Upload
ictseserv
View
595
Download
2
Tags:
Embed Size (px)
Citation preview
© 2011 The SESERV Consortium 1
A SESERV methodology for tussle analysis in Future Internet technologies
Introduction and motivation
Costas Kalogiros, Costas Courcoubetis, George D. Stamoulis{ckalog, courcou, gstamoul}@aueb.grAthens University of Economics and Business
September 2011
Socio-Economic Services for European Research Projects (SESERV)
European Seventh Framework CSA FP7-2010-ICT-258138
© 2011 The SESERV Consortium 2
Internet as a platform for stakeholders’ interactions
Internet Technology layer
InternetSocio-Economic layer
ISPsEnd-users ASPsRegulators
Socio-Economic layer is governed by laws of socio-economics, while technology layer by laws of physics
routerslinks switches
Internet protocols
Internet applications
Firewalls
middleboxes
3G towers
Real-world socio-economic transactions
Stakeholders with varying socio-economic interests
Technology choices(including investments, configurations)
Technology outputs (connectivity, QoS, mobility, security, etc.)
Technology components
…
servers
© 2011 The SESERV Consortium 3
Basic socio-economic technology cycle
Stakeholders’ strategies / policies with respect to a specific technology (functionality)
Adopt technology
Dimension resources
Configure technology
Use technology
ISP
Longer
Shorter
Adap
tatio
n tim
esca
le
At each stage conflicting incentives may exist at the socio-economic layer. The combination of actors’ strategies lead to a tussle outcome, which is characterized by the benefit that stakeholders get.
Internet Socio-Economic layer
tussle outcome
© 2011 The SESERV Consortium 4
Tussle evolution (1/2)
• If the tussle outcome is considered “unfair” by a subset of stakeholders, they can react by:• Leaving the system• Adopting another technology/ reconfiguring the selected one• Asking the regulator to intervene by restricting other
stakeholders’ policies• … hence making the outcome unstable
• Even though a tussle outcome can be considered “fair” by all stakeholders of that particular functionality, it can destabilize other functionalities (spillover effect to other tussles)
Analyzing the anticipated tussles can shorten unstable periods & help the long-term success of a technology
© 2011 The SESERV Consortium 5
Unstable outcome
Stable outcome Evolves
Affects
Functionality A
Functionality C
Functionality B
…
Legend
Tussle A2
Tussle C2
Tussle B1 Tussle B2
Initial state
Functionality
…
Functionality A
Functionality C
Functionality A
Functionality C
Functionality B
Tussle A1
Tussle C1
Functionality B
T0 (initial tussle outcomes)
T1 T2
Transfers
TA Time A
…
Tussle evolution (2/2)
© 2011 The SESERV Consortium 6
Purpose of Tussle Analysis• Defines a systematic approach for understanding the impact of
introducing new Internet technologies• Why a new technology is needed today?• What are the interests of existing stakeholders today?• What options do existing technologies offer to stakeholders?• What are the properties of existing outcome in terms of performance
& stability? • What would be the effect of a new technology to the ecosystem in
the future? • How would the interests of existing and new stakeholders be
affected?• How would the options of existing and new stakeholders be affected? • Can this technology help reaching a “fairer” outcome regarding this
functionality, or increase efficiency in case of an already stable outcome?
© 2011 The SESERV Consortium 7
Tussle analysis case study: bandwidth sharing
Functionality: bandwidth sharing
ISPs throttle bandwidth of p2p applications by using DPI technology.
p2p applicationsconfigured to perform traffic obfuscation
What if ISPs deploy congestion exposure technologies & congestion pricing schemes?
Functionality: VoIP service delivery
Regulator announces fines due to VoIP providers’ complaints
p2p users get disproportionate bandwidth share
p2p users configureapplications to open multiple TCP connections for thesame session
?
ISPs use DPI technology to degrade quality of rival VoIP services
ISP’s telephony services get preferential treatment
?
no discrimination
peer-to-peer (p2p) users
interactive users
interactive users get
disproportionate bandwidth share
ISPs
ASPs
Regulator
ISP(neutral entity)
fairbandwidth
sharing
Tuss
le o
utco
me
Stak
ehol
ders
’ st
rate
gies
/pol
icie
sSt
akeh
olde
rs’
stra
tegi
es/p
olic
ies
Tuss
le o
utco
me
ASP’s VoIP services get preferential treatment
p2p applications motivate multiple TCP connections for the same session
© 2011 The SESERV Consortium 8
Step 1: Identify all primary stakeholder roles and their characteristics for the functionality under investigation
Step 3: For each tussle assess the impact to each stakeholder and potential spillovers
Functionality I Functionality II
Step 2: Identify tussles among identified stakeholders
spillover new iteration
tussle tussle tussle tussle
A high-level view of the SESERV tussle analysis methodology
© 2011 The SESERV Consortium 9
A guide to applying the SESERV tussle analysis methodology
© 2011 The SESERV Consortium 10
More Information
Don’t hesitate to contact us if your research project is interested in performing a tussle analysis for understanding the existing (and/or future) socio-economic issues related to your proposed Future Internet technologies.
• http://www.seserv.org
• http://www.linkedin.com/groups?about=&gid=3870856
• http://www.twitter.com/seserv
© 2011 The SESERV Consortium 11
APPENDIX
© 2011 The SESERV Consortium 12
Design For Tussle
• A tussle-aware Internet protocol should• lead to a stable outcome by allowing all involved
stakeholders to express their interests and affect the outcome (“Design for Choice” Principle)• It does not impose one particular outcome because no one
can guess it
• avoid spillovers to other functionalities (“Modularize along the tussle boundaries” Principle)
Clark, D. D., Wroclawski, J., Sollins, K. R., and Braden, R.: Tussle in Cyberspace: Defining Tomorrow’s Internet. IEEE/ ACM Trans. Networking 13, 3, pp. 462-475, June 2005
© 2011 The SESERV Consortium 13
Towards achieving stable outcomes
• The “Design for choice” principle provides guidance in designing protocols that allow for variation in outcome. Useful properties are: • “Exposure of list of choices” suggesting that the stakeholders
involved must be given the opportunity to express multiple alternative choices and which the other party should also consider.
• “Exchange of valuation” suggesting that the stakeholders involved should communicate their preferences in regard to the available set of choices (for instance by ranking them in descending order).
• “Exposure of choice’s impact” suggesting that the stakeholders involved should appreciate what the effects of their choices are on others
• “Visibility of choices made” suggesting that both the agent and the principal of an action must allow the inference of which of the available choices has been selected. Clark, D. D., Wroclawski, J., Sollins, K. R., and Braden, R.: Tussle in Cyberspace: Defining Tomorrow’s Internet. IEEE/
ACM Trans. Networking 13, 3, pp. 462-475, June 2005
© 2011 The SESERV Consortium 14
Towards avoiding tussle spillovers to other functionalities
• The “Modularize the design along tussle boundaries” principle helps in identifying whether tussle spillovers can appear.
• A protocol designer can check any of the following two conditions:• “Stakeholder separation”, or whether the choices of one
stakeholder group have significant side effects on stakeholders of another functionality (another tussle space), for example creates economic externalities between stakeholders of different tussle spaces.
• “Functional separation”, or whether different stakeholders use some functionality of the given technology in an unforeseen way to achieve a different goal in some other tussle space, i.e., the functionality of technology A interferes (and possibly cancels) with functionality of technology B.
Clark, D. D., Wroclawski, J., Sollins, K. R., and Braden, R.: Tussle in Cyberspace: Defining Tomorrow’s Internet. IEEE/ ACM Trans. Networking 13, 3, pp. 462-475, June 2005