The Internet Telecom relationship

1
Telecoms and Internet Infrastructure Challenges
for Asia
Ministry of Communication Ministry of Planning
Royal Government of Bhutan
Presentation by Bertrand Bidaud
24 February 1999
2
The key challenges

• Structural
• Financial
• Technology

3
The Internet / Telecom relationship
Stage 1 providing the service
Stage 3 New market structure
Stage 2 Development
Demand gt Supply
Demand Supply
Demand Supply
Little competitive pressure the golden age
Market segmentation
Market gets very sophisticated
4
Phase 1 - Providing the service
US
Dial up
Routers, switches
Japan
Leased line
Other domestic ISPs
Frame Relay,...
ISP network (on telco lines)
Other ISPs networks
Telco access network
User Premises
5
Phase 2 - Development
Focus of Infrastructure (Internet Exchange,)
IX
Focus of Value Added Services (Certification
authority, web hosting, etc)
6
Phase 3 - New Market structure
Provide all telecom services (Voice over IP,..)
IX
Dark fiber Global Crossing, Hermes,...
IX
7
The incestuous relationship

• Telcos compete with their customers
• Good Telcos have deep pockets
• Bad Cross-subsidies mean competition does not
  perform its role
• ISPs expand their activity and eventually
  compete with their initial supplier
• UU-net (MCI-Worldcom)
• IIJ (Japan)
• Coopetition
• ISP Telco

8
The key challenges

• Structural
• Financial
• Technology

9
The ISPs cost structure
US
Rest of the world
International network
10
Why does the cost structure matters to policy
makers?
Catch 22

• For IT industry to develop, its needs low
  operating cost, AND quality network
• If added operating cost is transferred to the
  customer, IT industry at disadvantage
• If not, ISPs cannot invest in good quality
  network and services

11
The importance of low rates
New Revenue
Rates go down
Revenue or Cost
Revenue
New Profit
Cost
Profit
Volume
Demand
New Demand
12
International Internet Financing Things are
getting worse

• Bandwidth per user is increasing
• multimedia application
• lines are filled to the limit
• access to user is getting faster
• from modem dial up to ADSL/Cable modem
  (several Mbps per user)
• Currency devaluation
• International lines are (usually) paid in USD

13
International Internet FinancingHow did we get
there?
10
USA
Asia Pacific
90
Traffic Imbalances
14
International Internet FinancingHow did we get
there?
20
USA
Asia Pacific
80
Traffic Imbalances (end of 1997)
15
International Internet FinancingHow did we get
there?
N. American Share of Internet Users
100 90 80 70 60 50
Feb-97 Nov-97 Sep-98
Traffic Balanced soon
16
International Internet FinancingBig money?
Japan
gt500Mbps
Korea
gt200Mbps
USA
Others Asia
Asia Pacific
gt300Mbps
Estimated capacity - End of 1997
Estimated cost 150 millions The subsidy
from Asian users to N. American users 30
millions
17
International Internet FinancingBig money?
Japan
gt500Mbps
Korea
gt200Mbps
USA
Others Asia
Asia Pacific
gt300Mbps
Year 2000
Estimated cost 1 Billion The subsidy from
Asian users to N. American users 200 millions
18
International Internet FinancingBig money?
According to Telstra estimates, global subsidies
will amount to 5 Billions per annum And
growing!!
(FCC claims that US subsidies resulting from
outdated IDD settlement systems amount to 3.7
Billions)
19
International Internet FinancingPeering
50
USA
Asia Pacific
50
A false-good-idea
20
International Internet FinancingPeering
It wont happen
Asia Pacific
A false-good-idea
21
International Internet FinancingNew settlement
system
The user pays for the usage
20
USA
Asia Pacific
80
Volume based rate as in Singapore ADSL? Could
this be acceptable to N. American users?
22
International Internet FinancingAsian backbone?
Qwest backbone in the US
23
International Internet FinancingAsian backbone?
Ten 34 backbone in Europe
24
International Internet FinancingAsian backbone?
EUnet backbone in Europe
25
International Internet FinancingAsian backbone?
A-Bone IIJ as the Hub
26
International Internet FinancingAsian backbone?
STIX Singnet as the Hub
27
International Internet FinancingAsian backbone?
Many Hubs, - STIX - IIJ A-Bone - HKTelecom Net
Plus - Australia Big Pond - Internet KDD No
backbone
28
International Internet FinancingAsian backbone?
Submarine links being developped1
Asia Pacific Cable Network 2 KDD, Singtel, HK
Telecom, China Telecom Asia Pacific Cable
Network II NTT, MCI Worldcom. ATT,
Sprint Expected to reduce by 2/3 the cost of
transpacific links As well as Oxygen, Global
Crossing,...
The missing link? Terrestrial !
29
International Internet FinancingTransparency
For competition to lead to better price, to
better service information must be
available In the world of International
Internet connectivity, information is scarce
(confidentiality of agreements) Risk of
cartel type of behavior
30
International Internet FinancingActions
Telstra has been the most vocal (and earliest)
proponent of change (since 97) Several telcos
have issued a joint statement (Singtel, KDD,
Communications Authority of Thailand, Taiwan's
ChungHwa Telecom Co., Indonesia's PT Indosat,
Korea Telecom, Philippines Long Distance
Telephone Co. and Telekom Malaysia Bhd) ask the
US to share the cost of Internet
Infrastructure APEC Tel will look into the
issue at institutional level (Miyazaki-March
1999) Asia Pacific Internet Association studies
options on behalf of its members (Internet
professionals and ISP)
31
The key challenges

• Structural
• Financial
• Technology

32
Intelligent? Stupid?
1997 David Isenberg- The rise of the stupid
network The death of the 5 lanes theory
(mid-1997) - TV, Tel, CATV, Cellular,
Satellite IP ubiquity permits to rethink the way
networks are designed Key argument The cost!
33
Intelligent? Stupid?
IP
IP
IP
ATM
FR
ATM
SONET/SDH
OCS-WDM
OCS-WDM
IP over ATM over Sonet/SDH over OCS- WDM IP over
FR over Sonet/SDH over OCS- WDM IP over
Sonet/SDH over OCS- WDM IP over ATM over OCS- WDM
Do you really need that? What about IP over
WDM?
34
The IP paradigm

• Traditionally, networks are designed to optimize

Bandwidth
US Telco Fiber capacity (Source Gilder
Technology Report)
Processing power
Transistors per Intel Processor (Source Gilder
Technology Report)
35
The IP paradigm

• Why not considering that
• Bandwidth (at the backbone) is not the limit
• End user processors are powerful enough to handle
  any necessary real time conversion (for voice,
  video, etc)

Simple processing at the backbone level - no
quality control, or sophisticated channel
reservation Simple switching IP over WDM
36
The IP paradigm
Key success factor NO CONGESTION AT THE BACKBONE
Advantage Operating cost Versatility
37
Any development?

• US
• QWEST, Level3, ION (Sprint), ICN (ATT),...
• Europe
• Hermes, QWESTKPN, ...
• Asia
• Crosswave Communications Inc. (IIJ, Sony, Toyota
  ), Prism (Japan Telecom),...

But still IP over ATM, and IP over SDH
38
The Asian challenge

• Big numbers required to reach break even
  pointBig pipes, high number of heavy users,...

Estimated number of hosts connected to Internet
com 10,301,570 Commercial net 7,054,863
Networks edu 4,464,216 Educational mil
1,359,153 US Military jp 1,352,200 Japan au
750,327 Australia kr 174,800 Korea,
Republic Of tw 103,661 Taiwan, Province Of
China hk 72,232 Hong Kong sg 59,469
Singapore my 40,758 Malaysia th
25,459 Thailand id 10,691 Indonesia ph
7,602 Philippines
(July 98-http//www.nw.com/zone/WWW/dist-bynum.htm
l)
39
The key challenges

• Structural
• Financial
• Technology

Thank you