Peer-To-Peer by Oram, Andy -- Read -- Imperial Library of Trantor

Index

Peer to Peer Preface

Some context and a definition How this book came into being Contents of this book Peer-to-peer web site We’d like to hear from you

I. Context and Overview

1. A Network of Peers: Peer-to-Peer Models Through the History of the Internet

A revisionist history of peer-to-peer (1969-1995)

Usenet DNS

The network model of the Internet explosion (1995-1999)

The switch to client/server The breakdown of cooperation

Spam: Uncooperative people The TCP rate equation: Cooperative protocols

Firewalls, dynamic IP, NAT: The end of the open network Asymmetric bandwidth

Observations on the current crop of peer-to-peer applications (2000)

Authoring is not the same as publishing Decentralization Abusing port 80

Peer-to-peer prescriptions (2001-?)

Technical solutions: Return to the old Internet Social solutions: Engineer polite behavior

Conclusions

2. Listening to Napster

Resource-centric addressing for unstable environments

Peer-to-peer is as peer-to-peer does The veil is pierced Real solutions to real problems Who’s in and who’s out? Peer-to-peer is a horseless carriage

Follow the users

Users reward simplicity Listen to Napster

It’s the applications, stupid Decentralization is a tool, not a goal

Where’s the content?

PCs are the dark matter of the Internet Promiscuous computers

Nothing succeeds like address, or, DNS isn’t the only game in town

An explosion of protocols

An economic rather than legal challenge

All you can eat Yesterday’s technology at tomorrow’s prices, two days late 30 million Britney fans does not a revolution make

Peer-to-peer architecture and second-class status

Users as consumers, users as providers New winners and losers

3. Remaking the Peer-to-Peer Meme

From business models to meme maps

A success story: From free software to open source The current peer-to-peer meme map The new peer-to-peer meme map

File sharing: Napster and successors Mixing centralization and decentralization: Usenet, email, and IP routing Maximizing use of far-flung resources: Distributed computation Immediate information sharing: The new instant messaging services The writable Web Web services and content syndication Peer-to-peer and devices

Strategic positioning and core competencies

4. The Cornucopia of the Commons

Ways to fill shared databases

CDDB: A case study in how to get a manually created database Napster: Harnessing the power of personal selfishness The commons

II. Projects

5. SETI@home

Radio SETI How SETI@home works Trials and tribulations Human factors The world’s most powerful computer The peer-to-peer paradigm

6. Jabber: Conversational Technologies

Conversations and peers Evolving toward the ideal Jabber is created

The centrality of XML Pieces of the infrastructure

Identity Presence Roster

Architecture

Protocols Browsing Conversation management

Conclusion

7. Mixmaster Remailers

A simple example of remailers Onion routing How Type 2 remailers differ from Type 1 remailers General discussion

8. Gnutella

Gnutella in a gnutshell A brief history

Gnutella’s first breath Open source to the rescue

What makes Gnutella different?

Gnutella works like the real world

A Gnutella cocktail party A client/server cocktail party

Client/server means control, and control means responsibility The client is the server is the network Distributed intelligence Different from Freenet

Gnutella’s communication system

Message-based, application-level routing TCP broadcast Message broadcasting Dynamic routing Lossy transmission over reliable TCP

Organizing Gnutella

Placing nodes on the network

Gnutella’s analogues

The Gnutella horizon Cellular telephony and the Gnutella network Ethernet Cultivating the Gnutella network

Gnutella’s traffic problems

Host caches Returning the network to its natural state Private Gnutella networks Reducing broadcasts makes a significant impact

The policy debates

Napster wars Anonymity and peer-to-peer

Gnutella pseudoanonymity Downloads, now in the privacy of your own direct connection Anonymous Gnutella chat

Next-generation peer-to-peer file-sharing technologies

Gnutella’s effects

9. Freenet

Requests

Detail of requests The data store

Keys

Key types

Content Hash Keys (CHKs) Keyword Signed Keys (KSKs) Signature Verification Keys (SVKs)

Keys and redirects

Conclusions

10. Red Rover

Architecture

The hub The clients The subscribers

Client life cycle Putting low-tech “weaknesses” into perspective Acknowledgments

11. Publius

Why censorship-resistant anonymous publishing?

Publius and other systems in this book

System architecture Cryptography fundamentals

Encryption and decryption Secret sharing Hash functions

Publius operations

Publish operation Retrieve operation Update operation Delete operation

Publius implementation

User interface

Server software Client software

Publius MIME type Publius in a nutshell

12. Free Haven

Privacy in data-sharing systems

Reliability with anonymity

Anonymity for anonymous storage

Partial anonymity

The design of Free Haven

Elements of the system Storage Publication Retrieval Share expiration Document revocation Trading Receipts Accountability and the buddy system Communications channel Reputation system Introducers Implementation status

Attacks on Free Haven

Attacks on documents or the servnet Attacks on the reputation system Attacks on anonymity

An analysis of anonymity Future work Conclusion Acknowledgments

III. Technical Topics

13. Metadata

Data about data Metadata lessons from the Web

Resource description Searching

Resources and relationships: A historical overview

Foundations of resource description: Unique identifiers A contrasting evolution: MP3 and the metadata marketplace

Conclusion

14. Performance

A note on terminology Why performance matters Bandwidth barriers It’s a small, small world

An excursion into graph theory The small-world model

Case study 1: Freenet

Initial experiments Simulating growth Simulating fault tolerance Link distribution in Freenet The impact of free riding Scalability

Case study 2: Gnutella

Initial experiments Fault tolerance and link distribution in Gnutella The impact of free riding Scalability

Conclusions Acknowledgments

15. Trust

Trust in real life, and its lessons for computer networks Trusting downloaded software

Message digest functions Digital signatures Digital certificates Signature verification Open source software Sandboxing and wrappers

Trust in censorship-resistant publishing systems

Publius in a nutshell Risks involved in web server logging Anonymizing proxies Censorship in Publius

Using the Update mechanism to censor

Publius proxy volunteers

Third-party trust issues in Publius

Other anonymity tools

SSL Mix networks Crowds

Denial of service attacks

Quota systems CPU-based payment schemes Anonymous e-cash payment schemes

Legal and physical attacks

Trust in other systems

Mojo Nation and Free Haven The Eternity Service

Eternity Usenet

File-sharing systems

Napster Gnutella Freenet

Content certification

Trust and search engines

Distributed search engines Deniability

Conclusions

16. Accountability

The difficulty of accountability

Special problems posed by peer-to-peer systems Peer-to-peer models and their impacts on accountability Purposes of micropayments and reputation systems Junk mail as a resource allocation problem Pseudonymity and its consequences

Problems with pseudospoofing and possible defenses Reputation for sale—SOLD!

Common methods for dealing with flooding and DoS attacks

Caching and mirroring Active caching and mirroring

Micropayment schemes

Varieties of micropayments or digital cash Nonfungible micropayments

Extended types of nonfungible micropayments Nonparallelizable work functions

Fungible micropayments

Freeloading Fungible payments for accountability Micropayment digital cash schemes Making money off others’ work Anonymous macropayment digital cash schemes

The use and effectiveness of micropayments in peer-to-peer systems

Identity-based payment policies General considerations in an economic analysis of micropayment design Moderating security levels: An accountability slider

Reputations

Early reputation systems online

Codifying reputation on a wide scale: The PGP web of trust Who will moderate the moderators: Slashdot Reputations worth real money: eBay A reputation system that resists pseudospoofing: Advogato System successes and failures

Scoring systems

Attacks and adversaries Aspects of a scoring system Collecting ratings Bootstrapping Personalizing reputation searches Scoring algorithms Privacy and information leaks

Decentralizing the scoring system

Multiple trusted parties True decentralization

A case study: Accountability in Free Haven

Micropayments

The difficulty of distributed systems: How to exchange micropayments among peers Micropayments in the Free Haven context

Reputation systems Other considerations from the case study

Conclusion Acknowledgments

17. Reputation

Examples of using the Reputation Server Reputation domains, entities, and multidimensional reputations Identity as an element of reputation Interface to the marketplace Scoring system Reputation metrics Credibility Interdomain sharing Bootstrapping Long-term vision Central Reputation Server versus distributed Reputation Servers Summary

18. Security

Groove versus email Why secure email is a failure The solution: A Groove shared space Security characteristics of a shared space Mutually-trusting shared spaces

Anatomy of a mutually-trusting shared space The key to mutual trust

Mutually-suspicious shared spaces

Message fanout Fetching lost messages

Shared space formation and trusted authentication Inviting people into shared spaces The New-Member-Added delta message Key versioning and key dependencies Central control and local autonomy Practical security for real-world collaboration Taxonomy of Groove keys

19. Interoperability Through Gateways

Why unification?

Why not an ÜberNetwork? Why not an ÜberClient? Why not just use XML?

One network with a thousand faces Well-known networks and their roles

Freenet Gnutella Mojo Nation Free Haven and Publius

Problems creating gateways

Problems with inserts Problems with requests

Freenet Gnutella Mojo Nation Publius Free Haven

Gateway implementation Existing projects Conclusion Acknowledgments

20. Afterword

Precedents and parries Who gets to innovate? A clean sweep?

A. Directory of Peer-to-Peer Projects B. Contributors Index Copyright

← Prev
Back
Next →

← Prev
Back
Next →