.
Resolution
Width x Height
Video Rate Audio Rate
360p 0.5 – 1.0 Mbps
0.128 – 0.512
Mbps
480p (ED) 0.7 – 2.5 Mbps
720p (HD) 3.0 – 5.0 Mbps
1080p (Full HD) 4.0 – 8.0 Mbps
1440p 7.0 – 10 Mbps
2160p (4K UHD) 25 – 45 Mbps
Note: p stands for progressive scanning where all lines of each frame are drawn in sequence.
Best Video Format for Mobile: MPEG4, Best Video Codec for Mobile: H.264, Audio Codec: AAC
(Advanced Audio Codec).
Source: Sandvine Global Internet Phenomena, Asia-Pacific &
Europe, September 2015.
Busy Hour Traffic Distribution – APAC
(based on bits per second)
Example:
Radio cell sector has 1,000 users in the busy hour. Each watching 4 Minutes video stream (e.g.,
Youtube). Every 4 seconds a new video might be watched (0.25 per second).
On average over 4 minutes one would expect 60 simultaneous video streams, which at 720p would
be a BH demand of min. 180 Mbps or at 480p it would be min. 42 Mbps.
Wiki
4 min (duration) x 60 sec/min x ¼ videos/sec (arrival of new views)

Circuit-switched connections.
Circuit Switched (CS) Connections
A B
- End-2-End well defined circuit fully reserved
until user (or network) breaks connection.
- There is no changes to the circuit path
throughout a call.
- Capacity is reserved 100% throughout a call,
irrespective of activity.
- Highly reliable with less delay.
- If path is broken (due to quality or outage),
service is lost and needs to be built up again.
- It tends to be in-efficient in its use of
switching and transmission resources.
- Public Switching Telephony Networks (PSTN)
are circuit switched in nature.
- Today many PSTNs have migrated to VoIP and
(so-called) Next Generation (NG) IP switching
networks.
Communication paths
usually called circuit.
Wiki
Network resources 100% reserved when connection established.
No statistical advantage from variation in usage..
Connection establish
from A to B and kept
until user breaks it.

Packet-switched connections.
Packet Switched (PS) Connections
A B
- Uses Content (e.g., voice or data) broken up
in so-called (IP) packets & send over the
network.
- The path is changed based on quality, traffic
and priority.
- Network resources only used when user
packets send over the network.
- Can be very reliable although more delay
sensitive than circuit switched.
- Packet prioritization possible which can
provide almost same quality as a circuit
switched connections.
- PS Networks tend to be very efficient due to
the statistical nature of packet transmission.
- Most modern switching networks (e.g., LTE,
3G HSPA) are PS-based.
Wiki
Network resources used only when user uses.
Packets (of use) statistically distributed.
Packets of use can be transmitted
via many different paths
Uses broken down into
(IP) packets and send
into the PS network.
All (IP) packets re-assembled
(re-built) into a continues
stream of content.
123
1 1
1
1
2
2 2 2
3
3
3

What brings meat to your noodles?
(in Europe: we would say what brings the butter on your bread).

?
Write (maximum) 3 sentences
down describing what drives
your business.

Modern Telecommunication ... The Very Basic.

Telecommunications in General
Tele-communications is the exchange of information
(bits) over a distance by electronic or optical means.
A complete, single telecommunications connection
consists of at least two devices, each device
equipped with a transmitter and a receiver.
Ancient Greek
means at a distance.

We are (almost) all Mobile
High GDP
APAC
Europe
North
America
APAC EMEA
Latin
America
Sources: United Nations, Department of Economic & Social Affairs, Population Division. . Mobile Penetration is based on Pyramid Research and Bank of America
Merrill Lynch Global Wireless Matrix Q1, 2014. Index Mundi is the source for the Country Age structure and data for %tage of population between 15 and 64
years of age and shown as a red dotted line which swings between 53.2% (Nigeria) to 78.2% (Singapore), with an average of 66.5% (red dashed line).
Mobile Penetration Urban Population
2013

Most urban areas have
3G Mobile Broadband
High GDP
APAC
Europe
North
America
APAC EMEA
Latin
America
Sources: United Nations, Department of Economic & Social Affairs, Population Division. . Mobile Penetration is based on Pyramid Research. Index Mundi is
the source for the Country Age structure and data for %tage of population between 15 and 64 years of age and shown as a red dotted line which swings
between 53.2% (Nigeria) to 78.2% (Singapore), with an average of 66.5% (red dashed line).
3G Penetration Urban Population Urban Area
2013

Revenue slows down, Cost grows faster
… & that’s a problem for profitability!
REVENUE GROWTH EXCEEDS
GROWTH OF OPEX
OPEX GROWTH EXCEEDS
GROWTH OF REVENUE
CAGR 2007 to 2013
High GDP
APAC
Europe North
America
APAC EMEA Latin
America
Source: Bank of America Merrill Lynch Global Wireless Matrix Q1, 2014.

18 18
Telco at a Cross-road
SMS Revenue
In decline
Voice Revenue
In decline
Data Revenue
Slow to pick up
The traditional mobile access-based
business model of Voice, SMS & Data
inevitably will decline.
Access in decline
Monetizing
the 4th Wave

History at a glance.
• 7th March 1876 Graham Bell receives patent for the telephone.
• 10th of March 1876 Graham Bell makes the “first” telephone call.
• 1876 Bell makes first long distance call (6 miles ~ 10 km).
• 1876 telephone switchboard exchange invented (Puska, Hungary).
• 1877 First commercial telephony company (Germany).
• 1878 First US commercial telephony company (New Haven, Connecticut).
• 1880 Graham Bell invented the Photo-phone transmitting voice signal over an optical beam.
• 1917 Patent for a “pocket-size folding telephone with a very thin carbon microphone” (Finland).
• 1926 First transatlantic telephone call (London, UK – New York, USA).
• 1930 First experimental videophones.
• 1930s Concept of digital switching developed in Europe and USA.
• 1936 First video phone service (Germany).
• 1941 Multi-frequency dialing introduced.
• 1946 First commercial mobile phone call.

• 1960 First laser was built (USA).
• 1962 Telstar telecom satellite (TV pictures, telephone calls & fax images).
• 1965 First electronic switching system in commercial service.
• 1965 First working optical-fiber data transmission system (Borner, Germany).
• 1968 First fully digital central switch in commercial service (London).
• 1969 ARPANET (Advanced Research Projects Agency Network) Live.
• 1975 First commercial fiber communications systems developed.
• 1978 (D)WDM concept first published.
• 1980 First (D)WDM working system.
• 1990 First working worldwide web (WWW) as we know it today (Tim Berners-Lee).
• 2000 First commercial photonic crystal fibers.
• 2012 Record breaking 1.05 Peta bits per second over 52.4 km of 12-core optical fiber.

• 1917 Patent for a “pocket-size folding telephone with a very thin carbon microphone” (Finland).
• 1946 First commercial mobile phone call
• 1978 First NMT (Nordic Mobile Telephone) call made (Finland).
• 1991 First GSM phone call and service (Finland).
• 1992 First SMS Message (Vodafone UK). 1993 First commercial SMS service (Telia, Sweden).
• 2000 First commercial GPRS (General Packet Radio Service on GSM) services (first mobile packet data).
• 2001 First commercial UMTS (Universal Mobile Telephony Service) by NTT DoCoMo (Japan).
• 2003 EDGE (Enhanced Data rates for GSM Evolution) in commercial service.
• 2004 GSM surpasses 1 Billion Users.
• 2006 GSM surpasses 2 Billion Users.
• 2007 HSDPA (High Speed Data Packet Access on UMTS) launches.
• 2007 iPhone 1 (June 29th) – GSM/GPRS/EDGE.
• 2008 iPhone 3G (July 11th)
• 2008 Global Mobile Connections surpass 4 Billion.
• 2009 First Commercial LTE (Long Term Evolution) Network launches (Sweden & Norway, TeliaSonera)
• 2010 iPad 1 (April 3rd)
• 2010 Global Mobile Connections surpass 5 Billion.
• 2011 First LTE Network in South Asia Sri Lanka Telecom Mobiltel (96 Mbps).
• 2014 VoLTE (Voice over LTE) in commercial service (first ~ May 2014, Hong Kong, Singapore, USA).
• 2014 LTE-advanced in commercial service (first ~ June 2014).

Important drivers to consider.

Gordon Moore’s (co-founder of Intel) law:
transistor count double every two years.
GPU particular important for
development of Artificial Intelligence
& Virtual/Augmented Reality Apps.
The higher the
number of
Transistors the
higher the
performance
+10Yrs
× ~ 100

Average
Top-50
Max
Top-50
On average over period a
factor 2 in power reduction
per year has been achieved.
Very few
Data points
Approx. factor 10
improvement per 5 years
On average 20+%
improvement per year

Note: after 2005 GPU outperform the CPU processing
power in terms of GFLOPS so maximum performance
should be used to calculate the initial price!
Ca. factor 6
reduction per year
In cost of computing!

Last 5 years cost of Flash
has reduced a factor 10+
Last 3 years cost of SSD
has reduced a factor 2.5
Last 5 (10) years cost of Memory
has reduced a factor 2.8 (50+)

+10Yrs
× 1,000
Improvement in
storage &
memory capacity

Access technologies development.
Caution: Above does not consider contention ratio (e.g., 1:32 or 1:64), concurrent user demand, assuming vastly different
bandwidths to get to the speed, nor does this consider that the technologies have very effective ranges at optimal speed
plotted above. So it is a bit of an apple and orange comparison!
Last 10 years
more than
× 1,000
Improvement
in user speed

Last 10 years.
The amount of computer power performance
quantum leaped new applications (not possible prior).
The cost of computing and storage has reduced dramatically
Becoming cheap and un-locking boost in software-driven innovation.
Access technologies have improved with at least factor 10 in user speed
allowing for fast low-latency access to computing and storage on the go.
What before had to be built in expensive customized hardware can now be
supported by software on cheap off the shelf hardware.

Enablers
Drivers
Transistor
Count
×1,000 over period
Computing
cost
~ 1/6 per year
Last 10 Years
Cost of
Storage
~ 1/50 over period
Storage
Capacity
×1000 over period
Cellular Access
Speed
×1000 over period
Cloudization
Virtualization
NFV
SDN
SW replacing HW
functionalities
Data
DemandSW as a Service
Storage
Higher performance for much less cost
Technology
Progress

Global mobile subscriptions
per technology.
Ca. 7.2 Billion Subscriptions
Ca. 5.5 Billion Unique Users
75% of world population
Ca. 9.2 Billion
Ca. 5.3 Billion
Unique user has (on average)
1.3 subscriptions
38%
42%
20%
50+% from
Asia Pacific

Video content rules the internet
Global Monthly usage in Exa-Bytes (Million GB).
Note: Asia, North America & Western Europe makes up for 80% of the Total
Source: Cisco VNI 2013 – 2018; 2019 & 2020 is authors projection based on VNI.
30+ Billion
Full Movie DVDs
4+ DVD Movies
per person per month
150 Billion
Full Movie DVDs
20+ DVD Movies
per person per month
Mobile Ca. 20% of Total
IP Traffic
CDN 45+% of Total
IP Traffic
60+% of Total
IP Traffic in Metro
Exa-Bytes
Video-only traffic considered

APAC towards 2020 – Total IP Trafic.
Ca. 50+% of World Population lives here!
40% of Global IP Traffic
55% of Total* is Metro-based
60% Consumer IP Video Traffic
40% of Total is CDN-based
2020 APAC Projections:
Source: Pyramid Research 2013 – 2017; 2018 to 2020 is authors own projection.
15+% of Pop have fixed broadband
15% still on 2G
Up-to 20% likely to have LTE
*Total always refers to the Total IP Traffic.
Fixed Broadband Penetration
30% of total IP traffic from mobile.
Exa-Bytes

MEA towards 2020 – Video Traffic only.
Ca. 1 in 5 of World Population lives here!
4% of Global IP Traffic
17% of Total* is Metro-based
Exa-Bytes
74% Consumer IP Video Traffic
14% of Total is CDN-based
2020 MEA Projections:
Source: Pyramid Research 2013 – 2017; 2018 to 2020 is authors own projection.
5+% of Pop have fixed broadband
40% still on 2G
4+% likely to have LTE
*Total always refers to the Total IP Traffic.
Fixed Broadband Penetration
30% of total IP traffic from mobile.

Fundamentals of traffic growth.
Interest
- Growth of customers.
- Growth of traffic (per customer & total).
- Growth of revenue.
- Growth of profitability.
Technology
Adaptation
#Users
Usage
Adaptation
Usage per User
?
LIMITED?
×
Exponential?
S-curve-like?

Growth – technology adaptation.
Technology
adaptation
#Users
(is an IoT a user?)
LIMITED?
(maybe)
Population
Availability
Economics
2014: ca. 40% of APAC1
on 3G or better
2020: 70+% of APAC1
1 Pyramid Research APAC.
Subscriptions
APAC 2014
~ 1 sub per pop.
By 2020
~ 1.2 sub per pop.
What about
Internet of Things (IoT)?
1 Million IoT per km2
The 15 – 64 years

Growth – usage adaptation.
Cellular Usage
Adaptation
Usage per User
?
LIMITED?
(maybe)
Pricing
Use Cases
Technology
Convenience
Spectral capacity
Network Speed
Device performance
Transport infrastructure
20 hrs. per week TV viewing
@ 1Mbps unicast stream
20 GB per Month per user
2014: ~400 MB Cellular!
per Month per User in APAC
2020: ~5 GB Cellular!
Note: CELLULAR TV
Cloud
Cellular off-load

Global mobile revenue structure.
Total Revenue 2014
was ca. 1.0 Trillion US$
Total Revenue 2020
Expected to be ca. 1.4 Trillion US$
+4.5% pa (CAGR)
Note: SMS revenues are blended into the data revenues.
Note! Data + Voice Revenues

Global mobile revenue structure.
(an optimistic view)
ARPU Turn around
Note: SMS revenues are blended into the data revenues. This will in general pull down the pure mobile broadband data revenues.
<2 % of expected global
GDP per capita
<2 % of expected global
GDP per capita
Abbreviated as ARPU

Asia expectations.
+5.5% pa
~ 1.5 % of expected
Asia GDP per capita
~ 1.0 % of expected
Asia per capita
Mobile Asia by 2020:
• Ca. 5.0 B subscriptions.
• Ca. 3.7 B unique users.
• LTE ca. 20%
• 3G ca. 40%
• 2G ca. 40%
• ARPUU ca. 9 US$ / month
• ARPU ca. 5.8 US$ / month
• Total Revenue 0.5+ Trillion US$
• 50+% from data.
Total Revenue 2014
was ca. 0.4 Trillion US$
Total Revenue 2020
Expected to be ca. 0.5 Trillion US$
ARPUU – Average Revenue
per Unique User

44
Customer Economics to Consider.
0 – 25%
25%– 80%
Beyond 80%
ARPU Decline
Customer Growth Slows
Increasing Customer Acquisition Cost
Revenue stagnation &
decline
Profitability Pressure
Attractive
urban areas
All urban &
Sub-urban
areas
Rural Areas
Note: “Crossing the Chasm” is attributed to Geoffrey Moore from his book “Crossing the Chasm: Marketing and Selling High-Tech Products to Mainstream Customers”.

45
Time
USERS
ARPU
SERVICE REVENUES
Time
SMS REVENUE
VOICE REVENUE
DATA REVENUE
TOTAL REVENUE
DIGITZED REVENUES
(The 4th Wave*)
?
* The 4th Wave is attributed to CHETAN SHARMA, MobileFutureForward.
The very basics …
Mistakes & Mess
deadly for profitability!
Mistakes, Incompetence &
Mess don’t really matter!

Old world communication…
When 1 + 1 was 2
... Bla …
Bla bla bla
Mobile Network
We talked (a lot)
We SMS’ed (even more)
Rarely did we use the (mobile) web.

A new usage paradigm …
1 + 1 is no longer “just” 2
1
User
Multiple
Device
User & application initiated bandwidth demand.
Device & application (IP address, keep alive, …) driven signaling resources.
Many applications

50
MOBILE
CONVINIENCE
HOMEDigitized!
RETAIL
Content
SECURITY
DEFENCE
HEALTH
SURVAILANCE
SMART GRID
DATA MINING
TOURISM
PROFESIONAL
SERVICES
TRANSPORT
BIONICS
QoE
ENVIRONMENT

“Todays” Access
Telco Environment
The New
Telco Environment
The new competitive climate.
Enabled by Technology.

52
Mobile
1,400+
Billion US$
(55% Data)
Global Digitized Economy 2020
Fixed
440+
Billion US$
(60% BB)
Mobile Banking
400+
Billion US$
Public Cloud
370+
Billion US$
Mobile Health
60+
Billion US$
M2M
140+
Billion US$
Mobile App
30+
Billion US$
Mobile Digital Advertising
170+ Billion US$
(70+% of Total)
Smartphones
250+ Billion US$
Mobile Content
8+ Billion US$
Managed Cloud Services
4+ Billion US$
Sources: http://www.statista.com/ premium account. Typically
up-to 2020 has been projected based on available data. This
applies to the following page as well.

53
Mobile
1,400+
Billion US$
(55% Data)
Fixed
440+
Billion US$
(60% BB)
Mobile Banking
400+
Billion US$
Public Cloud
370+
Billion US$
Mobile Health
60+
Billion US$
M2M
140+
Billion US$
Mobile App
30+
Billion US$
170+ Billion US$
(70+% of Total)
Smartphones
250+ Billion US$
Mobile Content
8+ Billion US$
Another
Trillion Dollar+
Economy
in the most obvious
Digital Services
On top of Mobile.
Managed Cloud Services
4+ Billion US$
Sources: http://www.statista.com/ premium account. Typically
up-to 2020 has been projected based on available data. This
applies to the following page as well.

54
Mobile
1,400+
Billion US$
(55% Data)
Fixed
440+
Billion US$
(60% BB)
Mobile Banking
400+
Billion US$
Public Cloud
370+
Billion US$
Mobile Health
60+
Billion US$
M2M
140+
Billion US$
Mobile App
30+
Billion US$
170+ Billion US$
(70+% of Total)
Smartphones
250+ Billion US$
Mobile Content
8+ Billion US$
Another
Trillion Dollar+
Economy
in the most obvious
Mobile Digital
Services Entertainment &
Media
2,500+
Billion US$
Travel & Tourism
9,800+
Billion US$
(<10% Online)
Internet of Things
7,000+
Billion US$
Residential Financial
Transaction Volume
5,000+
Billion US$
(50+% Online Penetration)
Note: 2013 had Globaly
ca. 30% Internet Users
Healthcare
11,000+
Billion US$Medicine
1,400+
Billion US$

Spend 3 minutes
writing down 3 bullet points
of what makes
mobile & wireless
technologies attractive?

Writing down 3
treasured items
you would give up
for 1 year of internet access.

Value of internet1
What would you give up a year for internet access.
80 80
75 74
68
48
30
27
22
1 Source: The Boston Consulting Group Report on “The Internet Economy in the G-20”, March 2012.

Value of internet1
Need, Love and then taken for granted?
Perceived Value of Internet
(relative to GDP per Capita)
1 Source: Analysis based on The Boston Consulting Group Report on “The Internet Economy in the G-20”, March 2012.
0%
10%
20%
30%
40%
0% 25% 50% 75% 100%
Perceived Value of Internet
Price of Internet
0%
10%
20%
30%
40%
0% 2% 4% 6%
Internet penetration < 50%
Internet penetration > 50%
Japen & South-Korea
Taken for granted
Internet Penetration
The perceived value of internet drops as internet becomes a commodity
!

Normal price setting in mobile industry.
1 Most price levels are not
designed in isolation from
competition, In fact often
competition is the main
“inspiration” for pricing.
2 Quality could be speed but is
not exclusively so.
Price ( Volume ( Quality, Product, Time ) , Cost, Competition1, Regulation)
Volume
(eg Allowance
vs Unlimited)
Time
Possible
FUP based
feedback
COST
mainly driven by Quality & Product
Quality2
(eg speed,
latency, …)
Products
(eg Bundles)
Illustration
e.g., Interconnect tariffs,
roaming tariffs,
MTR (for voice), …

Dimensioning of pricing
+
Volume Time
ProductQuality
Mobile Data pricing policies
focus on Volume
Fixed Data pricing policies
focus on Speed
sometimes combined
with Volume limits.
Most WiFi pricing policies
focus on Time or
bundled with
mobile data plans
 Mobile bundled products
mainly Voice, SMS, and
Data.
 Fixed bundles with Media,
Broadband Data, Voice &
mobile access (if
available).Illustration

Changing the game!
New philosophies … new dimensions.
Volume
Time
Quality1
Product
Reduce
Cost of
Providing data
Differentiate on Quality.
Speed.
Latency.
Coverage.
Time.
Customer care & support, etc..
Always-Best-Connected
Leverage Fixed and Mobile.
Small Cell deployments.
WiFi / Femto-cell off-load, etc..
Product value add-on
VoIP.
Msg & notifications.
Internet Access.
Social media.
Mobile media player.
Handset, etc.. Illustration
1 Quality could be speed but is not exclusively so.

Pricing fundamental.
Cost
Minimum
Profit
Price
Range
Price Floor
Price Ceiling
Strategic price
Price
Quantity
Subject to Cost
Revenues
=
Price
×
Quantity
Price
leakage
Missing
volumes
Maximize Revenues
Illustration

Pricing fundamental.
Value
to
Customer
Benefits
Cost Profit
Price
Price – Cost = Profit
Benefits = Value / Price
Total
Value
Added
Cost
is a Function
of the Benefits
Price
Ceiling
Price
Floor
Value
Extracted
Value
Ceded
Illustration

Classical cellular pricing.
The old world thinking.
Myanmar Mobile Data Pricing:
Old school data pricing philosophy where data usage comes on-top of
what basic services (e.g., Voice & SMS) a customer has chosen.
Telenor most sophisticated with quality differentiated price depending
on speed range (i.e., up-to 500 kbps and up-to 2 Mbps). Neither MPT
or Ooredoo have quality differentiated pricing.
MPT appears to copy Telenor though does have a high cap data plan
(6.5GB) for 25 US$ (ca. 3.85 US$ per GB).
Telenor has the cheapest data prices at 3.7 US$ per GB if customer is
happy with up-to 500 kbps.
Group Study
Most expensive!
Highest Data Capacity per Customer
But Not the best in Quality class
US$ per GB

Pricing, quality & performance
Group Study
DL Speed UL Speed Ping DL Speed UL Speed Ping DL Speed UL Speed Ping DL Speed UL Speed Ping
ms ms ms ms
Telenor 2.20 1.07 222 2.81 1.35 179 1.41 0.99 142 3.57 1.22 144
Ooredoo 1.97 0.96 203 2.03 0.97 171 2.05 0.72 248 NA NA NA
MPT 1.53 0.71 249 1.82 0.97 189 2.62 0.46 266 1.71 1.06 175
MECTel 0.22 0.27 577 0.20 0.29 587 0.39 0.28 441 NA NA NA
Source: http://opensignal.com/coverage-maps/ for Yangon, Mandalay, Naypyitaw, etc..
Naypyitaw Average
MbpsMbps
Myanmar Average Yangon Average
Mbps
Mandalay Average
Mbps
Date: March 2016
Millions Telenor Myanmar Ooredoo Myanmar MPT Myanmar
US$ 2014 2015 2016 2014 2015 2016 2014 2015 2016
Customers 3.4 13.7 18 - 25 2.2 5.8 6.4 - 7.4 11.0 18.0 18 - 20
Employees 367 500 < 750 949 >1000 < 1000
ARPU 5.4 5.6 5.2 - 5.8 7.5 6.1 5.5 - 6.0
Revenue 39 615 1,200 - 1,600 52 292 450 - 500
Ebitda -68 244 550 - 900 -98 -21 0 - 50
Margin -1.8 40% 45% - 55% -1.9 -7% 0% - 10%
Opex 107 371 650 - 700 150 313 < 450
Capex 573 430 < 350 300 326 < 250
Capex / Revenue 15 70% 25% 6 112% 50%
Ebitda - Capex -641 -186 200 - 550 -398 -347 < -200
Sites 1,500 4,200 7,000 1,200 3,450 5,000
Data Users 52% 60% 80% 80% 80%
Market Share 20% 37% 42% - 48% 13% 15% 14% - 15% 66% 48% 38% - 44%
Source: https://www.telenor.com/about-us/global-presence/myanmar/ & http://ooredoo.com/uploads/misc/Ooredoo_at_a_Glance_Q4_2015_v2.pdf
Authors own prediction
1. Identify the operator who has the best quality in the below table
2. Go to http://opensignal.com/coverage-maps/ type Yangon into search block and on NetworkRank choose Advanced view and toggle
between Telenor, Ooredoo and MPT coverage. How big a %tage difference is there between Download speeds for No. 1, 2 and 3.
3. Who appears to have the best coverage in Yangon? Move to Mandalay or another region and repeat exercise.
The Table below provides an overview of financial results of Telenor and Ooredoo.
4. Compare Telenor &
Ooredoo performance
and identify the
strongest operator.
Explain why?
5. How many customers
would the weaker
operator need to arrive
at the same revenue of
the strongest?
ARPU = Average Revenue Per User
and comprises a blended figure
considering all services.

Data-centric price plans (1 of 2).
The Un-carrier price plan:
• Choose between 2, 6, 10 GB & Unlimited 4G Data plan.
You get beside your chosen data plan:
- 1 Phone.
- Unlimited Voice.
- Unlimited SMS.
- Unlimited data but at reduced speed beyond limit.
- Binge On: unlimited video on most popular streaming
services (Netflix, HBO, Hulu, etc..) … capped at 480p.
- Music Freedom: unlimited music streaming.
- Data Stash: rolls up-to 20 GB of unused 4G data forward.
- No recurring service contract.
$50/mo $80/mo $95/mo
One-off Price = $US 0
Monthly Price = P X GB
+ US$ 42.5 (fixed monthly fee)
Fixed monthly fee of US$ 42.5 covers all
the above beside the data plan.
Note: Unlimited ~ 14 GB @ 3.75 US$/GB
Unlimited average
consumption
US$ 42.5 takes care of all stuff not
covered by the variable 4G data pricing
Group Study

Data-centric price-plans (2 of 2).
handset_price
+ Customer Management Recurring Fee
+ PSMS (sms)
+ Pvoice(minutes)
+ Proaming_insurance(r.data, r.minutes, r.sms, r.geography)
+ Phandset_recovery_fee(terminal type)
Pvolume = 0.86 US$/GB × DataLimit < ∞
Pvolume < Punlimited = P∞ (i.e., unit price → 0)
Pspeed = 0 (in this example).
Source: http://techneconomyblog.com/2015/02/03/mobile-data-centric-price-plans-an-illustration-of-the-de-composed/
Illustration:
£ 26 or US$ 37.5
(for normal handsets)
= 0.86 US$/GB × DataLimit
Group Study

Target cost & design to cost.
Illustration
Revenue = $5.0 per Giga Byte (GB) unit sold (Marketing Wish)
Margin > 30% → Earn >$1.5 per $5.0
Target Cost < (1 – 30%) × $5.0 = $3.5 per GB (maximum)
The CFO View
Technology view:
Once off site investment $168,000 written of over 7 years ~ $2,000 per month.
Spectrum $500 per month per site
Site Lease $300 per month per site
Energy (+fuel) $500 per month per site
Transmission $200 per month per site
Operations $500 per month per site
Tech Cost per Site $4,000 per month → ~ $12,000 per month including all corporate cost.
3 sectors each of 2×10 MHz servicing with 3G
→ Capacity: 10 h traﬃc × 3600 s/h × 10 MHz × 1.4 Mbps/MHz/sector × 3 sectors × 1/8 Byte/bit
→ Capacity: 189,000 Mega-Byte per Day or ~ 4,000 GB per month
Bottom-up we get that minimum cost per GB is $3.0 per GB (i.e., $12,000/4,000 GB).
Group Study

Spectrum
Frequency in Hz
Carrier
Frequency
Channel
Bandwidth

72
The 3G traffic jam!
 3G capacity and quality crunch.
 Slow down migration from 2G3G,
migrate to LTE instead.
 New spectrum demand.
 Re-farming existing 900/1800 MHz
spectrum if possible (in time).
Empty 2G roads - in time?
 5 MHz in 3G will only take up ca. 1 MHz in
LTE.
 LTE mitigates the 3G capacity crunch.
 Re-farmed 2G spectrum too late for
mi ga ng the 3G capacity crunch →
migration to LTE a better option.

73
Spectrum management essential.
 Lots of Hz per customer … high speed!
 Alternative to fixed (xDSL) broadband.
 Higher speed than 3G/HSPA+.
Happy startup … plenty of quality.
 Geometrical growth in demand.
 Start-up quality difficult to maintain.
 Hz per customer drops dramatically.
 Demand for (much) more spectrum
 And many more capacity sites.
Tougher future … growth limitations.

Frequency spectrum – basics (1 of 3).
Frequency in Hz
Illustration (idealized)
Carrier Frequency
Fc
Channel
Bandwidth
B
Bandwidth Examples:
GSM 200 kHz
UMTS 5 MHz
LTE 1.25 – 20 MHz
With Bandwidth aggregation
(i.e., adding up bands)
substantially larger effective
bandwidth can be achieved.
Carrier Frequency examples:
GSM 900 MHz, 1800 MHz
UMTS 2100 MHz (900 MHz)
LTE 700MHz, 900 MHz, 1800 MHz, etc..

Road analogy of frequency & bandwidth.
Channel
Bandwidth
Width
of the Road~ The wider the road the more cars
can I support simultaneously
BChannel = WRoad
Coverage Length of
Carrier Frequency
L  1 / Fc
~
Length of the road
with a given Width
How long a distance
can I support a given
traffic volume of cars
How long a range can I
support a given traffic
demand of data
The wider the channel bandwidth
the more data traffic can I support

Paired spectrum.
Frequency in Hz
Frequency Division Duplex (FDD) - Illustration (idealized)
Carrier Frequency
FUplink
Channel
Bandwidth
BUL
Carrier Frequency
FDownlink
Channel
Bandwidth
BDL (can be > BUL)
Duplex
Separation
Channel
Spacing
or
Amplitude
Uplink:
from User to Network
Downlink:
from Network to User

Uplink
Out of City
Downlink
Into of City
Road analogy of FDD.
Frequency division duplex.

Un-paired spectrum.
Frequency in Hz
Time Division Duplex (TDD) - Illustration (idealized)
Carrier Frequency
FTDD
Channel
Bandwidth
BTDD
Guard period
Downlink
Uplink
Time
From User to Network
From Network to User
 Whether to use TDD or FDD depends primarily on spectrum availability.
 Both paired (FDD) & un-paired (TDD) technologies have benefits & disadvantages.

Downlink
Into of City
Road analogy of TDD.
Time division duplex.
Uplink
Out of City
360
seconds
120
seconds
Same road for
Inbound (DL)
as well as
Outbound (UL)
traffic

29-15 MHz @ 900
25-35 MHz @ 1,800
Typical 210 MHz – 215 MHz
For in-country merged operators can be 220 MHz - 225 MHz
Cellular frequency spectrum overview.
1 HSPA= HSDPA + HSUPA, 2 Including 10MHz for E-GSM, 3 Values of Spectral Efficiency tends to change a lot depending on antenna
technology and actual field data, 4 Typical value.
GSM / GPRS / EDGE UMTS R99 HSDPA
(Improved DL)
HSPA1
(Improved DL & UL)
LTE
0.032 – 0.128 0.064 – 0.384 0.384 - 4
2.5 (Avg.) to
14.4peak
30 (Avg.) to
170peak+
(450, 850,) 900,
1,800, (1,900) MHz
(AWS: 700, 1,400, 1,700,) 2,100 MHz,
2,600 MHz UMTS extension band (though also LTE candidate)
All Freq. UHF band &
700 – 2,600 MHz.
Min. 220 MHz
(target)
2352 MHz @ 900
2x75 MHz @ 1,800
260 MHz @ 2,100 and 270 MHz @ 2,600
AWS: 245 MHz @ 1,700
2n20 MHz
(n=1,2,3…)
Downlink Throughput (Mbps)
0.11 – 0.454 0.51 0.80 1.44 1.5 to 10
Downlink Spectrum Efficiency3 (Mbps per MHz)
Dominated by legacy infrastructure suppliers: Nokia Networks, Ericsson, Huawei, ZTE.
Note: FDD: Frequency Division Duplex Cellular Systems (i.e., operates in two separate frequency
bands; 1 for Downlink & 1 for Uplink).
Illustration only.

Coverage fundamentals
Frequency & length of traveling wave.
E.g., A male voice* reaches up-to double the distance of a female voice*
purely based on its lower frequency range (all else being equal).
(*) Male voice frequency 85 – 180 Hz vs Female voice frequency 165 – 255 Hz.
The effective reach
of a given wireless technology
is inverse related to
the carrier frequency.
E.g., lower frequencies cover
more length (& area) than high
frequencies.
The loss of signal power
over a given distance
is inverse related to the
square of the carrier frequency.

Spectrum benchmarking – coverage.
900 MHz
DL power
Coverage area
UL power (typical limitation for coverage)
Illustration
×9 ×6 ×4.5 ×1
900 MHz – 800 MHz (digital dividend)
2.6
GHZ
2.6 GHz
Available bandwidth for LTE
LargeVery small
LowHigh
190 MHz
2.1 GHz 1.8 GHz
2×60 MHz 2×75 MHz 2×35 MHz 2×30 MHz

Quiz
1. What carrier frequencies are the most valuable for coverage?
a) Low frequencies (<2100 MHz).
b) High frequencies (>1800 MHz).
c) Carrier Frequency itself is not valuable, bandwidth is the valuable property.
2. Which statements below are correct? (could be more than one!)
a) FDD divides the frequency spectrum up in individual codes.
b) FDD stands for Forestry Defence Department.
c) FDD divides a frequency spectrum into two bandwidth parts, with a frequency
separation between uplink use and downlink use.
d) TDD stands for Time Division Duplex.
e) In China TDD is the most popular implementation of LTE.
f) FDD is better than TDD.

Frequency link budget.
Transmit Tx
(path) Loss L ≤ 1
g × Tx
Receive Rx = L × g × Tx
Gain g ≥ 1
Convention:
in dB whereLink Budget

Spectrum efficiency.
How many bits per second can I transport per Hz of bandwidth.
Frequency in Hz
Illustration (idealized)
Channel
Bandwidth
B in Hz
Spectrum Efficiency =
Information rate (bits per
second) that can be support
by a given technology and
available bandwidth in Hz
In general, the higher spectral efficiency the better technology!

Road analogy of spectral efficiency.
Width
of the Road
WRoad
5 cars per second
Per
Width of Road
Old Road
10 cars per second
Per
Width of Road
Safe
distance
Width
of the Road
WRoad
New Road Next Generation Road
21 cars per second
Per
Width of Road
Width
of the Road
WRoad
Technology
Improvement
Technology
Improvement

Capacity fundamentals.
Unit Capacity = Bandwidth in MHz × Spectral Efficiency (in
Mbps
MHz − capacity unit
)
Dimension of Unit Capacity is Mbps/capacity-unit
B
Available BW in MHz per unit Spectral Efficiency in Mbps/MHz/ unit
N
Number of units

T-Mobile UK
& Orange TD-TV
Examples of frequencies & bandwidth.
TDD & FDD.
UL
(75 MHz)
DLUL
(35MHz)
DL
UL
(70 MHz)
TDD
(50 MHz)
DL
900 MHz 1,800 MHz
2,500 MHz
UL
(60 MHz)
DL
2,100 MHz
TDD part TDD
2,300 – 2,400+ MHz
part TDD / part FDD
This band provides interesting backhaul P2P options in some
Greenfield scenarios
3,400 – 3,800+ MHz
China: SD-CDMA alloc.
UL DL
400 MHz
DL UL
700 MHz
TDD
T
D
D
(20 MHz) (15 MHz)
(20 MHz)
Airway in China
TDD
BSNL in India

City coverage benchmarking.
For dense cities beside coverage being relative insensitive to frequency the effective cell range
decreases with increasing population density..
NYC
Den Haag
Houston
Leeds
LA
Chicago
Berlin
Hamburg
London
N
2
33
A
r 
A: Covered (hexagonal) Area
N: Number of cell sites
Houston
1
1 TMUS GSM spectrum at 1,900 MHz, and their 3G in the AWS1700 band.
0.20
0.40
0.60
0.80
1.00
0 2,000 4,000 6,000 8,000 10,000
City Pop Density (pop/km2)
GSM900
GSM1800
UMTS2100
Effective Cell Radius in km (City Coverage Characteristics)

Coverage (1 of 2)
Low-frequencies (<1,800 MHz) provides excellent coverage options while higher
frequencies with more available bandwidth gives higher speed performance.
Typical Cell Range (km)
0.01
1
10
100
1,000
LTE450
HSPA 2100
UMTS 2100
GSM900
GSM1800
GPRS
EDGE
LTE 2100
LTE 2600
Femto-cells/3.6GHz
WiFi
Note: Illustrational purpose only real cell sites can vary greatly as well as can the actual performance
Voice
10 km
0.1
Economical very attractive for sub-urban
to rural & deep indoor coverage
Economical attractive for urban
areas & capacity demand
indoor
Site throughput in Mbps vs cell range in km

Coverage (2 of 2)
Frequency and bandwidth determines the technical as well as economical
performance of a given access technology.
Typical Site Range (km)
Voice
0.1
1
10
100
1,000
High
Frequency Low
Frequency
High
Bandwidth
Low
Bandwidth
Below
900 MHz
Above
1,800 MHz
Below
10 MHz
Above
20 MHz
Site Throughput in Mbps (equivalent user capacity)

Capacity fundamentals.
CAPACITY Ci = BANDWIDTH Bi
MHz
× EFFICIENCY ηi
Mbps per MHz per Cell
× CELLS Ni
#
Business as Usual New spectrum New technologies New macro
×
Innovation Re-farming Improvements Small-cells
×
Radical Spectrum sharing Spectrum sharing Site sharing
Note: Sub−script i referes to a relevant cellular clutter area; thus the Total Capacity Ci

( )
= Bi×Ei×Ni

i(Areas)
VERY COSTLY
(VERY) COSTLY
EFFICIENT
(VERY) COSTLY
COMPLEX + EFFICIENT
COMPLEX BUT EFFICIENT, QoE CHALLENGE
BaU (COSTLY)
BaU (COSTLY)
B
× $ / MHz-pop
(e.g., 1 – 2 $/MHz-pop)
η
Modernization
Terminal subsidies
N
Cell splits / overlay
New Sites
Within technology up-to 20% gain
upto ceiling.
Between technologies x2-3 gain
Can be a signifiant capacity multiplier
and result in new site avoidance.
In urban areas can be
difficult to achieve
more density (new
sites).

93
ConnectivitySpectrum is the growth engine.
MHz
× EFFICIENCY Ei
× CELLS Ni
#
GSM
Ctot ≈ 40k
n x 0.2 MHz (TRX)
(n: 6 – 15)
~ 0.52 (GSM)
0.14 – 0.33 (EDGE)
80k – 100k
(Utilization ≈ 50+%)
Frequencies
Total Bandwidth
900 & 1800 MHz
Ca. 110 MHz
UMTS
Ctot ≈ 600k (×15 GSM)
n x 5 MHz (carrier)
(n: 2 – 4)
0.5 – 1.2 (average)
up-to 17 (peak)
80k -100k
(Utilization* ≈ 70+%)
Frequencies
Total Bandwidth
2100 (900) MHz
60 (95) MHz
LTE
Ctot ≈ 4,500k (×8 UMTS)
n x 5 MHz (carrier)
(n: 6 – 10+)
1.5 – 2.0 (average)
Up-to 30 (peak)
80k – 100k
Frequencies
Total Bandwidth
700 -900MHz, 1.8GHz, 2.5GHz
210+ MHz
Note: Above only FDD spectrum is considered. Bandwidth are represented by DL part (i.e., total BW = 2x(DL or UL for symmetric bands).
(*) pending on terminal type and application a single customer can in theory cause the a given cell to be highly utilized in terms of bandwidth resources.
LTE is supported over a very wide frequency range
from 450 MHz up to 3.6 GHz

Spectrum is a very valuable asset.
UK 2.1GHz ca. 35 Bn US$
NL 2.1GHz ca. 2 Bn US$
DE 2.1GHz ca. 38 Bn US$
IN 2.3GHz ca. 5.5 Bn US$
US 700 MHz ca. 20 Bn US$
Telenor paid $500M
for 2×5 + 2×10 MHz
Blended average of
$0.31/MHz/pop

UMTS Euphoria 2000.
- In March 2000 Mobile industry paid ca. 35 Billion
US$ for 3G spectrum (record).
- Almost $5 per MHz per pop.
- ×3 the total UK mobile revenue in 2000.
- ×8 the total UK mobile Ebitda in 2000 (estimate).
- On-top they would need to invest at least 20
Billion US$ in 3G network infrastructure.
- They would need minimum 8 – 10 Billion US$ FCF
per year (over 10 years) to reach an NPV 0.
- ×12 the total UK mobile FCF in 2000 (estimate).
- At least 15+ years to breakeven on cash.
- The internal business cases (at the time) would
have had to be very optimistic to finance what
was paid for the spectrum.
- Value should have far exceed the 35 Billion
US$ paid for spectrum including
deployment investments.
- In July 2000 Mobile industry paid ca. 2.0 Billion US$
for 3G spectrum.
- Ca. $2 per MHz per pop.
- Approx. the total NL mobile revenue in 2000.
- ×2.5 the total NL mobile Ebitda in 2000 (estimate).
- On-top they would need to invest at least 1.5 Billion
US$ in 3G network infrastructure.
- They would need minimum 0.5 Billion US$ FCF per
year (over 10 years) to reach an NPV 0.
- Approx. the total NL mobile FCF in 2000 (estimate).
- The internal business cases (at the time) was
optimistic but not as aggressive as in UK.
- Value should have far exceed the 2 Billion
US$ paid for spectrum including deployment
investments.
Telenor paid $500M
for 2×5 + 2×10 MHz
$0.31/MHz/pop (1/6 NL)
GDP per Capita is 30× lower

Quiz
1. A Technology using 900 MHz covers an area
a) Worse than for a frequency of 1800 MHz?
b) Better than for a frequency of 1800 MHz?
c) Too little information to answer question?
2. Which of the following set of parameters are important for providing cellular network
capacity?
a) Frequency, Number of rainy days, Number of Sites.
b) Number of Sites, Bandwidth and Number of Customers.
c) Number of Sites, Number of bits per second per Hz available from deployed technology and
Available Spectral Bandwidth.
3. My LTE customers demand 100 Giga bit per second (Gbps) in downlink. I
have 2 x 10 MHz available for LTE with an effective  of 2 bps/Hz/unit. How
many units do I need to support the demand?
a) 500 capacity units.
b) 250 capacity units.
c) 5,000 capacity units.

Facebook Drone Coverage
The Basic Economics
“The Drone Coverage Network is an exponential technology in the sense that it has the ability to
Disrupt existing terrestrial-based cellular coverage networks
by a factor of 10 or more on TCO and deployment-time.”
“Facebook’s ambition is to built 10,000 Acquila drones which could more than easily
cover all land based surface area.”

Coverage solutions.
Caution: drawing is not to scale!
Nano-satellite
FB Aquila Drone
Terrestrial
Cellular Tower
30 – 70 meter
10 – 50 km
< 2000 km
DataQuality
Cell Center
Drone Coverage
Terrestrial
Coverage
Distance from center
Practically
10 km range FB estimate ~ 80 km
Note: FB Aquila envision Laser
connectivity to house holds (beyond
fiber connectivity) although
mentioned but not described
wireless coverage as well.
=
GEO ~36,000
km
10 Gbps
Laser beam

Facebook Drone Coverage Network vs
standard MNO Cellular Network Coverage.
Drone Network Coverage is
10× more Capex Efficient.
10× more Opex Efficient.
Typically more than 10× faster deployment.
Highly scalable capacity provision & options.
Support all frequency ranges up-to mm-wave; thus also
standard cellular/wireless frequencies 700 MHz to 5 GHz.

Facebook Drone Coverage – Aquila.
Drone /
Unmanned Arial Vehicle (UAV)
10 – 50 km
Stratosphere
Wingspan 42m
<450kg
Up-to 80 km
< 20 thousand km2
• Envisioned as a constellation of 3 drones circling in
the stratosphere providing covered to an area of up
to 20 thousand km2.
• Up-time up-to 3 month (solar powered).
• Laser backhaul with 10 Gbps connectivity.
• Myanmar Example:
• Myanmar surface area is 676,578 km2 and we would thus
require ca. 30 Facebook drone constellations (i.e., 3
drones).
• Providing max WiFi speeds across coverage area.
• Cellular network providing up-to 80% geographical
coverage may require up-to 10,000 cellular sites, between
1.5 to 2.0 Billion US$ in Capex and easily several hundred
millions of US$ in annual Opex.
• Providing max speed in vicinity of tower and
increasing poor quality out to cell edge with 128 kbps
- 256kbps.
Unlicensed WiFi
or Cellular Freqs.*
10 Gbps
Laser backhaul
(*) Should Facebook acquire cellular frequencies or cm/mm-wave
frequencies this would likewise be straightforward to deploy via a
Drone-based coverage network.
Wiki Note: global pop per HH is ~3.5, world surface area 510 Million km2
of which ca. 150 million km2 is land area of which ca. 75 million km2 is
habitable. 3% is an upper limit estimate of earth surface area covered by
urban development, i.e., 15.3 Million km2.
Note: FB Aquila environs Laser connectivity to house holds
although mentioned but not described wireless coverage as
well.

1
0
2
102
Towards Next-Generation Telco Technologies
and Business Models.
Converged Access Internet of Things
Internet of Industries
Connected Vehicles
Converged Apps Vehicular Autonomy
E2E Latency < 1 ms
Very High Redundancy.
Medium BW requirements
E2E Latency ~ 1 - 50 ms
Very high availability
Elastic BW requirements
E2E Latency ~ 50+ ms
Privacy & Security protection
Extreme elastic BW requirements
Ultra-high Availability
Very high security required
Elastic E2E Latency requirements
FTTx + LTE  5G
Ultra-Efficiency requirements
Cloud & Virtualization
Seamlessness across all platforms
e.g., Fixed, Mobile & other screens
Ultra-Personalization
Industry
4.0IoA
IoT

Internet-of-Things (IoT).
A craze of very big numbers of small things connected.
- 4.3 Trillion US Dollar Market by 2024*.
- 2.4× that of the Mobile Industry Revenue.
- 27 Billion IoT units installed by 2024*.
- 4 × that of unique mobile users.
- Up-to 1 million IoT units per km2.
- An urban macro cellular site might have
to serve up 3 million IoT units.
IoT Requirements:
- (ultra) low device cost.
- (ultra) low power consumption.
- Near-zero maintenance.
- Very long battery lifetimes.
- Versatile connectivity.
- Elastic latency ( 1 ms to seconds).
- Elastic Bandwidth ( bps to Mbps).
- Massive scalability  106 per km2(*) https://machinaresearch.com/news/the-global-iot-market-opportunity-will-
reach-usd43-trillion-by-2024/

Quiz (1 of 2)
• By 2024 27 Billion Global IOT connections are expected.
• World population is expected to be 8 Billion with ca. 3.5 people per household (HH).
• Planet Earth total surface area is 510 Million km2.
• Land area is ca. 150 Million km2
• Ca. 75 Million km2 is habitable.
• Max. 15 Million km2 is covered by urban development.
1. How many IOT connections do you have per HH and per km2 considering the total surface
area of Planet Earth.
2. How many IOT connections do you have per km2 considering only land area.
3. How many IOT connections do you have per km2 considering only urban development.
4. Does the amount of IOT devices per Household seem realistic? Compare with the expected
number of mobile devices per HH?
5. Assume that a typical urban cell site area of is ca. 3 km2 how many IoT connections do you
get that a cell site would be required to support?

Quiz (2 of 2)
• By 2024 27 Billion Global IOT connections are expected.
• On average an active IOT connection will use 140 Bytes (approx. size of SMS).
• On average an IOT is active 60 times per minute (60x24x7x365:-).
1. How many Mega Bytes will an IOT connection consume per day?
2. How many Mega Bytes will an IOT connection consume per month?
1. Compare that to the global average smartphone data consumption 2015 (ca. 1,200 MB).
3. How many Giga Bytes will an IOT connection consume per year?
4. What is the total amount of Exa Bytes (i.e., Billion Giga Byte) consumed by of all IOT
connections?
5. If by 2024 the total data consumption excluding IOT is in the order of 400 Exa Bytes (i.e.,
400 Billion Giga Bytes), what would the proportion of IOT data consumption be if included
in the total data consumption?

Global IoT growth projections.
2014 – 2024.
~ 2 IoT Connections
per 3 people
~2.5 IoT connections
per Household
~13 IoT connections
per Household
Note: global pop per HH is ~3.5, world surface area 510 Million km2 of which ca. 150 million km2 is land area of which ca. 75 million km2 is habitable. 3% is an upper
limit estimate of earth surface area covered by urban development, i.e., 15.3 Million km2  300+ IoT per km2 in 2014 and 1,700 IoT per km2 in 2024.
~33 IoT connections
per km2 land area.
~180 IoT connections
per km2 land area

Global IoT revenues*.
(*) https://machinaresearch.com/news/the-global-iot-market-opportunity-will-reach-usd43-trillion-by-2024/
4.3 Trillion US$
CAGR 17%
0.9 Trillion US$
Ca. 160 US$/IOT/Yr
Ca. 13 US$/IOT/Yr
IoT Device Revenues  approx. 52 US$ per IoT.
IoT Connectivity (Access)  approx. 3.7 US$ per IoT per Yr.
approx. 0.3 US$ per IoT per month.
M2M/IoT Service  approx. 44 US$ per IoT per Yr.
approx. 3.7 US$ per IoT per month.
ALL REVENUES

Narrow Band (NB) IoT.
Many flavours to Internet-of-Things.

IoT structural ecosystem.
Infrastructure Transport Traffic Buildings Farm Emergency Factories
Edge Gateways.
Fiber Cellular WiFi Powerline Satellite Drone
Billing & Operations
Support Systems
Billing & Operations
Analysis
Service Platforms
1..N
QoS & Service
Management
Network
Management
Big Data Analytics
SmartCity
SmartFarm
Home Automation
Factory Automation
SmartRoads
eHealth
Etc etc…
End-users:
SME,
SOHO,
Municipality,
Private, …
User or
Service
Provider
owned
Telecoms,
Network
Providers,
MVNOs,
Others
Value Add
service
providers
Services

Industrial revolutions.
~1870s
Electrical.
Steam engines scales.
Industrial Iron making.
~1950s
Change from
mechanical &
electronic
technologies to
digital ones.
~1780s
Mechanical.
First factories & mass-production,
e.g., particular textiles.
~1880s
First Electric Power Plants
1st
2nd
3rd
4th Revolution?

Industry 4.0 (4th industrial revolution)
Whats in it for the Industry (Europe)?
- 50% of all capital investment until 2020.
- In Europe alone up-to 140 Billion Euro pa.
- 18% boost in productivity.
- 110 Billion Euro pa in additional revenues.
- Lead to massive transformations.
Definitions:
Cyber-physical system: workplace carriers,
assembly station & products.
Requirements:
- Interoperability with IoT & Internet.
- Virtualization  plant & simulation
models.
- Localized autonomy.
- RT capability ~ highly elastic latency reqs.
- Service orientation.
- Highly customizable / Modularity.

Network fundamentals (simplified).
Air-interface
• GSM
• UMTS/HSPA
• LTE/LTE-Adv
• WiFi
• 5G, etc…
Cell
• BTS
• Node-B
• E-Node-B
• AP
Cell
Access/Backhaul
• Microwave (MW)
• Leased Line (E1)
• Fibre / Cable
• Gigabit Ethernet
• Satellite
Backbone
• Fibre/IP
• Microwave IP
Aggregation
• BSC
• RNC
• IP Aggregation.
• Access Cloud.
Core Network
Data Center Cloud
IT VAS & BSS.
Packet Switched (PS/IP).
Circuit Switched (CS).
Signalling Network.
External World
- Foreign Public Land Mobile
Networks (FPLMN).
- Public Switching Telephony
Network (PSTN).
- WWW = The Internet.
Other mobile
networks
(FPLMN)
Fixed Networks
(PSTN)
WWW
MW
n×E1
Fiber
Fiber

What’s a Cloud?
Source: http://www.businessnewsdaily.com/4864-cloud-computing-terms.html, note: FW: Firewall, SW: Software.
Cloud Service are referenced to a computing service
that resides in a secure and (possible) centralized
remote (for the customer) platform.
Customers procuring a cloud service from a Cloud
Service Provider purchase only what they use leading to
substantial Capex and Opex savings (avoidance)
SaaS – SW as a Service:
Offering SW (residing in the
cloud) on a subscription basis
based on per-user basis.
Use-Case: Replaces traditional
on-device SW.
Examples: Microsoft Office 360,
Google Apps, Salesforce, Cisco
WebEx, SAP, ….
PaaS – Platform as a Service:
Providing strong infrastructure &
SW applications for building,
testing & launching new
applications.
Use-Case: Increase developer
productivity & utilization rates
while decreasing time-2-market.
Examples: SAP, Apprenda,
Openshift, AWS Elastic Beanstalk,
Cloud Foundry, Google App, …
IaaS – Infrastructure as a Service:
Provides computing, storage,
networking & services (e.g., FW).
Use-Case: Instead of purchasing
HW, users purchase IaaS based
on consumption (like electricity).
Examples:, Amazon Web
Services, Microsoft Azure, Google
Compute Engine, …
Examples pf Cloud-based services

Benefits of cloudization.
Networking
Storage
Servers
Virtualization
Operating System
Middleware
Execution
Data
Application
Legacy
All Internally
Managed
Networking
Storage
Servers
Virtualization
Operating System
Middleware
Execution
Data
Application
Infrastructure
as a Service
IaaS
Networking
Storage
Servers
Virtualization
Operating System
Middleware
Execution
Data
Application
Platform
as a Service
PaaS
Networking
Storage
Servers
Virtualization
Operating System
Middleware
Execution
Data
Application
Software
as a Service
SaaS
Source: http://news.microsoft.com/download/archived/presskits/cloud/docs/The-Economics-of-the-Cloud.pdf
Internally managedInternally managed
All being managed
by cloud provider(s)
It is possible to mix
cloud providers.
It is possible to mix
cloud providers.
Could be from
different provider

Basic business model for the cloud
Platform
as a
Service (PaaS)
Infrastructure
as a
Service (IaaS)
Software
as a
Service (SaaS)
Pay as you Use Pay as you Grow
Network
as a
Service (NaaS)
Customers can save or altogether avoid IT Infrastructure and substantially reduce need of IT Staff

The Top Cloud Providers.
As of 2014.
Infrastructure
as a Service (IaaS)
Platform
as a Service (PaaS)
Software
as a Service (SaaS)
Storage
as a Service
Security
as a Service
X as a Service
(XaaS)
Other Cloud Service examples

When you see a network cloud
Per Cloud Data Center:
- 500+ Million Euro Investments.
- 100s of thousand of servers.
- 100s of Megawatt power.
- 1000s of TFlops.
(Illustration only representing Google,
Facebook, Apple DCs).
E.g., Deutsche Telekom’s Biere
DC (2014) has approx. 15k
servers over 3k square meters.
(Data Center)
Cloud
100s thousand
homes equivalent

Virtualization (simply) defined.
Driven by vastly improved computing power at increasing lower cost
as well as cheap high performance storage.
Traditional Architecture
Resources underutilized (often significantly so).
Sharing resources difficult.
Virtual Architecture
Sharing resource easy,
secure & flexible.
Microsoft AppsWindows 10
Linux
Hadoop
Mac OS X
Photoshop
Windows 10
Microsoft
Desktop Apps

Basic requisites for economic benefits of
cloudization and virtualization.
Reliable, readily available and relative cheap energy & real-estate.
(Low cost) high quality fiber optical networks for local connectivity available.
Low cost high availability international (high) bandwidth
to tier-1 traffic destinations.

Economics of cloudization.
Biggest benefits relative to existing cost structure to be expected with
legacy data center providers, ISPs and businesses where ICT operation is
not a core business.
Study materials: http://news.microsoft.com/download/archived/presskits/cloud/docs/The-Economics-of-the-Cloud.pdf
Large-scale data centers (DC) should provide lower cost per server
from simple economics of scale considerations.
Aggregating computing demand provides statistical multiplexing gain
resulting in increased server utilization & efficiency.
Multitenant application model (e.g., Microsoft Office 365) reduces the app
management and server cost per tenant.

Caution on the Business Case
of Cloud & Virtualization.
Old
Legacy
DC Design
with
1:1
physical
server
to Services
(non-
virtualized)
Modern
DC Design
Fully
virtualized
60% to 70%
Cost reduction*
(*) Comparison should always be based on absolute cost at the same occupancy or efficiency rate. Be careful about relative TCO metrics only unless you can
check the absolute cost level as well. See also Cisco “The Economics of Cloud Computing” by Bill Williams.
Un-negotiated, as-is & older
HW (i.e., no optimization)
Newly negotiated, newest HW,
Optimized (high) efficiency
• Do understand your baseline!
• What is the Baseline Cost including TCO (Total Cost of Ownership)?
• What are the limits to optimization within legacy network?
• Include the potential of new sourcing.
• What is the cost per service? E.g., Server TCO, Storage TCO, Maintenance per Service, etc…
Modern
DC Design
Virtualized
Optimized
NG
DC Design
Virtualized
Optimized
< 20% Cost reduction*
At this time and age the above
scenario is relative rare.
At this time and age the above
scenario is the more normal one.
Newly negotiated, newest HW,
Optimized (high) efficiency
Un-negotiated, as-is & older
HW (i.e., no fully optimization)

Cost of basic DC Elements.
Always know the details of your baseline cost.
Networking
Storage
Servers
Virtualization
Operating System
Middleware
SW Execution
Data
Application
Facilities
Resources / FTE
Maintenance
Ancillary
3rd Party Services
Including transport cost
SAN, NAS, HDFS, etc..
mirror sites required?
COT or supplier specific
Might not be relevant in
baseline
Opensource or
supplier specific
Other cost elements
to be considered
End user SW & data
handling e.g., Salesforce.
Runtime.
Degree of
outsourcing
Note these cost elements
should minimum be
considered whether you
are a cloud provider or
cloud customer!
IaaS
PaaS
SaaS
Energy,
Water, etc..
Follows from Apps & OS
New Revenues?
Does Cloud –
Virtualization
leads to new
services?
Or more
efficient service
delivery?

Telco Cloud & Virtualization Economics.
Telco cost structure impact of migrating to cloud & virtualization.
Capex impact:
• Maximum 40% of Telco Capex likely to be positively impacted by Cloudization &
Virtualization (i.e., IT & Core).
• Of the 40% approx. 50% would be infrastructure & the remaining part Software-
driven (new development 20% & maintenance 80%).
• Transport (i.e., Backbone/interconnect) cost (up-to 20%) could increase.
• Really depends on the degree of centralization & whether traffic remains in country or
needs international transport.
• Upfront investment (unless outsourced to 3rd party) would be required to design
and build DC to host Telco IT and Core Network functionalities.
• Assuming that Telco builds up own Cloud capabilities: overall Capex benefits
(avoidance) should be expected to be relative minor as premium HW is replaced by
premium SW (although HW cost is reduced, the cost of SW tends to increase) and
the upfront investment to prepare DC for legacy migration cloud/virtualization.
• On IT specific Capex max 40% avoidance should be expected compared to a legacy
IT environment wo virtualization (i.e., Greenfield comparison).
• In complete Outsourcing model substantial part of the relevant Capex (i.e., 40%)
will disappear although additional transport invest may be required.
potentially
negatively
impacted
Potentially
negatively
impacted
Opex impact:
• Maximum 25% of Telco Technology Opex expected to be impacted by
Cloudization & Virtualization (i.e. IT & Core).
• Complete Outsourcing model: One need to consider the total cost including
the benefits of Capex Avoidance (see above). While part of the legacy Opex
should be expected to reduce (e.g., from personnel, energy, other) 3rd party
service cost would add Opex to the cost structure.
• Depending on how efficient legacy IT & Core were operated minor overall Opex
saving (<10%) could be expected.
• Assuming that Telco builds up own Cloud Capabilities: One should not expect
more than net 10% to 20% Opex savings on the relevant Opex (i.e., 25%) and
only after complete migration. During migration overall Opex is likely to be
higher than legacy only.
Group Study

Summary
Relative benefits to cost structure of cloudization & virtualization.
E-commerce businesses can avoid own IT infrastructure & need for substantial IT
staff, benefit from economical PAYG&U* licensing (i.e., very low barrier to start-up).
Legacy Data Center Providers & large non-Telco ISPs leverage scale of their
existing DC infrastructure (i.e., increased business on same infrastructure).
Telcos, where IT & Core Network cost structure is relative minor, will achieve
improvements on relevant total cost although overall it might not be a big change.
(*) PAUG&U: Pay As You Grow and Use.
Highest absolute economical benefits of cloud & virtualization are achieved in
Greenfield Scenarios with pure IT environments.

OSI goes soft.
Open Systems Interconnection model.
Data Link Layer
(frame, e.g., MAC, PPP,)
Network Layer
(packet, e.g., IPv4, IPv6, ..)
Physical
(raw bit stream, e.g., DSL, Ethernet, …)
Transport Layer
(segment TCP / datagam UDP)
Session Layer
(data, e.g., HTTP, SMTP, …)
Presentation Layer
(data)
Application Layer
(data)
1
2
3
4
5
6
7
Good reference: https://en.wikipedia.org/wiki/OSI_model, another good reference from Duke University: http://people.ee.duke.edu/~romit/courses/f07/material/
Data plane
Control plane
NFV
Network
Function
Virtualization
SDN
Software
Definable
Network
Raw
Transmission
Relievable Transmission
(start – end of stream)
Address, routing & control
(re)Transmission between
network points
back-and-forth transm.
between two node
Translate between
network services & app.
High-level API – closest to
the user SW.

Software definable network (SDN).
Driven by vastly improved computing power at increasing lower cost.
• Implement Control Plane in software.
• Decouple control plane from data plane.
• Using commoditize routers & switches.
SDN is Intelligence centralized in a
controller that manage commodity
devices controlled by imposed
policies & configurations
Application
Servers
Access
Aggregation
Core
Routers
Classical Data Center
Application
Servers
SDN Domain
Core
Routers
SDN-based Data Center
SDN Controller
Programmatic
(SW) control plane

Network Function Virtualization (NFV).
Driven by vastly improved computing power at increasing lower cost.
• Implement Data Plane in software.
• Decouple network elements from hardware.
• Using commoditize hardware (e.g., OTS servers).
NFV ensures that Virtual Devices
are configured remotely &
provisioned instantly on the OTS
server farms (in the Data Centres).
HSS
Policy
IMS
NAT DPI
SGSN SMSC
…
Service Provider (e.g., MNO) Domain
Customer Domains
Core
Router
Edge
Router
Fixed
Mobile WiFi
NetworkFunctions
Classical Architecture:
Mixed supplier
landscape with
proprietary
implementations.
Service Provider (e.g., MNO) Domain
Customer Domains
Core
Router
NFV
Service
Insertion
Point
Fixed
Mobile WiFi
NFV
NFV-centric Architecture:
OTS Server, Storage & switches
Software
HSS
Policy
IMS
NAT DPI
SGSN SMSC
Virtual Edge Router
Etc…
Off The Shelf

(*) Marc Andreesen, WSJ, 2011 (source: http://www.wsj.com/articles/SB10001424053111903480904576512250915629460). Recommended reading.

Front-end
Cloud
DC
Towards 5G in 2020.
5G / LTE
Small Cells
FTTH
Access
Cloud
DC
Software
Definable
Network
(SDN)
The Access Cloud is a
Data & Computing Center supporting
access & edge functionality
Max 50 km
Front-end
Cloud
DC
Back-end
Cloud
DC
Other
Networks
Depending on country size and size of
network more than 2 front-end
100 Gbps
10 Gbps
Access
Cloud
DC
Software
Definable
Network
(SDN)

GSM,
GPRS & EDGE
Global System for Mobile
communications
The 2nd, 2.5 & 2.75
Generation

Why GSM?
Prior to GSM
 Fragmented analogue systems (NMT, E-TACS, TACS, CNETZ, ...).
 No inter-operability between existing mobile standards.
 No inter-operability with fixed telephony networks.
 Poor scalability of existing technologies.
 National security concerns (The Russians).
 Common European Market requiring a Europe-wide standard.
 Growing demand for mobile telephony.
NMT: Nordic Mobile Telephony, E-TACS: European Total Access Communications System.

GSM – a 30 year old technology.
1982: 890 – 915 MHz (UL) and 935 – 960 MHz (DL) was reserved for a Cellular
System: 2×25 MHz (to be reserved across European Union member states).
1985: Decision to implement a digital cellular system.
1987:
Field trials completed in Paris.
GSM working group concludes that best technology would be based on Time
Division Multiple Access (TDMA) & Frequency Division Multiple Access
(FDMA).
Memorandum of Understanding initially signed by 12 countries.
1991:
First GSM call and service launched in Finland.
UK, France, Germany and Italy introduces digital services based on GSM
standard (phase 1).
1993: Explore GSM migration towards UMTS.

GSM operational requirements.
 High Audio Quality.
 High Spectral Efficiency.
 Identical systems in all countries.
 International roaming.
 Open architecture.
 Economical both in sparsely and in heavily populated areas.
 Integration with fixed digital networks (e.g., ISDN).
 Security features (e.g., Cold war era).
 New features, e.g., Short-Message Service (SMS), Data, and Fax.
 Easy system introduction.
 Low-cost infrastructure.
What was the
benchmark?
Explain why?
Explain why?
What benefits?
Explain why
important? Discuss what that
means?
ISDN: Integrated Services Digital Network … set of digital communications standards enabling simultaneous voice, video, data and other service using the
traditional circuits of a given PSTN. Basically ISDN provides also access also to packet switched data networks on top of old copper infrastructure.
Group Discussion

Basic business models of GSM.
Wireless Mobility Identity & security
Scalability Mobile TelephonyInterworking
PLMN
PSTN
PLMN: Public Land Mobile Network, PSTN: Public Switched Telephone Network, also known as Plain Old Telephone Service (POTS) network.
@kbps
SIM

GSM customer.
+
Mobile Station (MS)
= SIM + Device
Note! When operators count subscribers
they often do it by number of SIM cards
All Subscribers are assigned a
International Mobile Subscriber Id (IMSI)
• IMSI is the only absolute identity a subscriber
has within the cellular system.
• IMSI is not hardware specific, e.g., like IMEI or
MAC address.
• IMSI is used in GSM, UMTS & LTE.
MCC – 3 digits
Mobile Country Code
MNC – 3 digits
Mobile Network Code
MSIN – max 9 digits, first 3 digits = HLR-id
Mobile Station Identification Number
National Mobile Station Id (NMSI)
IMSI
MSISDN CC – 1- 3 digits
Country Code
NDC – Variable
Nat. Destination Code
SN – Variable
Mobile Station Identification Number
Mobile Station ISDN Number – max. 15 digits
Mobile (Telephone) Number
Unique customer id
IMEI International Mobile Equipment Id … unique identifier of equipment used by the subscriber.
Unique device id

GSM services.
Overview of Phase 1 to 2+ (most important items).
Tele Services:
• Mobile telephony (13 kbps voice).
• Incl. Half-Rate speech coding.
• Incl. Enhanced Full Rate
• Emergency calling (irrespective of subscription &
credit).
• SMS (i.e., 160 chars)
• Fax
Bearer Services:
• CS Data (300 – 9600 bps, 14.4 kbps in 2+).
• High-speed circuit switched data (HSCSD, 2+).
• General Packet Radio Services (GPRS, 2+) with
maximum speed of 115 kbps (theoretical).
• Enhanced Data for GSM Evolution (EDGE, 2++)
with maximum speed of 474 kbps (theoretical).
Network Features:
• Network Identity & time zone (NITZ, 2+)
• CAMEL Phase 1 (prepaid roaming).
• Support for Optimal Routing (SOR), particular
applicable to roaming.
Supplementary Services (most important):
• Call Forwarding (note doesn’t work for SMS).
• Call Barring (some phones supported for SMS).
• Calling Line Identification (CLI).
• Call Waiting.
• Call Hold.
• Multi-party communications (MPTY, up-to 5).
• Advice of Charge (AoC, online charging info).
• Unstructured Supplementary Services Data
(USSD, operator-defined individual services),
often used for prepaid charging and in roaming.
• Operator-determined barring (ODB, operator
restrict features for individual subscribers).
• Closed User Group (CUG).
• A few others …

GSM FDMA & TDMA.
Frequency & Time Division Multiple Access.
1
2
3
4
5
6
8
7
200 kHz
Channel
Bandwidth
8 Users
(or Time Slot)
per
Channel
F1, UL
902.3 MHz
Carrier Frequency
Mobile to Base Station (UL)
1
2
3
4
5
6
8
7
200 kHz
Channel
Bandwidth
8 Users
(or TS)
per
Channel
F1, DL
947.3 MHz (+45 MHz)
Carrier Frequency
Base Station to Mobile (DL)
20 MHz
890
MHz
915
MHz
935
MHz
960
MHz
In GSM this is also called a
TRX (2×0.2 MHz)
45 MHz
Duplex separation
Total of 125 Channels
Max, 1,000 Simult. Users
25 MHz / 0.2 MHz
125 Ch × 8 Users/Ch
Time Slot can be
transmitted at various
specified intensities
(or power levels)
1 Time Slot (TS) possible
reserved of signaling.

144
Connectivityspectrum is the growth engine.
MHz
× EFFICIENCY Ei
× CELLS Ni
#
GSM
Ctot ≈ 40k
n x 0.2 MHz (TRX)
(n: 6 – 15)
~ 0.52 (GSM)
0.14 – 0.33 (EDGE)
80k – 100k
Frequencies
Total Bandwidth
900 & 1800 MHz
Ca. 110 MHz
UMTS
Ctot ≈ 600k (×15 GSM)
n x 5 MHz (carrier)
(n: 2 – 4)
0.5 – 1.2 (average)
up-to 17 (peak)
80k – 100k
Frequencies
Total Bandwidth
2100 (900) MHz
60 (95) MHz
LTE
Ctot ≈ 4,500k (×8 UMTS)
n x 5 MHz (carrier)
(n: 6 – 10+)
1.5 – 3.0 (average)
Up-to 30 (peak)
80k – 100k
Frequencies
Total Bandwidth
700 -900MHz, 1.8GHz, 2.5GHz
210+ MHz

Basic GSM network architecture.
Supporting Voice and SMS and ISDN-data connections.
Cells
Cells
BTS
BTS
BSC
MW
Radio
Fixed Line
Core Network
HLR
EIR
MSC
IT – VAS (e.g., SMSC, VMS, IN, MMSC,
IVR,..) , Billing, Rating, CRM, …
PSTN
FPLMN
AUC VLR
Many BTS to 1 BSC
Many BSC to 1 MSC
Several Cells per BTS
Air-interface
Home Location Register
Subscriber details & services allowed
Authentication Center
Authenticate each SIM attempting to connect.
Equipment Identity Register
List of phones banned or monitored. Visitor
Location
Register
Temporary
Data from HLR
& MS.MS
MS
Interconnect to
external networks

IT & VAS landscape (data center).
Billing Platform
(post-paid billing)
Prepaid Rating &
Charging Platform
(Real-Time)
CRM: Customer Relationship
Management.
ERP: Enterprise Resource
Planning Platforms (e.g., SAP,
HR, Clarify…).
SMS-C: SMS Center
VMS: Voice Mail
System
IN: Intelligent
Network (for
prepaid)
Bulk SMS-C:
Wholesale SMS
M2M Server: Machine-
2-Machine Server
IVR: Interactive Voice
Response Server.
MMSC: Multi-Media
Messaging Center
<100 km on Fiber
SAN (secondary): Storage Area Network
SAN: Primary
Customer
Fraud
Voucher
Device
CEM Customer
Experience Management
OSS: Operations
Support Systems
BSS: Business
Support Systems
IllustrationTelco Network, Intra- & internet
Shops, Call Center(s), Etc..
Web Portal(s)

GSM elements & basic functions.
HLR VLR
MSCBSCBTS
Backbone
(transport: SDH)
Backhaul
(transport: MW, LL)
GW
PSTN
FPLMN
Air-interface
MS
MS – Mobile Station = SIM + Device.
- Authentication & Authorization.
- Transmit & receive voice & data
over the cellular system.
- Measuring surrounding cells for
optimum performance.
- Encoding & encryption of signals.
BTS – Base Transceiver Station.
- Transmit & receive radio signals
from MS over the air interface.
- Decode & decrypt signal from MS.
- Encode & encrypt signal to MS.
- Radio condition measurements
from MS.
- Each cell under a given BTS is
served only by that BTS.
Interconnect
(transport: SS7, SDH)
BSC – Base Station Controller.
- Radio Resource management for
BTS under BSCs control.
- Handover (inter-cell).
- Capacity & Quality optimization.
- Traffic concentration towards MSC.
- Operations & Management
interface for whole Base Station
Subsystem.
Base Station Subsystem (BSS) Network Switching Subsystem (NSS)

GSM elements & basic functions.
HLR VLR
MSCBSCBTS
Backbone
(transport: SDH)
Backhaul
(transport: MW, LL)
GW
PSTN
FPLMN
Air-interface
MS
MSC – Mobile Switching Center.
- Switching of all calls.
- Signaling (control)
- Paging.
- Location registration.
- Call setup of all MSs in area.
- Handover management.
- Dynamic resource allocation.
- Encryption.
- Billing for all subscribers in area.
- Interworking with other networks.
- SMSC Gateway.
- Possible to have several BSS
controlled by 1 MSC.
HLR – Home Location Register.
- subscriber related information.
- Record of supplementary services
subscription for each customer.
- Permission control granting access
to supplementary services.
- Some data are permanent, others
temporary changing depending on
customer movements & actions.
- Data Stored (perm): IMSI,
MSISDN, Roaming restriction,
supplementary services
parameters, AUC parameters, MS
category, …
Interconnect
(transport: SS7, SDH)
VLR – Visitor Location Register.
- Often integrated into the MSC.
- Supports MSC in storage and
retrieval of subscriber info within
its area.
- Info stored is temporary as it will
reside there only as long as
subscriber is within related MSC
area.
- Roaming customers will reside here
as well.
Data Stored: IMSI, TMSI, MSISDN, LAI,
MS category, Supplementary services
parameters, AUC parameters, MSC id, …
Base Station Subsystem (BSS) Network Switching Subsystem (NSS)

Basic GSM Architecture.
HLR VLR
MSCBSCBTS
Backbone
(m × 155 Mbps)
Backhaul
(n × 2 Mbps)
GW
PSTN
FPLMN
n × 0.2 MHz
 ~ 0.52 Mbps/MHz
e.g.,, 9 TRX per site 
~ 1 Mbps backhaul which
typically would result in
leasing an E1 (2Mbps) / T1
(1.5Mbps) or 2 or
4x2Mbps MW
Typically 155 Mbps (e.g., STM1)
could support between 45 – 55
BTS with 2 Mbps transport
solution (at 70% utilization).
STM - Synchronous
Transmission Module
defined on Synchronous
Optical Networking
(SONET) / Synchronous
Digital Hierarchy (SDH).
Can support bit rates of
up-to 38 Gbps (STM256)*
(*) SDH: https://en.wikipedia.org/wiki/Synchronous_optical_networking, STM1: https://en.wikipedia.org/wiki/STM-1

Basic GSM network structure.
PLMN Service Area
(e.g., Telenor Myanmar Mobile Network)
MSC VLR MSC VLR
MSC VLR MSC VLR
MSC Area
(e.g., MSC Yangon)
Location Area
LA1
LA3 LA2
LA5
LA6
LA4
Cell1 Cell2 Cell3
Cell4
Cell8
Cell12
Cell6
… …
… …
PLMN: Public Land Mobile Network, MSC: Mobile Switching Center, VLR: Visitor Location Register, LA: Location Area.

GSM Hierarchy.
PLMN
MSC Area
MSC Area
.
.
.
.
.
.
LA1
LA2
LA3
LA4
.
.
.
BSC1
BSC2
.
.
.
BTS1
BTS2
.
.
.
.
BTSn
BSCj
LAi
Cell1
Cell2
Cell3
Site location
3 Sector per BTS
Location
Regional assignment
Regional assignment
Operator’s footprint
Every BTS radio transmitter transmit a Location Area
Identity (LAI) = MCC + MNC + LAC
LAI
Mobile
Country
Code
(MM 95)
Mobile
Network
Code
(OM 05)
Location
Area
Code
(MNO def.)
MS
VLR HLR
Mobile Station (MS) location
LAI info resides here
Note: depending on the size of the BSC (i.e., number of cell associated with it) it is possible for a BSC to be supporting more
than one LA. Similarly within a BTS it is possible that a cells might be allocated to different LA’s (though in general should only
be the exception than the Rule).
For GPRS/EDGE a routing
areas is defined within LA

I have “forgotten” something … but what?

Subscriber Identify Module (SIM).
1991 20122008
• The SIM card is an integral part of the overall GSM System Architecture as well as
any cellular standard evolved from that such as UMTS and LTE.
• Without a working SIM the mobile device will not be able to connect to the
network (beside for emergency services).
• The SIM card uniquely identify the subscription & telephone number (MS-ISDN)
upon which billing or rating/charging (for prepaid) is based.
• Note 1 subscriber can have several SIM cards (and of course devices).
SIM securely store the user’s unique id associated the
particular cellular network:
IMSI: International Mobile Subscriber Identity.
IMSI =
MCC Mobile Country Code, e.g., 414 for Myanmar.
+ MNC Mobile Network Code, e.g., 05 for Ooredoo.
+ MSIN Mobile Subscription Identification Number (10 digit).

Adding data to GSM – GPRS & EDGE.
Cell
Cell
BTS
BTS
BSC
MW
Radio
Fixed Line
Core Network
HLR/AUC
(EIR) VLR
MSC
PSTN
FPLMN
WWW
SGSN GGSN
IT – VAS (e.g., SMSC, IN, VMS, IVR,..)
, Billing, Rating, CRM, CEM,…
Circuit Switched
(CS) Domain
Packet Switched
(PS) Domain

Coverage & capacity example.
Yangon city:
• Yangon Area ~ 600 km2.
• Yangon Population ~ 6 Million.
• Average Density 10,000 per km2.
• GSM effective range 1 km per site.
• Unit Coverage Area =

= 2.6 km2
• Ca. 230 sites to cover Yangon.
• Ca. 690 sectors with 3 sectors per site.
• Note: given population density site range best
practice should be lower than 1 km!
• Market share 25%.
• Customers ~ 1.5 Million.
• Minutes per Month per Customer ~ 100 Min. of Use (MoU).
• 5 minutes per day and ca. 1 minutes in Busy Hour (BH).
• Total minutes per BH 1.5 Million Minutes.
• With 690 sectors we have
• ~ 2,200 customers per sector (over BH).
• ~2,200 BH minutes per sector (demand).
• Operator has 2×5 MHz @ 900 MHz for GSM.
• 25 TRX (i.e., 5 MHz / 0.2 MHz) available in network.
• Re-use pattern of 7 implying 25/7 ~ 3.5 TRX per sector.
• At 2% GoS and 3.5 TRX (or carriers) per cell this will only supply
ca. 1,200 MoU per Sector about half the required capacity.
 Result in very poor quality with severe service level.
• Options:
• Built more sites: ~ 1,250 sectors, +560 sectors ~ 187 sites.
• More to Re-use pa ern of 4 → 6 TRX per Sector → 2,200
MoUs @ 2% GoS (however also a lot more interference).
See also: http://www.wirelesscommunication.nl/reference/chaptr04/cellplan/reuse.htm
& http://docplayer.net/697830-Dimensioning-and-deployment-of-gsm-networks.html
Group Discussion

Quiz
1. What initially made GSM so attractive to consumers?
a) Cool handsets (e.g., iPhone)
b) Wireless & enabling mobile.
c) You got your own SIM card.
2. Which statements below are correct? (could be more than one!)
a) GSM is based on TDMA and FDMA (8 timeslots for each 200 kHz carrier).
b) GSM was developed because engineers had nothing better to do.
c) GSM spectrum structure is TDD (time division duplex).
d) GSM strived for higher spectral efficiency (supporting large amount of customers)
using digital technology.
e) GSM introduced the SIM in order to improve identification of customers and provide
better security against unlawful tapping of conversations.
f) SIM stands for Subscriber Identity Module and uniquely identify the customer.
g) GSM voice services are based on circuit switched technology.
h) With GSM a unified mobile technology was created with sufficient scale and
economics that it would be affordable for most consumers and operators.

UMTS & HSPA.
Universal Mobile Telephony System &
High Speed Packet Access.
3rd & 3.5th Generation Mobile Systems.

Principles behind UMTS.
Wide-band Code Division Multiple Access (W-CDMA).
FDMA Traffic Channels: different frequency bands
are allocated to different users. Very common in
analogue telephony systems, e.g., NMT & AMPS.
TDMA Traffic Channels: different time slots & frequency
channels are allocated to different users. Common in 2nd
generation digital mobile systems e.g., GSMT & D-AMPS.
CDMA Traffic Channels: different users are
assigned unique code and usage transmitted
over the same frequency band, e.g., UMTS &
CDMA2000 (USA standard).
Wide-band simply distinguish system
from narrow-band (1.25 MHz)
Power
Power
Power
Source: based on Huawei presentation.

Understanding the Code in UMTS.
Chinese
Chinese
Chinese Danish Danish
Italian
Italian
Hungarian
Hungarian
“The Cocktail part analogy to CDMA”

Basic business model of UMTS.
Internet in your pocket!
Vision anno 2000 – 2001
(iPhone was still 7 – 8 years away!! Future = Nokia)

The smartphone … from 2008 onwards
The “killer” device and its “killer” applications…

164
Source: Pyramid Research.
Voice Revenue Growth suffers as
penetration of Broad Band increases!

15.7
Voice
SMS
Appenomics not so great for operators.
Apps “attacks” the highest margin. services.
22.4 1
12.7
1 Source 2010 & 2015 Pyramid Research, Western Europe.
2010A
MNO Centric
ARPU
2015E
Apps Centric
ARPU
(Free) OTT VoIP and Messaging Apps can lead to dramatic loss
of MNO revenue and margin.
15.7
Voice
3.4
SMS
3.3
Data
6.5
Voice
6.2 1
Data
?By 2015
more than 70%
of users have
a smartphone
+
9.7+
Data
ARPU
ARPU
Death to SMS.
VoIP @ Home
& Work.
43%
Missing

166
Mobile broadband…
feels like this?
INDUSTRY FEAR

168
NOTE: WiFi is just a bridge to better cellular small-network systems become main stream with controlable spectrum assets and E2E Customer Experience Management.
@ Work
(2 – 4 Cells)
@ Home
(2 – 3 Cells)
On the
Go
@ Home
(1 – 2 Cells)
On the
Go
00:00 10:00 12:00 22:0017:006:00 8:00
voicedata
Small Cells
14:00
Femto Cell Femto Cell
On-load potential
WiFi
+18 month
The digital consumer.

Customer cellular trends to be considered.
Source: The “Cellular Data Usage” distribution based on detailed data mining study, Mature Market Illustration.
80% of a customers traffic is delivered to no more than 3 cells
Up-to 30+ Cells
2 – 4 Cells 1 -3 Cells

UMTS spectrum and bandwidth.
Frequency
in MHz
Channel
Bandwidth
5 MHz
Channel
Bandwidth
5 MHz
Duplex Separation
190 MHz
Power
1,920 1,980 2,110 2,170
60 MHz
12 UMTS Channels of 5 MHz
UL DL
Other MHz bands are also in use for UMTS:
• 1,850 - 1,910 (UL) + 1,930 - 1,990 (DL) (USA PCS-band)
• E-GSM 880 – 915 (UL) + 925 – 960 (DL).
• Most other bands initially earmarked for UMTS are
being deployed for LTE.
• Both 2100 and GSM900 bands are being re-farmed to
LTE or planned to be.
UMTS DL Spectral efficiency:
• R99 η = 0.15 Mbps/MHz
• HSDPA η = 0.9 Mbps/MHz
• 2015 η = 1.5 Mbps/MHz
UMTS2100
Typically operators have between
2 to 3 UMTS Channels

171
MHz
× EFFICIENCY Ei
× CELLS Ni
#
GSM
Ctot ≈ 40k
n x 0.2 MHz (TRX)
(n: 6 – 15)
~ 0.52 (GSM)
0.14 – 0.33 (EDGE)
80k – 100k
Frequencies
Total Bandwidth
900 & 1800 MHz
Ca. 110 MHz
UMTS
Ctot ≈ 600k (×15 GSM)
n x 5 MHz (carrier)
(n: 2 – 4)
0.5 – 1.2 (average)
up-to 17 (peak)
80k – 100k
Frequencies
Total Bandwidth
2100 (900) MHz
60 (95) MHz
LTE
Ctot ≈ 4,500k (×8 UMTS)
n x 5 MHz (carrier)
(n: 6 – 10+)
1.5 – 3.0 (average)
Up-to 30 (peak)
80k – 100k
Frequencies
Total Bandwidth
700 -900MHz, 1.8GHz, 2.5GHz
210+ MHz
(450 MHz – 3.6 GHz)

172
Connectivity
UMTS voice capacity
leapfrogging voice capacity compared to GSM.
Max. 8 voice users
Per MHz/Sector
Max. 12 voice users
Per MHz/Sector
Max. 48 voice users
Per MHz/Sector
Source: “Voice over LTE (VoLTE)” by Miikka Poikselka, Harri Holma, et all (Wiley 2012).
UMTS

UMTS maximum DL speed performance.
Only achievable under ideal conditions and the effective real
experience of customers will be a lot lower.
5 MHz 10 MHz
15MHz
20 MHz
to
40 MHz
40 MHz
Time
R99
R11

UMTS maximum DL performance.
Only achievable under ideal conditions and the effective real
experience of customers will be a lot lower.
Max. Speed in Mbps = Max. Spectral Efficiency η in Mbps/MHz × Available Spectrum in MHz
Normal Range of MHz
Max. Limit for most MNOs

Basic UMTS network architecture.
Cell
Cell
Node-B
Node-B
RNC
MW
Radio
Giga-bit Ethernet /
Fibre Optical connections
HSS VLR
MSS
NG-IT – VAS, Billing, Rating, CRM, Big
Data (structured & unstructured), …
PSTN
FPLMN
WWW
MGW
Policy AAASGSN
GGSN
RRUs
RRUs
RRU: Remote Radio Unit moves the Rx & Tx from cabinet
to or near the antenna.
Inc. AUC
Node Server
Baseband
OSS
Circuit Switched
(CS) domain
Packet Switched
(PS) domain

Packet Switched (PS) Domain
CS
Core
Basic UMTS architecture.
HSS VLR
MSS
RNCNode-B
IP Backbone
(10 – 100 Gbps)
Fiber
IP Backhaul
(typically ~ 100+ Mbps)
Fiber or MW
MGW
PSTN
FPLMN
Bandwidth = n × 5 MHz
mean up-to 1.2 Mbps/MHz per sector
peak up-to 17 Mbps/MHz per sector
Circuit Switched (CS) Domain
Cloud-RAN
(optional)RRUs
• Moving traditional
functionality (e.g.,
Baseband or BS
Server) from site
to “centralized”
cloud.
• Radio Resource
pool
SGSN
IP /
Core
WWW
Not required
for Cloud-RAN
RRU: Remote Radio Unit, RNC: Radio Network Controller, Radio HSS: Home Subscriber Server, VLR: Visitor Location Register, MSS: Mobile Switching Sever, MGW:
Mobile Gateway, SGSN: Serving GPRS Support Node, GGSN: Gateway GPRS Support Node, AAA: Authentication, Authorization & Accounting.
GGSN
AAAPolicy
optional
@ 15 MHz DL on 3 sectors →
Average Site throughput ~ 54 Mbps.
Peak Site throughput ~ 378 Mbps (~ 10 ms)

Quiz
1. UMTS was develop with what in mind?
a) Higher data rates supporting internet in your pocket.
b) To make Apple & Google happy.
c) An simple extension of GPRS & GSM.
2. UMTS is based on?
a) FDMA only (frequency division multiple access).
b) TDMA (time division multiple access) & FDMA only.
c) CDMA (code division multiple access).
3. The frequency carrier bandwidth of UMTS is?
a) 200 kHz.
b) 1.25 MHz, 3 MHz, 5 MHz, 10 MHz, 15 MHz and 20 MHz.
c) 5 MHz.
4. What has happened to voice usage and revenue after introduction of UMTS?
a) Voice traffic & voice revenue has increased after UMTS.
b) Voice traffic & voice revenue has decreased after UMTS.
c) Voice traffic has kept increasing but voice revenue has decreased after UMTS.

& LTE-advanced.
Long Term Evolution (advanced).
4th Generation Mobile Systems.

179
UMTS growth pains.
Initially with UMTS.
Tougher future … UMTS growth limitations.
 Need for much higher spectral efficiency.
 Need for more economical technology.
 Need for much more spectrum.

Why do we need LTE?
”The writing on the wall” … The urgency of getting something better than UMTS.
UMTS spectral efficiency cannot make up for
increased mobile data demand1
Mounting cash pressure resulting in
End-to-profit exposure for MNOs with UMTS
Time
Spectral supply
Spectral demand
For UMTS only
Breaking Point
2014 - 2016
Western Europe assessment
Business
model
breakdown
Mounting cash
pressure
CAGR 33%
CAGR 26%
Illustration
The real “scissor effect”:

LTE status as of End-2015.
Source: http://gsacom.com/paper/spectrum-bands-used-in-480-commercially-launched-
lte-networks/
More than half
(552 Million)
of that was
gained during
2015!
Comprising
157 Countries.
54+% from APAC

The LTE Vision (2004).
Care should be taken as the LTE Vision was derived comparing HSPA Release 6 (2H2014).
• Spectral efficiency 2 – 4 × HSPA which was ca. 0.9 Mbps/MHz.
• Achieved using OFDMA (Orthogonal FDMA) in DL and DS-FDMA (Single Carrier FDMA) in UL.
• For details see “LTE for UMTS – Evolution to LTE-advanced” by Harri Holma & Anti Toskola (Wiley)
• Peak data rates exceed 100 Mbps DL & 50 Mbps UL.
• 10+ × the performance of HSPA at the time.
• Round trip time (RTT) of 10 ms.
• HSPA RTT was 40+ ms at the time.
• Highly flexible spectrum bandwidth; 1.5 – 20 MHz.
• Remember UMTS/HSPA came in 1 flavor, 5 MHz.
• Optimized end user’s terminal power efficiency.
History:
• Standard approved in End-2007.
• First commercial LTE networks launched 2010.
• By January 2016; 1+ Billion subscriptions & 480 commercial networks.

The LTE advanced requirements.
• Carrier aggregation of contiguous and non-contiguous spectrum allocations.
• 100+ Mbps peak for highly mobile users (350km/h).
• 1+ Gbps peak for nomadic and stationary users.
• Scalable system bandwidth up-to 100 MHz.
• Asymmetric bandwidth assignments for FDD (e.g., 20 MHz UL / 40 MHz DL).
• and a lot more really cool stuff.
History:
• Standardization started in 2008. Standard approved 2011.
• First commercial handsets in 2014/2015 (1 in 2013).
• 116 operators (24%) launched LTE-advanced (primarily for carrier aggregation).

FDD-LTE commercial launches.
GSM legacy band of 2×75 MHz usually
easy to re-purpose.
(all-round coverage, good capacity)
Lots of spectrum (70MHz) available,
usually virgin spectrum.
(good for small cells & urban coverage
with good capacity)
Relative scarse but in cellular sense
usually virgin spectrum.
(very good coverage, limit capacity)
In use for UMTS and thus relative
little capacity available now
(overlay coverage to existing UMTS)
Asia and South America.
old broadcast / analogue tv spectrum
(very good coverage & good capacity)

187
MHz
× EFFICIENCY Ei
× CELLS Ni
#
GSM
Ctot ≈ 40k
n x 0.2 MHz (TRX)
(n: 6 – 15)
~ 0.52 (GSM)
0.14 – 0.33 (EDGE)
80k – 100k
Frequencies
Total Bandwidth
900 & 1800 MHz
Ca. 110 MHz
UMTS
Ctot ≈ 600k (×15 GSM)
n x 5 MHz (carrier)
(n: 2 – 4)
0.5 – 1.2 (average)
up-to 17 (peak)
80k – 100k
Frequencies
Total Bandwidth
2100 (900) MHz
60 (95) MHz
LTE
Ctot ≈ 4,500k (×8 UMTS)
n x 5 MHz (carrier)
(n: 6 – 10+)
1.5 – 3.0 (average)
Up-to 30 (peak)
80k – 100k
Frequencies
Total Bandwidth
700 -900MHz, 1.8GHz, 2.5GHz
210+ MHz

188
ConnectivityLTE voice capacity
LTE voice capacity is on par or substantial better than UMTS.
Max. 8 voice users
Per MHz/Sector
Max. 12 voice users
Per MHz/Sector
Max. 48 voice users
Per MHz/Sector
Source: “Voice over LTE (VoLTE)” by Miikka Poikselka, Harri Holma, et all (Wiley 2012).
LTE

Basic business models of LTE.
More of the same same just better!
Fixed substitution Smartphones Faster 4 Zero Xtra
Scalability ConvergenceLarger Data Plans

Basic LTE network architecture.
Cell
Cell
BTS (GSM)
Node-B (3G)
eNode-B (LTE)
pool
Fibre (1 – 10 Gbps)
HSS
NG-IT – VAS, Billing, Rating, CRM, Big
Data (structured & unstructured), …
RRUs
RRUs
RRU: Remote Radio Unit
Cloud
RANFibre (1 – 10 Gbps)
SGSN
Evolved
Packet
Core
(All-IP)
S-GW
Policy
RADIUS
MME
Most LTE EPC functions likely to end
up as NFV in MNO DC.
P-GW
IMS
IT Service Cloud
Backend Data Center
Connect to
2G & 3G
Shared with
2G/3G CS domain
Expanded
AAA function
e.g., PCRF
More complex
antenna solutions
→ higher spectral
efficiency
VoIP /
VoLTE

LTE terminal categories (backup).
Up-to 3GPP Release 11
(September 2012)
Cat 1 Cat 2 Cat 3 Cat4 Cat5 Cat6 Cat7 Cat8 Cat9 Cat10 cat11 cat12
3GPP Release 8 8 8 8 8 10 10 10 11 11 11 11
DL Speed Mbps 10 50 100 150 300 300 300 3000 450 450 600 600
UL Speed Mbps 5 25 50 50 75 50 100 1500 50 100 50 100
Max. #DL-SCH transport blocks Rx
in a TTI
10,296 51,024 102,048 150,752 302,752 299,552 299,552 2,998,560 452,256 452,256 603,008 603,008
Max. # bits of a DL-SCH transport
blocks Rx in a TTI
10,296 51,024 75,376 75,376 151,376 tbd Tbd tbd
Total # of soft channel bits 250,368 1,237,248 1,237,248 1,827,072 3,667,200 3,667,200 3,654,144
359,827,7
20
5,481,216 5,481,216 7,308,288 7,308,288
Max. # of supported layers for
spatial multiplexing in DL
1 2 2 2 4 2 or 4 2 or 4 8 2 or 4 2 or 4 2 or 4 2 or 4
Max. # of bits of an UL-SCH
transport block Rx in a TTI
5,160 25,456 51,024 51,024 75,376 tbd Tbd tbd
Support for 64 QAM in UL No No No No Yes No Yes Yes Yes Yes Yes Yes
Support for 256QAM in DL No No No No No No No No No No optional Optional
MiMo DL Optional 2x2 2x2 2x2 4x4 4x4 4x4 8x8 4x4 4x4 4x4 4x4
Note: In 3GPP Rel 12 a new set of terminal categories was defined: Cat0, Cat13 – Cat16.

What will 5G bring?
Standardization still work-in-progress.
• Indoor user experience of 1 Gbps DL & 500 Mbps UL.
• User experience @ cell edge: 300 Mbps DL & 50 Mbps UL and mobilities
up to 100 km/h.
• Latency from 1 - 10 ms.
• IoT support 1 – 100 kbps, E2E latency seconds – hours at connection
densities of up to 200 thousand per km2. Target device autonomy
(lifetime) up-to 15 years.
• Full Conversion of fixed and mobile services.
• Elastic and dynamic service slicing (i.e., “on-the-fly” service provisioning
end-2-end).
Standardization timelines ~ initial submission June 2019 and detailed
submission by October 2020.
Source: https://www.ngmn.org/uploads/media/NGMN_5G_White_Paper_V1_0.pdf

1
9
5
195
Towards 5G and beyond.
Converged Access Internet of Things
Internet of Industries
Connected Vehicles
Converged Apps Vehicular Autonomy
E2E Latency < 1 ms
Very High Redundancy.
Medium BW requirements
E2E Latency ~ 1 - 50 ms
Very high availability
Elastic BW requirements
E2E Latency ~ 50+ ms
Privacy & Security protection
Extreme elastic BW requirements
Ultra-high Availability
Very high security required
Elastic E2E Latency requirements
FTTx + LTE  5G
Ultra-Efficiency requirements
Cloud & Virtualization
Seamlessness across all platforms
e.g., Fixed, Mobile & other screens
Ultra-Personalization
Industry
4.0IoA
IoT

5G Operators
Service Revenues
forecasted to exceed
65 Bn USD
by 2025
(5 Years after launch)
LTE was roughly
170 Billion USD
after 5 years.
Source: http://www.juniperresearch.com/press/press-releases/5g-revenues-forecast-to-exceed-65bn-usd-by-2025
Total MNO
Revenues by 2020
~ 1,400 Billion USD..

Basic business model
Light-speed
for $0-Access Internet of Things Ultra-low latency
Entertainment Full ConvergenceCloud services
Fixed
Mobile
TV
Tactile Networks

Ultra-low latency (tactile) networks.
100 ms > 20 ms < 5 ms
Seeing movements
without feeling them
1 seconds
Edge-Computing will be important to
achieve ultra-low latency communication
<2 meter <2 meter
100 km/h100 km/h100 km/h
Electronic Protection/Brake System
(in platooning)
Between cars requires
communications ~ 1 ms or lower.
Within a car:
Wheel Synch <5ms
~ 10 – 15 cm
Example: http://www.peloton-tech.com/Average reaction times: ~250 ms to a visual stimulus, 170 ms for audio
stimulus, 150 ms for a touch stimulus.

Latency & Technology
Latency in milliseconds (ms = 1/1000th second)
5G
target
LTE TTI RTT
2 ms (No HARQ!)
Note: network measurements are likely to yield (a lot) worse
latency results than above would indicate.
Light in Optical fiber travels ca. 200 km in 1 ms
(some variation depending on wavelength).
Car driving 150 km/h moves ca. 40 mm in 1 ms
(Note: the breaking distance on dry road would be ca. 125 meters and
would take several thousand ms to achieve).
A steel ball falling to the ground (on Earth)
would have moved 5 micro meter (that’s 5
millionth of a meter) in 1 ms.
In a 1Gbps stream, 1 μs correspond to ca. 125
Bytes worth of data.

Why low latency is important (1)
Baseball … 5 ms between winning and loosing.
160 km/h
4 cm per ms
~ 250 ms eye – catch
coordination
(reaction time)
With 5 ms delay in sight and actual position
would miss the ball with 20 cm @ v = 4 cm per ms
Gone by

Why low latency is important (2)
Driving … maybe not such a big deal?
340 km/h *
Ca. 65 meters
(< 9%)
Thinking time
Ca. 680 meters
Braking distance
Alert driver on a dry road!
The 65 meters could be saved by having an A.I. driver by autonomous driving
Note: (*) by 50 km/h (5 ms = 7 cm) thinking distance is ca. 10 m (40%) and braking distance 15 m (60%). In autonomous driving/breaking up-
to 10 m or 40% of the stopping distance could be cut away
5 ms = 0.47 meter

Connected Cars – Strive of <5 ms latency.
Network Assisted Driving.
Connected
Car (Edge) Cloud
Backend Cloud
(Operator Data Center)
Insurance
Company
Government
Base Station
HSPA/3G (<50 ms)
LTE/4G (10 -20ms)
5G (<10 ms)
<50km*
SSD: Solid-State Drive, HSPA: High Speed Packet Access, LTE: Long-term Evolution, RTT: Round Trip Time.
(*) G.655 λ ~ 1310 nm – 1625 nm, η ~ 1.47, distance of 204 km/ms, (**) convolutional neural network used to identify traffic situation (lane & vehicular detection).
Today 2 – 5+ ms
Queuing delay.
0.5 ms
Transport RTT
<5 ms
For wheel Synch (e.g., EPS, ABS)
In-car pc:
SSD seek time <0.2 ms
Processing < 30 ms (e.g., cnn**)

Illustration
Beyond today’s cellular networks.
A critical enabler for the Digitized Society!
Router
Edge
Cloud
Front-end Data Center (FE-DC)
Cloud SDN
with NFV
FE-DC
PSTN
FPLMN
Edge Computing
LTE/5G
100 Gbps
10 Gbps
Target Min 100+ Mbps at consumer edge.
Delay supported down to 1 millisecond.
Network slicing supporting large span of service categories & specifications.
FE-DC: Front-end Data Center, SDN: Software Defined Network, NFV: Network Function Virtualization, PSTN: Public Switching Telephony Network, FPLMN:
Foreign Public Land Mobile Network, LTE: Long Term Evolution.
Backend
Data
Center

205
Network Architecture
New business models will appear.
NG-DC, Cloud, NVF, SDN, Metro-CDNs, NG-transport!
Disruptive Business Models
3rd parties
delivers
core
network
functionality
(e.g., HSS,
PCRF, etc..)
3rd parties
delivers BSS /
OSS cloud
services (off-
the-shelf)
3rd party, Sharing, Wholesale, media companies,
MNO/MVNO CDN & SDNs.
NG-DC: Next-generation data center, NFV: Network Function Virtualization, SDN: Software Defined Network, CDN: Content Delivery Networks, MVNO: Mobile
Virtual Network Operator, MNO: Mobile Network Operator, PCRF: Policy & Charging Rules Function, P/S-GW: Packet/Serving Gateway, HSS: Home
Subscription Server, MME: Mobile Management Entity, IMS: IP Multimedia Subsystem, BSS: Business Support System, OSS: Operations Support Systems.
Network Slicing being very important!

“Softwared” Network …Cloud & Virtualization
Cisco VNI: 60+% of all IP Traffic is projected to be Metro-based and never floods the long-haul networks. Expectations are that this ration will not change much over
period to 2020. BELL LABS “METRO NETWORK TRAFFIC GROWTH: AN ARCHITECTURE IMPACT STUDY” indicates as much as 75% by 2017 is Metro-based.
DC: Data Center, QoE: Quality of Experience, ROI: Return on Investment, SDN: Software Definable Network, NVF: Network Virtual Function, EaaS: Everything as a
Service, CDN: Content Delivery Network, MSAN: Multiple-Service Access Node, v-ABC: virtualized-ABC, RAN: Radio Access Network, BBU: Base Band Unit. Gary Lee
“Cloud Networking “, 2014. Matt Portnoy “Virtualization Essentials:, 2012.
Metro
Transport
NG Core
Metro
Access &
Aggregation
Localized
Data Centers
(optional)
Metro
Data Centers
(optional)
(Inter)National
Data Center(s)
SDN
NVF v-CDN
v-GW v-VAS
v-RAN-BBU
Private-Cloud
Public-Cloud
EaaS, vMSS, vGW,
vBSS, …
Caching, storage, …
Public-Cloud
Infrastructure-
Cloud
End-User QoE, ROI & Transport
Cost Optimization required
Increased Operational
Efficiency
Improve Cost per Byte
DC Localization push
Due to huge growth in video demand
v-MSAN
Caching (e.g., video), local storage
Driven by growth in IP Video Streaming DemandQoE
$$$
$$$
$$$
QoE
$$$

Connectivity
100 Gbps
DWDM
10
GbE
Illustration
Transport … Terastream … The Endgame?
OLT
MSAN
Cellular
Wholesale
NG-CN
Data Center(s)
w. Cloud, XaaS, SDN/NFV, CDN, …
Note: Croatia Telekom (Deutsche Telekom), Croatia’s largest telecommunications company, first one to deploy the new TeraStream cloud-enabled transport & data
center architecture (in partnership with Cisco). MSAN: Multi-Service Acccess Node, OLT: Optical Lite Termination. Sources: Peter Lothberg (DTAG)
https://ripe67.ripe.net/presentations/131-ripe2-2.pdf & http://www.internetsociety.org/deploy360/blog/2014/01/videoslides-case-study-of-terastreams-ipv6-
implementation-ripe67/ , NG-CN: Next-Generation Core Network, MSAN: Multi-Service Access Node, OLT: Optical Line Termination, XaaS: Anything as a Service, SDN:
Software Definaed Network, NFV: Network Function Virtualization.
Small cell
Node
Small cell
Small cell
Small cell
Intelligence & Services
Moves to Cloud & End-User Devices
Target Min 100+ Mbps at consumer edge
R1
R2
Pure Switching
& IP Routing

Mobile Emerging markets
Growth slowing down.
Revenue 2009 – 2012 + 48% pa.
Revenue 2012 – 2017 9% pa.
Opex 2012 – 2017 + 12% pa.
& 3G has to get started 
- Top-line pressure (voice & sms).
- Increasing Opex pressure.
APAC Example
Long-term development troublesome.
Emerging Market Growth on expense of profitability
 Long-term development troublesome.
134%
88% 111
%
76%
68%
112%
55%
60%
60%
10%
20%
25%
Mob
%
3G%

Network economics.
Cost distributions illustrations - Capex.
Indication of heavy
fiber investments
Indication of leased
infrastructure
Might mean less
MW focus
Typically between
15% – 20%
Typically between
10% - 15%
Reduction of Capex
benefits cash
position of company!
(RAN)

Network economics.
Cost distribution illustrations – Opex.
Indication diesel
fuel dependency
Indication of leased
infrastructure
Indication of owned transport
– high MW reliance.
Typically below 10%
Typical range
5% - 15%
Typical range
15% - 25%
Reduction of Opex
benefits Profitability
of Company!

100%50%0%
Sites
0%
20%
40%
60%
80%
100%
CumulatedRevenue(Traffic)
The ugly tail … the low profitability areas.
Should drive sharing in low-traffic areas
50% revenue ≈ 10% sites
Low profitability
sites
Top 30% sites ≈ 80% revenue.
50% sites takes
less than 10% revenue
Illustration
Rural-like areasSub-urbanUrban

Why sharing a part or all of a 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”).

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 214

Expectation management.
The full sharing potential.
Total Opex
100%
Technology Opex
NT 14%
RAN
saving
RAN 7%
Cluster Opex
40%
Illustration
Expect up-to 35% saving on Tech Opex
Up-to 5% on Total Corporate Opex
Termination cost 1.5 – 3+ × of Opex savings
Integration Capex synergetic with BaU Capex
Instant Cell split potential  Enhanced Capacity
Spectral efficiency gains (>10%+)

Frequency
(MHz)
Site
(acq. + build)
Radio
(electronics)
Backhaul
(transport)
Backbone
(transport)
Core
(switch & control)
BSS
(bill & care)
Capex
prevention
Efficiency
enabler
40%-60% < 35% up-to 50% up-to 50%
Partly
possible
Less likely
Opex
prevention
Efficiency
enabler
< 35% ca. 35%
scale
discount
scale
discount
Partly
possible
Less likely
Regulatory
complexity
HIGH LOW LOWER LOWER LOWER HIGH HIGH
Network Sharing can provide better
economics and market timing …
BTS &
NODE-B
eNodeB
BSS
BSS
MNO 1
Core
MNO 2
Core
plmn 1
plmn 2
plmn 1 + plmn 2
(optional)
217Dr. Kim Kyllesbech Larsen, Network Sharing Fundamentals.

Anatomy of access 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.

Anatomy of modern network sharing options.
Aggressive sharing and still guaranties competitive differentiation,
operator independency and vast consumer quality improvements.
• Sharing: Network including data centers infrastructure will be shared.
• Not shared: Core Services & Functionality that provides respective operator’s customers with
differentiated brand, services, applications, handsets, rate plans, etc.
• Result: An ultra-efficient network providing an elastic range of services with high profitability than
an unshared network would be able to do. Increased capacity and efficiency.
SERVICES incl. NFV / SDN
DATA CENTER(s) CLOUD INFRA (min. IaaS) – COULD SHARE (SOME) NFV & SDN FUNCTIONS
SHARED ACCESS INFRASTRUCTURE incl. FRONTEND DATA CENTERS & TRANSPORT
BACKBONE FIBER & DWDM
SERVICES incl. NFV / SDN
SHARED FREQ.
Services & Functionality is Software
residing in COTS DC hardware
Individual Services & Functionality is Software
residing in shared COTS DC hardware

Most of a network can be shared.
Radio Access Network Part.
Concrete block to
stabilize tower
Soil
MW
Antenna
Cabinet
Concrete block to
stabilize tower
Soil
35 m
MW
Antenna (3×sectors)
Cabinet
DWDM
DWDM
Fiber Connectivity
Diesel
Electricity
Steel Tower
Concrete
Site Location
Site
Location
Radio Access Site
Dark
Fiber
~ 50% – 70% of Capex
~ 50% – 80% of Tech Opex
(unshared situation)
25% - 30%+ Opex
pending diesel
share of energy
30% - 40%+ Opex
pending lease model.
Up-to 50%+ of Total Capex
pending lease or no lease model.
At least 30%+ of
Total Capex
Equipment
10% – 15% of Opex goes to
Operations & Maintenance
At least 10%+ of
Total Capex but
Costly on per site
basis

Rollout Phase Steady State
Stages of sharing benefits.
Best sharing strategy depends on business cycle & technology age.
 High Capex prevention.
 Opex prevention.
 Cash optimized startup.
 Best network.
 Little Capex benefits.
 Opex savings.
 Significant write-off.
 High re-structuring cost.
 Extended coverage.
 High Capex prevention.
 Opex savings.
 Minor write-off.
 Re-structuring cost.
 Instant cell split (more capacity).
 Better network.
< 5 years 5+ years
UMTS - GSMLTE
> 5+ years
UMTS
Passive sharing: e.g., Site Lease & Civil Works, Antennas,..
Mast/Tower sharing: e.g., Ancillary & Rack/Cabinet, shelter sharing.
Transport sharing: e.g., Backhaul & Backbone – particular fiber-based and Microwave radios.
Active sharing: e.g., Frequencies, TRXs, PAs, Baseband, CPU, ports, Cloud Servers, ….
Modernization
GSM – UMTS
(LTE piggybacking)
Illustration

Restructure cost
can be significant.
Although contract
termination can be
less costly due to
longer operational
period.
Termination
•Site lease.
•Site restoration.
•Service
Contracts.
•Personnel cost
Other
•JV overhead
•Legal, etc..
Restructuring cost
can be low if little
legacy
infrastructure is
present.
If decision for
network sharing is
taken in the
renewal /
obsolescence
phase write-off
exposure can be
relative light both
for equipment and
site-build.
As most of the
network has been
deployed at this
stage the write-off
exposure can be
significant even if
equipment can be
re-used.
Relative low
exposure if little
legacy
infrastructure is
present.
Economics of sharing benefits.
Rollout Phase
Bulk (>80%) of sites and
nodes to be deployed.
Steady State
80% of coverage and sites
deployed. Mainly capacity
additions and coverage
maintenance.
Modernization/
Obsolescence
Active element / node
replacement, technology
migration. Site
consolidation.
Passive sharing
• Site build
• Mast
• Rack / Ancillary
Active sharing
• MW/Fiber
• Electronics
• Spectrum
• Resources
Passive sharing
Low Capex level
Active sharing
Passive sharing
Medium Capex level
Active sharing
Substantial Capex
Capex Synergy
= Low = High
Synergy potential
Opex prevention
• Site lease
• Non-telco services
• Telco services
• Energy
• Resources
Opex saving if
absolute number
of site locations
are reduced.
Primarily Opex
prevention in case
of site number
expansion.
Opex saving if
absolute number
of site locations
are reduced.
Primarily Opex
prevention in case
of site number
expansion.
OPEX SynergySharing stages Restructure Cost
= Low = HighCost exposure
Write-off
Illustration

Typical modelling scenarios to consider
for Network Sharing.
NETWORK SHARING EFFECTIVE SYNERGY %SHARED START STOP CAPEX #NOT SHARED Model
TOWER SHARING 40 METER 1 35% 35% 100% 2014 9999 0 0 TowerCo: Capex->Opex swap
ROOFTOP SHARING 0 0% 50% 0% 9999 9999 1 0 Own Asset
IN-BUILDING SYSTEMS (IBS) 0 0% 50% 0% 9999 9999 1 0 Own Asset
POWER SHARING
ANCILLARY SHARING - BBU 1 25% 25% 100% 2014 9999 0 1,000 TowerCo: Capex->Opex swap
ANCILLARY SHARING - GENERATOR 1 25% 25% 100% 2014 9999 0 1,000 TowerCo: Capex->Opex swap
MANAGED SERVICES POWER 1 0% 50% 0% 2014 9999 0 NA TowerCo: Capex->Opex swap
TRANSPORT SHARING
BACHAUL SHARING 0 0% 20% 0% 9999 9999 1 NA Own Asset
BACKBONE FIBER-OPTIC SHARING 0 0% 20% 0% 9999 9999 1 NA Own Asset
METRO FIBER-OPTIC SHARING 0 0% 20% 0% 9999 9999 1 NA Own Asset
ACTIVE SHARING
SPECTRUM SHARING 0 0% 0% 0% 9999 9999 1 NA Own Asset
RAN SHARING 0 0% 50% 0% 9999 9999 1 NA Own Asset
ANTENNA SHARING 0 0% 50% 0% 9999 9999 1 NA Own Asset
CABINET SHARING 0 0% 50% 0% 9999 9999 1 NA Own Asset
I&C SHARING 0 0% 50% 0% 9999 9999 1 NA Own Asset
RNC SHARING 0 0% 50% 0% 9999 9999 1 NA Own Asset
BACKBONE DWDM SHARING 0 0% 10% 0% 9999 9999 1 NA Own Asset
METRO DWDM SHARING 0 0% 10% 0% 9999 9999 1 NA Own Asset
SHARED SERVICE
FIELD MAINTENANCE 0 0% 50% 0% 9999 9999 1 NA Own Asset
FACILITY MANAGEMENT 0 0% 35% 0% 9999 9999 1 NA Own Asset
SPARE PART MANAGEMENT 0 0% 25% 0% 9999 9999 1 NA Own Asset
COMMON NOC 0 0% 35% 0% 9999 9999 1 NA Own Asset
TOWERCO TENANCY TARGET BY START PER YR
2-TENANT TARGET 1 50% 2014 2014 50% UNTIL TARGET IS REACHED
3-TENANT TARGET 0 0% 9999 9999 0% UNTIL TARGET IS REACHED

Case Study
Student Exercise on
Network Sharing.
Group Discussion

Starting point.
Planning assumptions:
• Need to built 100 sites along 200 km highway.
• Cost of site infrastructure xxx US$ (can support up-to 3 operators).
• Site rental is xxx US$ per year (this includes infrastructure maintenance = non-Telco)
• Energy cost is xxx US$ per year per site where of transport cost (for diesel) is xxx US$.
• Cost related to Telco operations per site is xxx US$ per year.
• Cost of electronics & antennas (incl. installation) xxx US$, antennas xxx US$.
• Cost of diesel generator for 1 operator xxx US$ and xxx US$ for 2 operators.
• Connect all sites to fiber.
• Microwave radios not feasible / available for backhaul connectivity.
• Average Site distance from main fiber path is 500 meters.
• Cost of fiber xxx per xx cores per km, you can add xx additional cores for xxx per km.
• Per 50 km you need optical switching equipment of xxx US$ (for up-to 2 operators).
• Maintenance cost of fiber xxx per year of total fiber investment.
• Maintenance cost of optical switching equipment xxx per year of total investment.
Group Study

1st task: Make a drawing that roughly depicts problem that needs to be solved!
2nd task: Split unit cost given in starting point between Capex (once off) and Opex
(recurring) cost elements.
3rd task: Calculate the standalone (not shared) total Capex and Opex an Operator
would have.
4th task: Assume you share all that can be shared calculate the total Capex and Opex
an operator would have.
5th task: Summarize your findings (Table or Chart or Textual etc..). How much can an
operator save by sharing? What would you recommend your board?
Assignment – prepare 5 slides.
Group Study

Basic Project Management
Basic Intro.

Project Management.
Project management is the discipline of
initiating, planning, executing, controlling, and closing
the work of a team to achieve specific goals and
meet specific success criteria.
(Wikipedia)

Project Management.
At least to parties in any project delivery …. You & Your Customer.
You Your Customer
• Cost okay compared to value.
• Delivery time-lines compared to internal
projects deliverables okay.
• Feeling of Trust (with 1-sided penalties).
• Contract signed.
• Cost to deliver within target.
• Confident that timelines can be met
(considering risks & penalties).
• Feeling of trust.
• Contract signed.
PM
CEO
CTO + Team
Sales CTO + Team
PM
Sourcing
Legal Legal
CEO
PM
Staff
Staff
Staff
Depending on organization setup PM reporting line can
be to different levels.
Staff
Staff
PM
Staff
Staff
Staff
Depending on organization setup PM reporting line can
be to different levels, but typically reporting to CTO.
Plan Dept.
Build Dept.
Operate Dept.
CTO
Dotted reporting lines
Functional reporting lines

Corporate You
Quest for
Perfection
Primal Desires
Rationality &
Moderation
e.g., “I want to get rich”,
“I want power”, “I want …”,
“me, me, me”, …
e.g., “Its is important that my
customer respects me”, “It is
important to have a long
term relationship with my
customer”, “I strive for the
highest quality”,
“getting buy-in from my
team ensures good delivery
and performance”.
e.g., The ego recognizes that other people have
needs and wants too, and that being selfish is not
always good for us in the long run.

You.
What is your product?
What is your Business Model
What’s your Operational Risks?
Strategy – today & tomorrow
Why you?
What is your (target) cost?
What capital needs do you have?
What’s your goals?
Do all of you know you?
What value do you offer?
Who are you?
What can you deliver?
and when?

The customer.
- Who is your customer?
- Do you have more than one?
- What about internal in your own company?
- What does your customer want?
- If you have more customers, do they want
the same thing?
- Do you have clear customer expectations?
- Do you manage customer expectations?
- How valuable is your product to your
customer(s)?
- Do you understand your competition and
their role with your customer?

Quiz
Write down 3 bullets points that you
think is the most important to consider in
a project.
Group Discussion

Organization & project management.
The functional organization.
Ex. Planning
Ex. Built
Ex. Engineering
The Classical organization :
Influence on project
execution;
• Project manager’s
authority: Little / None.
• % of Organization
assigned to full time
project work: ~ None.
• Project managers role:
usually part-time.
• Common titles: Project
Coordinator or Project
Leader.
• Project Admin Staff: If
existing usually on part-
time basis.
Source: http://www2.fiit.stuba.sk/~bielik/courses/msi-slov/reporty/pmbok.pdf
Almost Escalations
ends up here
Often lacks clear roles and responsibilities. Very little project
memory is attained from project to project

The project organization.
Project focused organization:
Ex. Planning
Ex. Built
Ex. Engineering
execution;
authority: High to
almost total.
project work: at least
80%.
usually full-time.
Manager or Program
Manager.
• Project Admin Staff:
Full time but synergetic
between project
managers.
Rare and only very
serious escalations
Full control of all
aspects of project
Yangon-NPT
West
Yangon-NPT
East
MDL -
Myawaddy
Clear roles and responsibilities. High degree of project
memory is attained from project to project.

The composite organization.
So-called Composite organization:
Ex. Planning
Ex. Built
Ex. Engineering
execution;
authority: Relative
High.
project work: at least
80%.
usually full-time.
Manager or Program
Manager.
• Project Admin Staff:
Full time but synergetic
between project
managers.
Internal escalations (can be frequent).
External escalations should be rare
Reasonable control of all
aspects of project
Usually clear roles and responsibilities. High degree of project
memory is attained from project to project. Combines functional
organization with a matrix organization.
Most internal & external
escalations ends up here!

Project Management Methodology.
Initiation Execute Closing ?
Are you getting smarter after each project?
Do you get more efficient after each project?
Are your customer getting more happy after each project?
Etc..

When does it happen?
• Project Initation tasks overlap with Planning tasks.
• Pass when the majority of the initiation task is done.

Design Phase.
Initiating
processes
Planning
processes
Controlling
processes
Executing
processes
Closing
processes
Implementation Phase.
Initiating
processes
Planning
processes
Controlling
processes
Executing
processes
Closing
processes
Prior
Phases
Next
Phases
Interaction ... rolling wave planning.
Its not rocket science but its not easy either!

How to recognize a poorly initiated
project?
PM cannot answer
straight questions
about the project
Project Team can not
agree on what it is
supposed to be doing
The stake-holders
are confused
There is no definable
contract between the
project & its
stakeholders
Initiating
processes

Why project initiation is important.
 It reduces wastage: your time, your
stake-holder’s time, project team’s
time, money
 It obliges your stake-holders to think
harder about they want before you
commit to deliver
 It gives your stake-holder confidence
that you are approaching the project
professionally
 It defines what you need to do to be
successful
INCREASED
PROBABILITY
OF
SUCCESS
Benefits of Initiation.
Initiating
processes

Getting the project team together.
Know your stakeholders and team members. Realize that on both
your own side and your customers you will find these types.
Source: Dr. Einecker Personal Management Training (2009).
Each member can be
characterized by the
behavioural type.
The type is also a clue as to
how to manage them.
Note: don’t call your colleagues by the
types that you might assign to them
Initiating
processes

Power Interest Attitude Type Technique
+ + +
Saviour
Do whatever is needed to keep them on
your side. Pander to their needs.
- + +
Friend
Use as a confidant or sounding-board.
They may be able to help with saboteurs.
+ + -
Saboteur
Engage in order to disarm them. Be
prepared to clear up their mess.
- + -
Irritant
Make sure they don’t waste valuable
resources. Be prepared to put them in
their place.
+ - + Sleeping
Giant
Engage in order to awaken.
- - + Acquaintance Keep informed.
+ - -
Time Bomb
Understand their concerns, and defuse
before the bomb goes off.
- - -
Trip Wire
Understand their concerns to avoid
tripping up.
Source: Einecker 2009 & Lucidus 2005.
Initiating
processes
Ways to manage personal complexities.
Techniques to consider to manage stakeholders & team members.

The Planning process.
Planning
processes
Scope
Planning
Scope
Definition
Activity
Definition
Resource
Planning
Activity
Sequencing
Activity
Duration
Estimation
Cost Estimation
Schedule
Development
Cost
Budgeting
Project Plan
Development
Core processes that should be considered.
Facilitating processes supporting the planning.
Quality
Planning
Organizational
Planning
Staff
Assignment
Procurement
Planning
Sourcing
Planning
Risk
Identification
Risk
Quantification
Risk Response
Development
Communication
Planning

The planning process.
Example: how to build a cellular tower structure?
Concrete block to
stabilize tower
Soil
MW
Antenna
Cabinet
Concrete block to
stabilize tower
Soil
Project 1 in Scope
Equipment
35 m
MW
Antenna (3×sectors)
Cabinet
DWDM
DWDM
Fiber Connectivity
Diesel
Electricity
Steel Tower
Concrete
Site Location
Site
Location
Endgame
Project
3
Project
2
Project
4

Approach to expert planning.
Minimum, Most Likely and Maximum judgement.
Source: the above expert curves have been generated with Palisade Decision Suite 7.0 (Trial Version)
see http://www.palisade.com/. For more on Expert Functions see also http://www.riskamp.com/beta-pert.
Question to the Expert:
Given an event what is the Most Likely outcome, what is the Minimum outcome and
what do you think the Maximum outcome is?
Example: How long does it take to prepare a acquired site for tower installation?
1. Building permit; Most Likely (ML) 21 days, never been less than 7 days (Min) and also never no longer
than 35 days (Max).
2. Site preparation (e.g., remove vegetation, shield building area, flatten out ground, etc…); ML 5 days, Min
2 days and Max 10 days.
3. Digging (e.g., for foundation); ML 15 days, Min 5 days and Max 50 days (in case of heavy rainfall).
4. Concrete foundation; ML 14 days, Min 5 days (dry weather, dry soil), Max 20 days (wet weather/wet soil).
+ + + =
1 2 3 4
From 30 to 95 days
with 59 on average
Group Study

The planning process.
Example: how to build a cellular tower structure?
Concrete block to
stabilize tower
Soil
Project 1 in Scope
Steel Tower
Concrete
Site
Location
Steel tower:
- Order Mon 4 Jan, 2015
- Shipping P(+14,+20,+25) d
- Transport P(+10,+15,+30) d
- Import P(+2, +5,+10) d
- Transport to warehouse +1 d
- Transport to site +1 d
- Tower Build +1 d
- Ready for Telecom Inst.
- Min 1 Feb 2015
- ML 5 Mar 2015
- Max 12 Apr 2015
Import permit Tower:
- Application Wed 6 Jan
- Process P(+5,+10,+15)
- Import Permit Granted Date
- Min 10 Jan 2015
- ML 16 Jan 2015
- Max 22 Jan 2015
Site acquisition:
- Start Mon 14 Dec. 2014
- Land Lease Contract P(+15,+20,+30) d
- Permitting P(+7,+21,+35) d
- (could allow for permit failure).
- Permit Granted Date
- Min 07 Jan 2015
- ML 24 Jan 2015
- Max 12 Feb 2015
Site preparation:
- Start Permit Granted Date.
- Site preparation P(+2,+5,+10)
- Digging P(+5,+15,+50) d
- Concrete P(+5,+14,+20) d
- (add seasonal dependency).
- Site ready for Tower Installation Date
- Min 30 Jan 2015
- ML 03 Mar2015
- Max 10 Apr 2015
1 2
3
Group Study

The Execution process.
Execution
processes
Project Plan
Execution
Facilitating processes supporting the Project Plan Execution.
Information
Distribution
Sourcing
Source
Selection
Team
Development
Contract
Administration
Quality
Assurance
Scope
Verification

The Controlling process.
Control
processes
Performance
Reporting
Facilitating processes supporting the Project Plan Execution.
Scope Change
Control
Risk Response
Control
Schedule
Control
Quality Control
Cost
Control
Overall
Change Control
Communication
Control

Closing.
Enjoying the laurels after a job well done?

Always (always!) do a Post-mortem ...
Even when successful and the “patient survived”
& yes honesty is important!

“A postmortem is a meeting held shortly after
the completion of every movie in which we
explore what did and didn’t work and attempt
to consolidate lessons learned.”
Companies, like individuals, do not become
exceptional by believing they are exceptional
but by understanding the ways in which they
aren’t exceptional. Postmortems are one route
into that understanding.”
Source: Catmull, Ed (2014-04-08). Creativity, Inc.: Overcoming the Unseen Forces That Stand in the Way of True Inspiration
“In truth, to most people postmortems
seem a bit like having to swallow some kind
of bad-tasting medicine. They know it’s
necessary, but they don’t like it one bit.”
Consolidate what’s been learned.
Teach others who weren’t there.
Don’t let resentments fester.
Use the schedule to force reflection.
Pay it forward … ask the right questions going forward

Quiz
Write down the first 3 things you think
about when you hear the word mistakes.
Group Discussion

On mistakes.
Mistakes aren’t a necessary evil. They aren’t
evil at all. They are an inevitable consequence
of doing something new (and, as such, should
be seen as valuable; without them, we’d have
no originality). And yet, even as I say that
embracing failure is an important part of
learning, I also acknowledge that
acknowledging this truth is not enough. That’s
because failure is painful, and our feelings
about this pain tend to screw up our
understanding of its worth. To disentangle the
good and the bad parts of failure, we have to
recognize both the reality of the pain and the
benefit of the resulting growth.
Source: Catmull, Ed (2014-04-08). Creativity, Inc.: Overcoming the Unseen Forces That Stand in the Way of True Inspiration.

7 key criteria to project success…..
Objective Related Activities
Stakeholders Stakeholders are committed. Identifying, evaluating, informing, and influencing individuals
and groups who are affected by or who influence the project.
Business
Benefits Business benefits are realised.
Estimating, measuring, and monitoring the benefits the client
will gain from the project. Benefits are understood as
including the expected result and the associated costs, both
financial and non financial, of achieving that result.
Work, Schedule &
Quality
Work, schedule & quality
should be predictable.
Controlling the production and acceptance of project services
and deliverables, and ensuring their provision to specified
performance and acceptance requirements.
Team Team is high performing.
Identifying, mobilizing, and developing the people required
for the project team. Obtaining and maintaining appropriate
space, equipment, and other resources required to
successfully complete the project.
Scope Scope is realistic & managed.
Agreeing, maintaining, and amending the boundaries of the
project.
Supplier
Performance
Supplier is performing to
(contractual) expectation.
Establishing, agreeing, and monitoring the supplier
performance (such as financial reward, knowledge transfer,
and skill development). Maintaining and protecting the
supplier’s interests in relation to the project.
Risks Risks are mitigated.
Identifying and evaluating risks and issues. Developing
avoidance, mitigation, and resolution activities to counteract
those risks and issues.

Modern Telecommunication ... The Very Basic.

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Modern Telecommunication ... The Very Basic. (20)

More from Dr. Kim (Kyllesbech Larsen) (13)

Recently uploaded (20)

Modern Telecommunication ... The Very Basic.