Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Introduction

1. Index

2. Simple P2P

3. Index Poisoning

4. Purpose of Index Poisoning

5. Cost-Effective

Information containing the name, owner, size

and the other attributes of the file

Index

Introduction

A

B

query

Index

Simple P2P

Introduction

Introduction

Index Poisoning

A technique which alters the index of illegal files

so that they could not be reached by any peer which wishes to access them through P2P

Introduction

Purpose of Index Poisoning

To reduce the probability of correct identification

as much as possible by reducing the number of correct copies of the index of illegal files

Introduction

Cost-Effective

1. Network cost

Evaluated by the total number of messages and the load of agent to poison index

2. Spatial cost

Evaluated by the number of altered copies existing in the network

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Proposed Scheme

Overview

Proposed Scheme

Overview

The identification of upstream, midstream, and downstream peers can be realized by  the behavior of each peer concerned with the index propagation

More concretely, this can be achieved by the software, Winny

Proposed Scheme

Overview

Proposed Scheme

Step 1: Injection to remove the seed of propagation

• Target at the illegal file of the file owners

• Overwrite a copy of the index held by the file owner with an altered one

• Stop the new propagation of the correct index

Proposed Scheme

Step 2: Injection to remove propagated copies

• Inject copies of altered index to upstream peers by setting a long lifetime

• The altered index will then be received by many peers

Proposed Scheme

Step 3: Injection to remove propagated copies

• Inject copies of altered index to downstream peers by setting a Short lifetime

• Aim to overwrite the index of the illegal files previously received.

• No propagation, and therefore shorter lifetime will be more cost-effective

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Evaluation

Tools

• PeerSim
  - an open source P2P simulator

• Winny
  - a monitor to compare the performance of the proposed scheme

Evaluation

Configuration

• Number of peers: 1000

• Kind of files: 500

• Number of keys initially held by each peer: 2

• Maximum number of keys held by each peer:
  - 500 for downstream peers
  - 300 for midstream peers
  - 100 for upstream peers

• Number of peers of each type:
  - 235 are downstream peers
  - 485 are midstram peers
  - 280 are upstream peers

• Default lifetime of index: 1500 seconds

• Simulation time: 3870 seconds

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Results

Time variation of the number of correct copies vs simple scheme

Results

The effect of the injection to downstream peers

Results

Impact of the update interval of file owner

Results

Impact of the update interval of file owner

Results

Hit rate of queries requesting the correct index of illegal file

Results

Impact of the update interval of file owner

Results

Impact of the update interval of file owner

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Outline

1. Introduction

2. Proposed Scheme

3. Evaluation

4. Result

5. Conclusion

Conclusion

The proposed scheme could attain almost the same performance with a simple (but expensive) scheme in which altered index with a sufficiently long lifetime is directly injected to all peers in the system

Presenter: Denffer (趙之宇)

2017@Department of Computer Science, University of Taipei

Thank you