Network Economics -- Lecture 2: Incentives in P2P systems and reputation systems Patrick Loiseau EURECOM Fall 2012.

Slides:



Advertisements
Similar presentations
Network Economics -- Lecture 1: Pricing of communication services Patrick Loiseau EURECOM Fall 2012.
Advertisements

Network Economics -- Lecture 4: Auctions and applications Patrick Loiseau EURECOM Fall 2012.
6.896: Topics in Algorithmic Game Theory Lecture 20 Yang Cai.
CPS Bayesian games and their use in auctions Vincent Conitzer
Nash’s Theorem Theorem (Nash, 1951): Every finite game (finite number of players, finite number of pure strategies) has at least one mixed-strategy Nash.
This Segment: Computational game theory Lecture 1: Game representations, solution concepts and complexity Tuomas Sandholm Computer Science Department Carnegie.
Effort Games and the Price of Myopia Michael Zuckerman Joint work with Yoram Bachrach and Jeff Rosenschein.
1 Combinatorial Agency with Audits Raphael Eidenbenz ETH Zurich, Switzerland Stefan Schmid TU Munich, Germany TexPoint fonts used in EMF. Read the TexPoint.
DARWIN: Distributed and Adaptive Reputation Mechanism for Wireless Ad- hoc Networks CHEN Xiao Wei, Cheung Siu Ming CSE, CUHK May 15, 2008 This talk is.
Infinitely Repeated Games. In an infinitely repeated game, the application of subgame perfection is different - after any possible history, the continuation.
Chapter 14 Infinite Horizon 1.Markov Games 2.Markov Solutions 3.Infinite Horizon Repeated Games 4.Trigger Strategy Solutions 5.Investing in Strategic Capital.
EC941 - Game Theory Lecture 7 Prof. Francesco Squintani
Course: Price of Anarchy Professor: Michal Feldman Student: Iddan Golomb 26/02/2014 Non-Atomic Selfish Routing.
How Bad is Selfish Routing? By Tim Roughgarden Eva Tardos Presented by Alex Kogan.
A Lightweight Currency-based P2P VoD Incentive Mechanism Presented by Svetlana Geldfeld by Chi Wang, Hongbo Wang, Yu Lin, and Shanzhi Chen.
On the Economics of P2P Systems Speaker Coby Fernandess.
Game-theoretic analysis tools Necessary for building nonmanipulable automated negotiation systems.
Benjamin Johnson Carnegie Mellon University Are Security Experts Useful? Bayesian Nash Equilibria for Network Security Games with Limited Information TRUST.
Improving Peer-to-Peer Networks “Limited Reputation Sharing in P2P Systems” “Robust Incentive Techniques for P2P Networks”
1 Algorithmic Game Theoretic Perspectives in Networking Dr. Liane Lewin-Eytan.
Sogang University ICC Lab Using Game Theory to Analyze Wireless Ad Hoc networks.
Modelling and Performance Analysis of BitTorrent-Like Peer-to-Peer Networks.
Game-Theoretic Approaches to Multi-Agent Systems Bernhard Nebel.
Yale Lectures 21 and Repeated Games: Cooperation vs the End Game.
Lecture 1 - Introduction 1.  Introduction to Game Theory  Basic Game Theory Examples  Strategic Games  More Game Theory Examples  Equilibrium  Mixed.
Distributed Multiagent Resource Allocation In Diminishing Marginal Return Domains Yoram Bachrach(Hebew University) Jeffrey S. Rosenschein (Hebrew University)
Service Differentiated Peer Selection An Incentive Mechanism for Peer-to-Peer Media Streaming Ahsan Habib, Member, IEEE, and John Chuang, Member, IEEE.
Beyond selfish routing: Network Formation Games. Network Formation Games NFGs model the various ways in which selfish agents might create/use networks.
Modeling and analysis of BitTorrent-like P2P network Fan Bin Oct,1 st,2004.
Selfish Caching in Distributed Systems: A Game-Theoretic Analysis By Byung-Gon Chun et al. UC Berkeley PODC’04.
A Payment-based Incentive and Service Differentiation Mechanism for P2P Streaming Broadcast Guang Tan and Stephen A. Jarvis Department of Computer Science,
Keeping Peers Honest In EigenTrust Robert McGrew Joint work with Zoë Abrams and Serge Plotkin.
Algorithmic Issues in Non- cooperative (i.e., strategic) Distributed Systems.
Free-rider problem in peer-to- peer networks Sumitra Ganesh.
Algorithmic Issues in Non- cooperative (i.e., strategic) Distributed Systems.
How Bad is Selfish Routing A survey on existing models for selfish routing Professor John Lui, David Yau and Dah-Ming Qiu presented by Joe W.J. Jiang
Anonymizing Web Services Through a Club Mechanism With Economic Incentives Mamata Jenamani Leszek Lilien Bharat Bhargava Department of Computer Sciences.
Price of Anarchy Bounds Price of Anarchy Convergence Based on Slides by Amir Epstein and by Svetlana Olonetsky Modified/Corrupted by Michal Feldman and.
1 Issues on the border of economics and computation נושאים בגבול כלכלה וחישוב Congestion Games, Potential Games and Price of Anarchy Liad Blumrosen ©
Allerton 2011 September 28 Mathias Humbert, Mohammad Hossein Manshaei, and Jean-Pierre Hubaux EPFL - Laboratory for Communications and Applications (LCA1)
A Game Theoretic Framework for Incentives in P2P Systems --- CS. Uni. California Jun Cai Advisor: Jens Graupmann.
Free-riding and incentives in P2P systems name:Michel Meulpolder date:September 8, 2008 event:Tutorial IEEE P2P 2008.
Chapter 12 Choices Involving Strategy Copyright © 2014 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the prior written.
Yitzchak Rosenthal P2P Mechanism Design: Incentives in Peer-to-Peer Systems Paper By: Moshe Babaioff, John Chuang and Michal Feldman.
Beyond Routing Games: Network (Formation) Games. Network Games (NG) NG model the various ways in which selfish users (i.e., players) strategically interact.
1 Efficiency and Nash Equilibria in a Scrip System for P2P Networks Eric J. Friedman Joseph Y. Halpern Ian Kash.
A Non-Monetary Protocol for P2P Content Distribution in Wireless Broadcast Networks with Network Coding I-Hong Hou, Yao Liu, and Alex Sprintson Dept. of.
Ivan Osipkov Fighting Freeloaders in Decentralized P2P File Sharing Systems.
Game-theoretic analysis tools Tuomas Sandholm Professor Computer Science Department Carnegie Mellon University.
Cs286r Victor Chan Scrip Systems Victor Chan. CS286 Victor Chan Agenda  Scrip Systems  Peer to Peer Systems  Scrip Systems for P2P Networks  Adobe.
EE 685 presentation Utility-Optimal Random-Access Control By Jang-Won Lee, Mung Chiang and A. Robert Calderbank.
Incentives for Sharing in Peer-to-Peer Networks By Philippe Golle, Kevin Leyton-Brown, Ilya Mironov, Mark Lillibridge.
Lecture 5 Introduction to Game theory. What is game theory? Game theory studies situations where players have strategic interactions; the payoff that.
Beyond selfish routing: Network Games. Network Games NGs model the various ways in which selfish agents strategically interact in using a network They.
Beyond selfish routing: Network Games. Network Games NGs model the various ways in which selfish users (i.e., players) strategically interact in using.
Oligopoly Theory1 Oligopoly Theory (11) Collusion Aim of this lecture (1) To understand the idea of repeated game. (2) To understand the idea of the stability.
6.853: Topics in Algorithmic Game Theory Fall 2011 Constantinos Daskalakis Lecture 22.
Repeated Game Modeling of Multicast Overlays Mike Afergan (MIT CSAIL/Akamai) Rahul Sami (University of Michigan) April 25, 2006.
Peer-to-Peer Networks 15 Game Theory Christian Schindelhauer Technical Faculty Computer-Networks and Telematics University of Freiburg.
Institutional Analysis Lecture 6: Collective Action.
The Price of Routing Unsplittable Flow Yossi Azar Joint work with B. Awerbuch and A. Epstein.
Network Formation Games. NFGs model distinct ways in which selfish agents might create and evaluate networks We’ll see two models: Global Connection Game.
Network Formation Games. NFGs model distinct ways in which selfish agents might create and evaluate networks We’ll see two models: Global Connection Game.
Game theory basics A Game describes situations of strategic interaction, where the payoff for one agent depends on its own actions as well as on the actions.
Microeconomics 1000 Lecture 13 Oligopoly.
Contents of the Talk Preliminary Materials Motivation and Contribution
Instructor: Ruta Mehta TA: TBA
Unit 4 SOCIAL INTERACTIONS.
Game Theory in Wireless and Communication Networks: Theory, Models, and Applications Lecture 2 Bayesian Games Zhu Han, Dusit Niyato, Walid Saad, Tamer.
The Price of Routing Unsplittable Flow
Presentation transcript:

Network Economics -- Lecture 2: Incentives in P2P systems and reputation systems Patrick Loiseau EURECOM Fall 2012

References Main: – N. Nisam, T. Roughgarden, E. Tardos and V. Vazirani (Eds). “Algorithmic Game Theory”, CUP Chapters 23 (see also 27). Available online: Nisan_Non-printable.pdf Nisan_Non-printable.pdf Additional: – M. Chiang. “Networked Life, 20 Questions and Answers”, CUP Chapters 3-5. See the videos on

Outline 1.Introduction 2.Free-riding and reputation system 3.Hidden actions

Outline 1.Introduction 2.Free-riding and reputation system 3.Hidden actions

P2P networks First ones: Napster (1999), Gnutella (2000) – Free-riding problem Many users across the globe self-organizing to share files – Anonymity – One-shot interactions  Difficult to sustain collaboration Exacerbated by – Hidden actions (nondetectable defection) – Cheap pseudonyms (multiple identities easy)

Incentive mechanisms Good technology is not enough P2P networks need incentive mechanisms to incentivize users to contribute – Reputation (KaZaA) – Currency (called scrip) – Barter (BitTorrent) – direct reciprocity

Extensions Other free-riding situations – E.g., mobile ad-hoc networks, P2P storage Rich strategy space – Share/not share – Amount of resources committed – Identity management Other applications of reputation systems – Online shopping

The P2P file-sharing game Peer – Sometimes download  benefit – Sometimes upload  cost One interaction ~ prisoner’s dilemma CD C D 2, 21, -1 0, 0-1, 1

Prisoner’s dilemma Dominant strategy: D Socially optimal (C, C) Single shot leads to (D, D) – Socially undesirable Iterated prisoner’s dilemma – Tit-for-tat yields socially optimal outcome CD C D 2, 2-1, 3 0, 03, -1

P2P Many users, random interactions Direct reciprocity does not scale Feldman et al. 2004

P2P Direct reciprocity – Enforced by Bittorrent at the scale of one file but not over several files Indirect reciprocity – Reputation system – Currency system

How to treat new comers P2P has high turnover Often interact with stranger with no history TFT strategy with C with new comers – Encourage new comers – BUT Facilitates whitewashing

Outline 1.Introduction 2.Free-riding and reputation system 3.Hidden actions

Reputation Long history of facilitating cooperation (e.g. eBay) In general coupled with service differentiation – Good reputation = good service – Bad reputation = bad service Ex: KaZaA

Trust EigenTrust (Sep Kamvar, Mario Schlosser, and Hector Garcia-Molina, 2003) – Computes a global trust value of each peer based on the local trust values Used to limit malicious/inauthentic files – Defense against pollution attacks

Attacks against pollution systems Whitewashing Sybil attacks Collusion Dishonest feedback See later in this course

A minimalist P2P model Large number of peers (players) Peer i has type θ i (~ ”generosity”) Action space: contribute or free-ride x: fraction of contributing peers  1/x: cost of contributing Rational peer: – Contribute if θ i > 1/x – Free-ride otherwise

Contributions with no incentive mechanism Assume uniform distribution of types

Contributions with no incentive mechanism (2) Equilibria stability

Contributions with no incentive mechanism (3) Equilibria computation

Contributions with no incentive mechanism (4) Result: The highest stable equilibrium contribution level x 1 increases with θ m and converges to one as goes θ m to infinity but falls to zero if θ m < 4

Overall system performance W = ax-(1/x)x = ax-1 Even if participation provides high benefits, the system may collapse

Reputation and service differentiation in P2P Consider a reputation system that can catch free-riders with probability p and exclude them – Alternatively: catch all free-riders and give them service altered by (1-p) Two effects – Load reduced, hence cost reduced – Penalty introduces a threat

Equilibrium with reputation Q: individual benefit R: reduced contribution T: threat

Equilibrium with reputation (2)

System performance with reputation W = x(Q-R)+(1-x)(Q-T) = (ax-1)(x+(1-x)(1-p)) Trade-off: Penalty on free riders increases x but entails social cost If p>1/a, the threat is larger than the cost  No free rider, optimal system performance a-1

FOX (Fair Optimal eXchange) Theoretical approach Assumes all peer are homogeneous, with capacity to serve k requests in parallel and seek to minimize completion time FOX: distributed synchronized protocol giving the optimum – i.e., all peers can achieve optimum if they comply “grim trigger” strategy: each peer can collapse the system if he finds a deviating neighbour

FOX equilibium

Outline 1.Introduction 2.Free-riding and reputation system 3.Hidden actions

Hidden actions in P2P Many strategic actions are not directly observable – Arrival/departure – Message forwarding Moral hazard: situation in which a party is more willing to take a risk knowing that the cost will be supported (at least in part) by others – E.g., insurance

Principal agent model A principal employs a set of n agents: N = {1, …, n} Action set A i = {0, 1} Cost c(0)=0, c(1)=c>0 The actions of agents determine (probabilistically) an outcome o in {0, 1} Principal valuation of success: v>0 (no gain in case of failure) Technology (or success function) t(a 1, …, a n ): probability of success

Read-once networks One graph with 2 special nodes: source and sink Each agent controls 1 link Agents action: – low effort  succeed with probability γ in (0, 1/2) – High effort  succeed with probability 1-γ in (1/2, 1) The project succeeds if there is a successful source-sink path

Example AND technology OR technology

Contract The principal agent can design a “contract” – Payment of p i ≥0 upon success – Nothing upon failure The agents are in a game: The principal wants to design a contract such that his expected profit is maximized

Definitions and assumptions Assumption: – t(1, a -i )>t(0, a -i ) for all a -i – t(a)>0 for all a Definition: the marginal contribution of agent i given a -i is Increase in success probability due to i’s effort

Individual best response Given a -i, agent’s i best strategy is

Best contract inducing a The best contract for the principal that induces a as an equilibrium consists in – for the agents choosing a i =0 – for the agents choosing a i =1

Best contract inducing a (2) With this best contract, expected utilities are – for the agents choosing a i =0 – for the agents choosing a i =1 – for the principal

Principal’s objective Choosing the actions profile a * that maximizes his utility u(a,v) Equivalent to choosing the set S * of agents with a i =1 Depends on v  S * (v) We say that the principal contracts with i if a i =1

Hidden vs observable actions Hidden actions: if a i =1 and 0 otherwise If actions were observable – Give p i =c to high-effort agents regardless of success – Yields for the principal a utility equal to social welfare  Choose a to maximize social welfare

(POU) Price of Unaccountability S * (v): optimal contract in hidden case S 0 * (v): optimal contract in observable case Definition: the POU(t) of a technology t is defined as the worst-case ration over v of the principal’s utility in the observable and hidden actions cases

Remark POU(t)>1

Optimal contract We want to answer the questions: How to select the optimal contract (i.e., the optimal set of contracting agents)? How does it change with the principal’s valuation v?

Monotonicity The optimal contracts weakly improves when v increases: – For any technology, in both the hidden- and observable-actions cases, the expected utility of the principal, the success probability and the expected payment of the optimal contract are all non-decreasing when v increases

Proof

Proof (2)

Consequence Anonymous technology: the success probability is symmetric in the players For technologies for which the success probability depends only on the number of contracted agents (e.g. AND, OR), the number of contracted agents is non-decreasing when v increases

Optimal contract for the AND technology Theorem: For any anonymous AND technology with γ = γ i = 1-δ i for all i – There exists a valuation finite v * such that for any v v *, it is optimal to contract with all agents (for v=v *, both contracts are optimal) – The price of unaccountability is

Remarks Proof in M. Babaioff, M. Feldman and N. Nisan, “Combinatorial Agency”, in Proceedings of EC POU is not bounded! – Monitoring can be beneficial, even if costly

Example n=2, c=1, γ=1/4 Compute for all number of agents – t – Δ – Utility of principal

Optimal contract for the OR technology Theorem: For any anonymous OR technology with γ = γ i = 1-δ i for all i – There exists a finite positive values v 1, …, v n such that for any v in (v k, v k+1 ), it is optimal to contract k agent. (For v v n, it is optimal to contract n agent and for v=v k, the principal is indifferent between contracting k-1 or k agents.) – The price of unaccountability is upper bounded by 5/2

Example n=2, c=1, γ=1/4 Compute for all number of agents – t – Δ – Utility of principal

Illustration Number of contracted agents