WHAT'S NEW?

[November 19, 2001] Schedule updated. Due date for final paper posted.

[October 29, 2001] Some new links added to the list of commercial vendors.

[October 25, 2001] A few updates to the schedule.

[October 15, 2001] A variety of small changes to the schedule, plus addition of the Citations section.

[October 8, 2001] Art Hitomi's slides are now available.

[October 4, 2001] .pdf versions of intro slides and Greg's slides now available.

[October 2, 2001] Greg Bolcer's slides from Monday, October 1, are now available.

[September 27, 2001] Posted my lecture slides from week one, and updated a few more of the references.

[September 25, 2001] Added links to some Gnutella items

[September 20, 2001] Public unveiling of the website.

Instructors

• Professor Richard N. Taylor
  • Electronic Mail: taylor@ics.uci.edu
  • Office: ICS2-203
  • Office Phone: (949) 824-6429
  • Hours: Monday and Wednesday: 11-12:20. Please make appointments by email, however, in order to guarantee the meeting.
• Fax: (949) 824-1715
• Mailing Address:
  Department of Information and Computer Science,
  University of California, Irvine
  Irvine, California 92697-3425

Overview

We will examine the peer-to-peer phenomenon, looking at the underlying technical issues and notional applications. We will attempt to identify the key open topics, assess their importance in the long term, and develop a research agenda for tackling those issues.

The class will operate as a research seminar: I will lecture occasionally, but there will be numerous presentations by students. The focus of class time will be discussion of what's been learned outside of class.

Grading will be based on participation in the class, and a term paper +/- a project. (That means, you'll almost assuredly have to write a term paper, and that paper in many cases will reflect work you've done in a project.) Projects can certainly be worked on in teams. We might even be using p2p technologies to support the teams! It is highly unlikely that there will be a final exam.

Prerequisite Knowledge

Students should have some background in software engineering including exposure to the basic ideas of software architecture. This is a research seminar and students will be expected to participate actively in class discussions and presentations. Some form of project will take place and to be successful with that students will need to be able to independently acquire, install, adapt, and experiment with externally developed p2p systems. Students will also need to independently discover, obtain, and analyze publications germane to the topic, and present the results of those investigations to the class.

Research Issues

There's a lot of research issues in this domain. The following list gives you some idea.

• The relationship between requirements and architecture-based development, with a focus on mapping between them, or transformation, or development: all this as applied particularly to the P2P domain
• Supporting groups of people with peer-to-peer (p2p) infrastructure. "WebDAV grows up." Supporting little processes; social issues; quick setup and tear-down.
• Search technologies in the p2p world.
• Designing, modeling, and implementing p2p applications
  • formal models
  • an architecture based approach; xADL + richer groups, process, and security
  • responsibility, ownership, and intellectual property models
• Analyzing p2p architectures
• Pervasive p2p applications. Think ubiquitous computing as implemented by p2p applications and micro-networking.
• P2P infrastructures and network issues/security issues.
• Instant messaging, p2p, and process. Protocols for this, infrastructure, and interoperability.
• Events: XML rep and interchange in the p2p space and beyond
• The Ultimate Peer-to-Peer Protocol (TuP3)

Comparison Factors

The following list is incomplete, but illustrates the kinds of factors we'll be looking at in evaluating peer-to-peer architectures

• Overview
  • Contact Information
  • Overview
  • Best Feature
  • Most Telling Criticisms
  • Business Model
  • Target Companies
  • Investors
  • Terms and Conditions
  • History
  • Product Trajectory
  • Source of Info
• Peer Applications
  • Searching
  • Synchronization
  • Collaboration
  • Decentralized Work
  • Remote Control
  • Dispatching
  • Backup
  • E/M-commerce
  • Home Networking
• Clients
  • Main GUI
  • Coupled/Decoupled
  • Others
  • Ease of Use
  • Installation
  • Learning Curve
  • Drag & Drop
  • Auditing
  • Extensions
  • Electronic Wallet
• Networking
  • Disconnected
  • Mobile
  • Security
  • Firewall
  • Writeable
  • Two-way Routing
  • Replication
  • Discovery
  • Wireless Access
• Scalability
  • Broadcast/Anycast
  • Namespaces
  • Event Notification
  • Web Integration
  • Publish/Subscribe
  • Embedded
  • SyncML/XML
  • WML/XHTML
• Robustness
• Extensibility
• Open Source

Example P2P systems and infrastructure

We'll categorize and examine several of these...

• Magi
• Groove.Net (and since their site is such a mess, you might zip to http://devzone.groove.net/library/ )
• eMikolo
• Lightshare
• Flycode (Apple Soup)
• Kalepa
• MojoNation
• Napster
• Gnutella
• Freenet
• Aimster 1.0
• Scour
• Static (is now Blue Falcon)
• GameSpy a bit off-beat, I'll admit, but you might try looking at this link
• iMesh
• Pointera
• GoneSilent (Infrasearch)
• Popular Power (now out of business, but the original site is still there)
• United Devices
• Radio Userland
• KnowNow
• Infraworks
• Agranat
• US Software
• Enfish
• Engenia (and they have now acquired the assets of Roku)
• Kazaa
• Morpheus/MusicCity
• Omniprise
• Think Desk
• Avaki
• HotComm

Textbooks

• Peer-to-Peer: Harnessing the Power of Disruptive Technologies, Edited by Andy Oram. O'Reilly.2001
• Other readings from journals and conferences, as assigned.

You might be interested in looking at:

• A p2p "research" report from O'Reilly (actually, just a little ad for it) And you thought textbooks were expensive!

Schedule (Subject to change)

Academic Dishonesty

I've never had an instance of cheating in a graduate seminar in my life. Nonetheless I suppose someone might plagiarize sooner or later. Thus: "Cheating in ICS 280 will be punished in accordance with University policy and ICS policy. Please familiarize yourself with those documents. Note that University policy states that faculty have the responsibility of "assigning an appropriate grade to a student who engages in academic dishonesty." That appropriate grade, for this class, is an F. Cheating is wrong. Don't do it."

Citations

Here's some background documents that may be relevant to your projects.

1. "Peer-to-peer: A security nightmare or a secure opportunity" The Magi folks.
2. "An efficient and open implementation of the Minstrel broadcasting infrastructure" Stefan Haberl, TU Wien. (I have this document... let me know if you want to borrow it.)
3. "An implementation of the Millicent micro-payment protocol and its application in a pay-per-view business model." Michael Puhrerfellner. TU Wien. (Ditto)
4. "The JEDI event-based infrastructure and its application to the development of the OPSS WFMS" G. Cugola, E. DiNitto, and A. Fuggetta. IEEE TSE. (If you want the real article, you'll have to do to the IEEE TSE site.)
5. The February issue of CACM. It has a set of articles on securing network software applications (like p2p) and on intellectual property (yes, less relevant to the class).
6. "P2P Networking: An information-sharing alternative" IEEE Computer, July 2001, vo. 34. no. 7.
7. Two KnowNow whitepapers on Event-based Notification. ISENS and EBI