Current Developments in Peer-to-Peer

Mark Luker, EDUCAUSE Bruce Block, RIAASamuel Haldeman, Penn State Eric G. Ferrin, Penn State

P2P File Sharing and Online Music Services

Technology Task Force

Joint Committee of Higher Education and the Entertainment Industry

Joint CommitteeGraham Spanier, President, Pennsylvania State University John L. Hennessy, President, Stanford University Charles Phelps, Provost, University of Rochester Dorothy K. Robinson, Vice President and General Counsel, Yale University Molly Corbett Broad, President, University of North Carolina Cary Sherman, President, Recording Industry Association of America Jack Valenti, President and CEO, Motion Picture Association of America Roger Ames, Chairman and CEO, Warner Music Group Sherry Lansing, Chairman, Paramount Pictures Matthew T. Gerson, Senior VP, Government Relations, Vivendi-Universal Irwin Robinson, Chairman, National Music Publishers Association Chairman and CEO,

Famous Music

Three Task Forces in• Legal issues and legislation• Education• Technology

Technology Task Force

• Charge – Educate the community on the technology issues and options related to the illegal distribution of content using P2P

• Activities• RFI #1 – Controlling illegal file sharing • RFI #2 – Legal, online content services• Share information from pilot projects

RFI #1

• Charge – Document the range of available products and services for the control of illegal file sharing using P2P

• RFI organized by 11 criteria

• 19 responses

• Reviewed and documented by committees

Criteria

• Network architecture• Scalability• Protocol identification• Granularity of protocols• Content identification• Examination of network packets or file content• Distribution systems• Resilience of the technology to countermeasures• Testing and installed base• Competitive approaches• Third-party components

www.educause.edu/issues/rfi

Wide range of proposals

• Restrict bandwidth by location, time, …• Restrict bandwidth by application• Filter “announcements” of available content• Filter copyrighted songs and movies• Detect songs and movies on disk • Detect applications on disk• Monitor all activity on PC• Automate processing of “cease and desist”

Institutional Choices

• Choice of application is up to institutions• Culture and environment• Business considerations• Individual agreements with vendors• Pilot projects

• In conjunction with efforts in• Education• Ethics• Law

Present Implementations

• Recent surveys• 66+% shape bandwidth

• Location, application, time, …• “Top talkers”

• 89% use policy and education• 100% comply with DMCA notices

• “Homegrown” implementations• Block “announcements” of available content• Prohibit servers in residence halls

Technology Challenges

• Hard to discriminate violations of law

• Very large numbers of cases

• Technology (and policy) issues of privacy

• Arms race

• Solutions will require education and policy in addition to technology!

Long Term Solutions?

• Legal, online content• New services• New business models• Disruptive innovations

• RFI #2

• www.educause.edu/issues/rfi

Napster/Penn State Initiative

Samuel Haldeman

Pennsylvania State University

What was needed

• In addition to education and enforcement we also provided a viable, acceptable, legitimate alternative.– Viable: Was it robust enough?– Acceptable: Did it have the features and service

students have come to expect– Legitimate: Was it legal, was it fair?

The Napster Service

• Superior song content and access

• Simple to learn, use, and navigate

• Tremendous features and amenities

• Community-like design

• Adopted Shibboleth Technology

First Five Hours

• Over 4,200 students registered for service

• Over 10,000 hits to napster.psu.edu

• 51,000 songs streamed

• 42,000 songs downloaded

• Fewer than 10 errors/complaints

Since January 12th

• 7,500 registered users– 64% of University Park Students– 49% of non-University Park students

• Averaging more than 100,000 songs streamed or downloaded per day

• Fewer than 100 errors or complaints

Input from napster.psu.edu

• 41 total suggestions/complaints/questions– 6 complained about OS availability– 4 suggestions to make it available to alumni– 6 questions about registration– 12 questions regarding the policy– 13 other (include bandwidth limitations,

interest from other schools, etc.)

Input from focus groups

• Nice user interface: simple to learn, navigate, and use

• Plenty of features and amenities– Appreciate streaming and downloading, radio

• Not enough popular artists– Have to pay for newer albums/tracks

• Not compatible with Mac OS, Windows ‘98 & ME, and Linux

Positives Reactions

• Students want to use the service

• Simple, easy to use

• Features and amenities make a difference

• “Sharing” is not necessarily important

• Legal approaches can work– Students are anxious, excited, optimistic

Negatives Reactions

• Not all students can have access

• Not all students can use the service

• Newer tracks and albums have to be bought

• Partnerships seem more like sell-outs

What we can learn

• Circumventing or revolting against the system isn’t what students want

• OS availability and content are two biggest complaints

• Access NOT ownership

• Biggest motivating factor: cost– Plain and simple, if it’s free it’s better

Action

• What should be done, and by whom?– Is it our responsibility?

• It might be, but in any case we can…– Emphasize to the computer industry that standards

would help– Emphasize to artists and labels music can and should be

distributed legally electronically– Promote legal, legitimate file sharing on campus– Investigate services that may work for your campus– Enforce the DMCA and university policy

Success?

• Never guaranteed– Failure is.

• How can we measure the success?– Analytically?– Subjectively?– Ever?

The edge of the network is the new frontier

•It is personal•It is social•It is nomadic•It is connected…

The edge of the network is the new frontier

•It is personal•It is social•It is nomadic•It is connected… … and disconnected

Agenda

• LionShare background

• Provide an overview of LionShare

• Discuss Current/Future Efforts

LionShare Origins

• Visual Image User Study (VIUS)• Hosted in University Libraries at Penn State• A two-year user study sponsored by Mellon

Foundation• Looking at how fac/staff/students use digital

images in teaching, research and service• Several prototypes identified

– Peer-to-Peer was identified as one possible solution - LionShare prototype

VIUS Identified These Problems

• Need for tools to manage personal collections

• Rapid movement from analog to digital

• Difficulty in finding appropriate resources

• Difficulty merging public/private collections

• Need for faculty/student/dept’s to manage large collections

• Need for copyright and access control

Why Use P2P for LionShare?

• Encourages collaboration– Student, faculty members and departments

• Helps manage the digital media explosion– Digital consumer devices

• Provides common organizational structure– Metadata and standards

• Flexible and Scalable– Customizable for different needs

LionShare Uses

• Media organization (offline use)

• Publish personal media collections

• Person-to-person collaboration

• Group projects

• Departmental collaboration

• Formation of user communities

• Publication of academic collections

LionShare Design Goals

• Media management

• Simple, intuitive interface

• User-defined sharing

• Authenticated access to the network

• Standard descriptive metadata structure

• Leverage Open Source

LionShare Architecture

• Based on Limewire Open Source project

• Modified version of the Gnutella protocol– P2P + Client/Server Architecture– Decentralized + Centralized Topology

• Integrated Authentication with Kerberos

LionShare Architecture

• LionShare adds the concept of a PeerServer

PeerServer• Local aggregator• Adds persistence to P2P• Can function as gateway • Web interface

LionShare PrinciplesThe Three A’s

• Authentication– Kerberos

• Authorization– Access Control

• Accountability– Non-anonymous network

• Userid associated with shared files• Activity logging

LionShare Topology

• Private P2P Network• Hybrid topology (P2P+Client/Server)

– PeerServers• Users can publish metadata and/or files to a server

to remain shared on the P2P network even though the user is not connected to the network

– PeerServer Uses• Off-line sharing• Remote backup• WWW publication possibilities

Peer Peer

Institution“A”

P2PNetworks

LionShare Conceptual Design

LionShare Conceptual Design

PeerServer

Peer Peer

Institution“A”

P2PNetworks

LionShare Conceptual Design

PeerServer

Peer Peer

Institution“A”

AuthenticationService

P2PNetworks

Current Development Status

• Accomplishments to date– Basic architectural design complete– LionShare App Alpha Release

• MIT Kerberos compatibility• XML schemas for learning object description• Protocol customization

– LionShare PeerServer prototype• Alpha pre-release

– Mellon development grant just awarded

Future Development Plans

• Hardening the LionShare Application

• Re-engineering the PeerServer

• Federation of the LionShare protocol– Shibboleth-like implementation

• Connecting to Fixed Repositories– IMS DRI Spec/OKI OSIDs– Access to Large Collections

LionShare Conceptual Design

PeerServer

Peer Peer

Institution“A”

AuthenticationService

P2PNetworks

LionShare Conceptual Design

PeerServer

Peer Peer

InstitutionalBoundary

PeerServer

Peer Peer

Institution“A”

Institution“B”

AuthenticationService

AuthenticationService

P2PNetworks Trust fa

bric[Shibboleth-like]

LionShare Conceptual Design

PeerServer

Peer Peer

InstitutionalBoundary

PeerServer

Peer Peer

Institution“A”

Institution“B”

AuthenticationService

AuthenticationService

P2PNetworks Trust fa

bric[Shibboleth-like]

LionShare Conceptual Design

PeerServer

Peer Peer

InstitutionalBoundary

PeerServer

Peer Peer

Institution“A”

Institution“B”

AuthenticationService

AuthenticationService

P2PNetworks Trust fa

bric[Shibboleth-like]

OSIDTranslator

Gateway

Gateway

FixedRepository

(Merlot, Careo, EdNA)

Future P2P Directions

• NomadicP2P networks– wireless hand-helds P2P networks– Personal Learning Management Systems

• Personal ePortfolio Management Systems

• Personal Back-up Solution

• Distributed Academic Resource Systems

• Security and encryption

• End-to-end diagnostics

Co-development Effort

• Open Source Project

• Partnering to develop these ideas

• Some clearly out of scope for current effort

• Website will develop some of these ideas

• Informal community approach

LionShare Team

• Penn State University

• Internet 2 Middleware and P2P WGs

• eduSource Canada/Simon Fraser U.

• MIT - Open Knowledge Initiative (OKI)

• And others– Dartmouth, Florida, Georgia Tech…

Contact

• Interim LionShare website:– http://lionshare.its.psu.edu

• Mike Halm mjh@psu.edu

• Alex Valentine asv108@psu.edu