MTRMS: MutualTrustbased Reputation Management System

1
PST 2006
MTRMS Mutual-Trust-based Reputation Management
System in Community-based Peer-to-Peer
Environment
EunJoung Byun and et. al. Distributed Systems
Laboratory, Korea University vision_at_disys.korea.ac
.kr
2
Abstract
Problems
Propositions
Effects
Decrease Malicious Dynamic Peer
Activities Stable Reliable P2P Environment

• Peers Features
• Malice,
• Volatility,
• Anonymity
• Reliability,
• Performance
• Degradation

Trust Relationship Novel Reputation MTRMS
3
Contents
1
Introduction
2
Related Work
System Environment
3
3
Markov Modeling
4
Mutual Trust
1
Mutual Reputation
2
Meta Reputation
5
Performance Evaluation
6
Conclusions
4
1. Introduction

• Peer-to-Peer(P2P) Computing
• Distributed computing paradigm
• Self-organize into network topologies
• Without requiring the intermediation or support
  of a global centralized server or authority
• Low cost solution to distributed computing
• Considerations on Trust for reliable P2P
  computing
• Malicious and Fraudulent Peer feedback
• Anonymity and low accountability among peers
• Handle autonomous dynamic characteristics of
  peer

5
1. Introduction
Malice
Volatility
Uncertainty
Mutual Trust-based Reputation Management System
Reputation Information
6
1. Introduction
Caste
Mutual Trust
Mutual Reputation
Meta Reputation
7
2. Related Work

• Reputation Management System in P2P Computing
• PeerTrust
• community-based reputations to trustworthiness of
  peers
• decentralized reputation management framework
• based on feedback, satisfaction, and number of
  transactions
• EigenTrust
• distributed reputation management algorithm
• with a global unique trust value based on the
  past experience
• Gupta
• reputation system for decentralized unstructured
  P2P networks
• Debit-credit reputation computation
• NICE
• local trust inference
• distributed search
• system generates cookies to record direct trust
  between peers
• PET
• recommendation is used to reflect the opinions of
  other peers
• recommendation information collected by the
  feedback
• REGRET

8
3. System environment

• Grand Trust Coalition (GTC)
• Community-based P2P network

9
4. Mutual Trust-based Reputation Management System

• Mutual Trust-based Reputation Management System
  (MTRMS)
• Mutual Trust (MT)
• Provides generic approach to estimate degree of
  trust
• Reputation Rate (RR)
• Supports prediction based on past experience
• Quality of donation
• Entropy of ratio of reputation value
• Mutual Reputation (MuR)
• Dependency of reputation information
• Mutual information of past reputation experience
• Meta Reputation (MeR)
• Evaluates accuracy of recommendation
• Caste
• Indicates social position (level) from reliable
  third party

10
4. Mutual Trust-based Reputation Management System

• Mutual Trust

Reputation Rate
Mutual Reputation
Meta Reputation
11
4. Mutual Trust-based Reputation Management System
12
4. Mutual Trust-based Reputation Management System
13
4. Mutual Trust-based Reputation Management System
14
4. Mutual Trust-based Reputation Management System
15
4. Mutual Trust-based Reputation Management System

• Mutual Trust

Mutual Trust
Mutual Trust
Mutual Trust
16
4. Mutual Trust-based Reputation Management System
17
5. Performance Evaluation

• Simulation
• Transaction Success Rate
• Comparison of MTRMS and No Trust

18
6. Conclusions

• Conclusions
• Future Works
• Implementation on the top of real P2P System

Mutual Trust-based Reputation Management System
19
Questions
Thank You !
Q A