Transcript Internet interconnection - Pompeu Fabra University

Manuel Palacin ([email protected])
INTERNET INTERCONNECTION
& NET NEUTRALITY
12th November 2010
Content
 Net Neutrality



What is Net Neutrality?
BEREC public consultation
Search Neutrality
 Network Interconnection









Definition
Types
Alternative Interconnections models
BW Stock Market
Scenarios
Operators Action Plan
Objective
What is new in this approach
Open questions
What is Network Neutrality?
 Network Neutrality is a principle that is based on
the network access providers can not prioritize
access of contents to users or companies over
others
 The net neutrality does not prohibit the
prioritization of flows according to the needs of
users, but attempts to prevent access providers
discriminate these flows based on content,
origin or destination of traffic

i.e. If a network access provider reaches a financial agreement with Yahoo,
it can prioritize this search over Google search. User access to Google
website is not banned, but it will be so slower that the user could end up
giving up and switching to another browser.
3
BEREC Position
 BEREC (Body of European Regulators for
Electronic Communications)
 BEREC responses (30th Sept 2010) 14 questions
http://www.erg.eu.int/doc/berec/bor_10_42.pdf
 KEY problems that the resolution identifies if
operators separate from neutrality principles:
 Discrimination and anticompetitive effects
 Long-term consequences in the Internet economy
 Problems in transparency and QoS
4
Other Point:
Search Neutrality
 Net Neutrality is in the air, but what happens
with Searchers?
 Are neutral doing their searches or discriminate
depending on their interests?
 Google was accused of modifying its Search
Engine and GoogleAddWords to penalize some
sites
5
Network Interconnection
 Definition:
 physical connection between a carrier's facilities
and the equipment belonging to its customer
 connection between two (or more) carriers
6
Interconnection Types
 Peering - Bill And Keep (BAK)
 Bilateral agreement signed by two
operators or ISPs to exchange traffic for
their customers in pure reciprocity
 Transit:
 An operator agrees to carry traffic from
another operator to a third operator or
interconnection point
 Calling Party Network Pays
 Calling operator (caller) pays the total
cost of point-to-point connection.
7
Exists any Interconnection
Alternative?
 Maybe an interconnection stock market?
 Operators offer ALL or EXCEEDED BW from their




networks with an associated QoS (LSA)
Operators offer Dynamic BW Fares (LSA) in real time
for LONG or SHORT periods (hours, days, exact date)
The prices depend on the demand like in the stock
market  DYNAMIC
Operators subscribe to LSAs with better conditions 
tradeoff(LessPrice-BestQoS) in real time
Requires a new model of acquisition of BW more flexible
 REAL TIME
8
BW Stock Market





Based on supply and demand model using Revenue Management
(Airlines Business Model)
Band Width stocks are from a limited period of time (hours, days, ...)
using LSAs
Can be negotiated QoS (more complex than Best Effort)
Maximize network usage to obtain better operating performance.
Objective: to occupy 100% of resources
Can appear a new currency based in BW units instead of money. A
“Xbps” unit TODAY does not have the same value as YERSTERDAY or
TOMORROW
BW STOCK MARKET
SELLER
OPERATORS
LSAs:
LSAs:
-1 Gbps BE 20-01 h: 500€
-2 Gbps AF1 1/1/2011:
1000€
-1 Gbps BE 20-01 h: 400€
-2 Gbps AF1 1/1/2011:
1200€
B
W
B
W
BUYER
OPERATORS
B
W
B
W
B
W
B
W
B
W
B
W
B
W
B
W
B
W
B
W
LSAs:
-1 Gbps BE 20-01 h: 600€
-2 Gbps AF1 1/1/2011: 800€
B
W
B
W
B
W
B
W
B
W
B
W
TIME
9
Scenarios (I)

Mobile Market  2 proposals

Mobile Virtual Network Operator (MVNO)
 Buy FIXED BW
 Buy VARIABLE BW On Demand

Operators
 When an Operator is saturated or in NO coverage areas
 Buy VARIABLE BW On Demand

Possible consequences:



MVNO are not restricted to only 1 operator  competence in prices
MVNO can buy BW depending on their necessities  granularity
Operators can buy BW temporarily in saturated or NO coverage areas
 Can detect new HOT POINTS to invest with new equipments

Operators can find a new Business Opportunity selling BW to Third Parties
BW STOCK MARKET
MOVISTAR
75% 25%
ORANGE
50%
50%
VODAFONE
80% 20%
Why don’t we
use this extra BW?
10
Scenarios (II)

Wired Market  Many points of view

Access Networks
 Big operators
 They are connected to their Core Networks
 Small operators
 They are connected to their Core Networks
 They could need extra BW
 Community operators
 They can have rigid agreements with one Operator
 They can buy BW on demand of different Operators
ACCESS
NETWORK
FTTH Town
CORE
NETWORK
11
Scenarios (III)

Wired Market  Many points of view

Core Networks
 Big operators
 Peering between operators of same traffic volume
 Offer BW stocks (instead of money) of their network to small operators when passes
through their “small” networks
 Small operators
 Peering between operators of same traffic volume
 Buy BW slots of big network when passes through a bigger networks
 Use BW stocks to pay from previous agreements
 International carriers
 They can offer variable BW prices but their prices maybe are NOT competitive because
they don’t have competence
BIG
1
SMALL
1
BIG
2
SMALL
2
CARRIER
12
SMALL Operator Example
BUY EXTRA BW from:
BW Stock Market
SMALL
1
BIG
1
BIG
2
PEERING
SMALL
2
CARRIER
13
Operators action plan
 Peering  between operators of same size

Used by Operators that passes more or less the same traffic amount
 Transit  between operators of different size


Fixed and rigid agreements
Easy to develop
 BW Stock Market


Dynamic agreements
Difficult to develop
 Transit vs. BW Stock Market

Markets where prices are fixed/planned/regulated are less elastic than
dynamic ones based on supply/demand with competence
 BW Stock Market will be better than transit only without mono
 Mixing Peering with BW Stock Market

Use peering when is possible and BW Stock Market when not or to
obtain extra BW
14
Objective
SOCIAL
Competence
More Employments
Reduce Prices
Evolution of
Information Society
Demand Increases
Resources ~100%
ECONOMICAL
Expand Market
Operator buys
new resources
(equipments,
infrastructure,
interconnections)
Telcos Providers,
Manufactures, ...
Create Market
New Possibilities  New Services
15
What’s different/new in this
approach?
 Network Brokers Studies are basically negotiators of
QoS in end-to-end connections that choose between
fixed prices
 BW stock market proposes not how to negotiate (because
is the same ideatradeoff price/QoS), it aims for a
dynamic BW market model
 Operators are NOT putting BW prices in a dynamic
and real-time way
 Operators sellers could offer BW-slots for a
determined time-interval varying the price
 Operators buyers could buy in advance because they use a
learning algorithm based on historical statistics
 We can test if Revenue/Yield Management methodologies
can fit in the BW business environments
16
Open questions
 Is this model more effective in terms of competition than
Peering or transit? In which situations (small/big
operators)?
 Maybe can be used as a complement to these models for
selling exceeded Band Width?
 It is needed more than 2-3 interconnections points to avoid
monopolies or oligopolies?
 Is it needed COMPETENCE to make run this model?





It is a sustainable model in terms of maintenance?
Will small operators, that cannot grow up, die?
Will BW prices be reduced?
Is this model creating more demand?
Can interconnection agreements be done using an
hypothetic BW currency?
17
Bibliography










DYNAMIC EFFECTS OF NETWORK NEUTRALITY. Johannes M. Bauer. Michigan
State University. TPRC 2006
Draft Work Programme 2011 BEREC Board of Regulators.
BEREC Response to the European Commission’s consultation on the open
Internet and net neutrality in Europe. BEREC
The Evolution of Internet Congestion. Steven Bauer, David Clark, William Lehr.
Massachusetts Institute of Technology
Internet Peering. Jean-Jacques Laffont, Scott Marcus, Patrick Rey, Jean
TiroleSource. The American Economic Review, Vol. 91, No. 2
Network Externalities, Competition, and Compatibility. Michael L. Katz; Carl
Shapiro. The American Economic Review, Vol. 75, No. 3
Two Papers on Internet Connectivity and Quality. Roberto Roson. Dipartimento di
Scienze Economiche, Università Ca’Foscari di Venezia
Pricing and brokering services over interconnected IP networks. D. Di Sorte, G.
Reali.Dipartimento di Ingegneria Elettronica e dell’Informazione (DIEI), University of
Perugia
Dynamic Price Competition with Fixed Capacities. Victor Martinez-de-Albeniz,
Kalyan Talluriy. UPF and IESE
Pricing, Provisioning and Peering: Dynamic Markets for Differentiated Internet
Services and Implications for Network Interconnections. Nemo Semret, Raymond
R.-F. Liao, Andrew T. Campbell, and Aurel A. Lazar, Fellow, IEEE
18
Thank you!
Any question?