Towards Digital Rights Protection in BitTorrent-like P2P Systems Xinwen Zhang Samsung Information Systems America Dongyu Liu and Songqing Chen George Mason.

Towards Digital Rights Protection in BitTorrent-like P2P Systems

Xinwen ZhangSamsung Information Systems America

Dongyu Liu and Songqing ChenGeorge Mason University

Zhao ZhangIowa State University

Ravi SandhuUniversity of Texas at San Antonio,

MMCN 2008

BitTorrent P2PBitTorrent P2PBy the end of 2004, BitTorrent (BT) was accounting for as much as By the end of 2004, BitTorrent (BT) was accounting for as much as 30% of all Internet traffic.30% of all Internet traffic.– P2P traffic is 60% of all Internet trafficP2P traffic is 60% of all Internet traffic– Data from CacheLogicData from CacheLogic

compared to E-mail, FTP and the Web in general.compared to E-mail, FTP and the Web in general.

TrendsTrendsBT Keeps GrowingBT Keeps Growing– In May 2006, the average torrent had 817,588 people In May 2006, the average torrent had 817,588 people

participating. 12 months later, that figure had jumped to participating. 12 months later, that figure had jumped to 1,357,318 seeders and leechers: a 1,357,318 seeders and leechers: a 66 percent66 percent year-over- year-over-year growth rate. year growth rate.

http://arstechnica.com/news.ars/post/20070705-p2p-traffic-shifts-away-from-music-towards-movies.htmlhttp://arstechnica.com/news.ars/post/20070705-p2p-traffic-shifts-away-from-music-towards-movies.html

– P2P applications account from anywhere between 50 P2P applications account from anywhere between 50 percent and 90 percent of all Internet traffic percent and 90 percent of all Internet traffic

BitTorrent accounted for between 50 percent to 75 percent BitTorrent accounted for between 50 percent to 75 percent in 2006 in 2006

– http://arstechnica.com/news.ars/post/20070903-p2p-responsible-for-as-much-as-90-percent-of-all-http://arstechnica.com/news.ars/post/20070903-p2p-responsible-for-as-much-as-90-percent-of-all-net-traffic.htmlnet-traffic.html

P2P file sharing Traffic shifts from small file-sizes P2P file sharing Traffic shifts from small file-sizes to huge file-sizesto huge file-sizes– From music to movies, TV shows, and full albumsFrom music to movies, TV shows, and full albums

Current StatusCurrent StatusThere are some legal content sharingThere are some legal content sharing– Open source software distributionsOpen source software distributions

However, most content shared with BT are However, most content shared with BT are not copyright-protected.not copyright-protected.

No practical DRM mechanism built for BTNo practical DRM mechanism built for BT– Due to its highly distributed environment and Due to its highly distributed environment and

uncontrolled peer activitiesuncontrolled peer activities

Motivation & ChallengesMotivation & ChallengesEnable DRM in BT-like P2P systemEnable DRM in BT-like P2P system– Leverage efficiency of BT for legal content Leverage efficiency of BT for legal content

distributionsdistributions– Enable new business model with P2PEnable new business model with P2P

Challenges: Challenges: – Existing DRM schemes are client-server modelExisting DRM schemes are client-server model– Open platforms of peersOpen platforms of peers– Should be no demanding infrastructure

changesP2P users are “loath to change”.

Approach & FeaturesApproach & FeaturesApproach:Approach:– An re-encryption crypto scheme based on An re-encryption crypto scheme based on asymmetric

encryption algorithm– Each piece is re-encrypted at runtime before a peer uploads it

to any other peer so that the decryption keys are unique for both different peers and difference pieces.

Features: – Leverage the lightly-centralized tracker site to store and Leverage the lightly-centralized tracker site to store and

distribute re-encryption keysdistribute re-encryption keysTypically is trusted by content provider or ownerTypically is trusted by content provider or owner

– Any user can take part in the content distribution while only legitimate users can access the plaintext

BitTorrent’s Basic IdeaBitTorrent’s Basic IdeaTo best utilize parallel downloading, file is treated as many To best utilize parallel downloading, file is treated as many piecespieces::– Using Using SHA-1SHA-1 to ensure every piece’s data integrity to ensure every piece’s data integrity– Using “Using “rarest firstrarest first” strategy for piece selection to speed up ” strategy for piece selection to speed up

distributiondistribution

Using a Using a tit-for-tattit-for-tat algorithm (choking for the lowest speed algorithm (choking for the lowest speed uploader) to encourage uploadinguploader) to encourage uploading

Typically there’re Typically there’re Web serversWeb servers to to serve .torrent metainfo fileserve .torrent metainfo file

TrackerTracker: maintain all peers’ IP/port : maintain all peers’ IP/port information, so each peer can find information, so each peer can find othersothers

Downloaders Downloaders (leecher/seed) as peers (leecher/seed) as peers connected to each otherconnected to each other

Content DistributionContent DistributionOriginal File(s)

.torrent file(containing a valid tracker address)

Distribute .torrent file:Post .torrent file to any web server for people to download

Computing using CompleteDir

Serving as a seedas long as possible

Running BitTorrent Downloader with the .torrent file

Content Publisher

Overall ArchitectureOverall Architecture

Web page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

TrackerWeb Server

.torr

ent

Overall ArchitectureOverall Architecture

Web page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

Tracker

Get-announce

Web Server

Overall ArchitectureOverall Architecture

Web page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

Tracker

Response-peer list

Web Server

Overall ArchitectureOverall Architecture

Web page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

Tracker

Shake-hand

Web Server

Shake-hand

Overall ArchitectureOverall Architecture

Web page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

Tracker

pieces

pieces

Web Server

Overall ArchitectureOverall Architecture

Web page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

Tracker

piecespieces

pieces

Web Server

Overall ArchitectureOverall ArchitectureWeb page with link to .torrent

A

B

C

Peer

[Leech]

Downloader

“US”

Peer

[Seed]

Peer

[Leech]

Tracker

Get-announce

Response-peer list

piecespieces

pieces

Web Server

Enable DRM for BTEnable DRM for BTOverall DRM requirements: Overall DRM requirements: – Encrypted contentEncrypted content– Flexible policy management (e.g., through license file)Flexible policy management (e.g., through license file)– Usage trackingUsage tracking

Three different DRM modelsThree different DRM models– Model 1: encode policy in object

Static policy, no update, no tracking– Model 2: separated policy file with object

Same object copy for all users (same encryption key)Break-once-run-everywhere

– Model 3: separate policy from object, and each user obtains a different copy of the object

each object is encrypted with a different keyOr each encrypted object has a unique IDStrong trackingStrongest DRM schemeStrongest DRM scheme

Case Study: Windows Media Digital Rights ManagementCase Study: Windows Media Digital Rights Management

Enable Strong DRM in BTEnable Strong DRM in BT

Overall requirements: Overall requirements: – Confidentiality of data transition between any

two peers– no-linkable content secrecy– Immune to passive attacks with intercepted

messages.– Immune to compromised peers.– Immune to collusion between any number of

peers.

Some Naive SolutionsSome Naive SolutionsSymmetric key algorithm– Each pair of peers shares a symmetric key– Key management is an issue– a peer can share a file piece only after

decryption, which makes the system vulnerable to the attack that a malicious peer could upload/distribute plain file pieces to others

Traditional public-key algorithm– Each peer has a public/private key pair.– Same problems

Secure BT scheme for Strong DRMSecure BT scheme for Strong DRM

Secure BT SchemeSecure BT Scheme

Peer k Peer j

Web Server

Tracker SiteWeb page withlink to .torrent file

Peer ID

Public Key

TS Keys

1 g^s1 r1,1,r1,2,…r1,N 2 g^s2 r2,1,r2,2,…r2,N … … … j g^sj rj,1,rj,2,…rj,N k g^sk rk,1,rk,2,…rk,N

1. G

et .

torr

ent

file

3. R

espo

nse-

peer

list

2. G

et-a

nnou

nce

6. Re-encryption K

ey: g^(ri,k s

k -ri,j s

j ))

4. Shake hand

7. Cipher Piece: mig^(ri,ksk)

5. UploadR

equest: (i, Pj , P

k )

Security of the SchemeSecurity of the Scheme

PerformancePerformanceCryptography performanceCryptography performance– PEnc is only performed by the initial seed.PEnc is only performed by the initial seed.– PDec is only performed when playingPDec is only performed when playing– Selective encryption and pre-processing can increase Selective encryption and pre-processing can increase

the speed. the speed.

PerformancePerformanceCommunication overhead measurenent in PlanetLab:Communication overhead measurenent in PlanetLab:– 4 dedicated seeds are set up with an uploading speed of 200

KB/s.– Randomly selected 120 PlanetLab nodes are used as

downloaders,from Asia, Europe, and United States.

– The object is a 640-MB file. Both the seeds and the tracker are running Celeron CPU 2.4 GHz with 1 GB memory, and Linux Fedora 2.6.9 and Python 2.3.4.

– Change piece size to simulate different download traffic from seeds.

512KB vs. 120 nodes256KB vs. 240 nodes128KB vs. 480 nodes64KB vs. 960 nodes32KB vs. 1920 nodes

PerformancePerformanceSlightly increase on tracker response time for single Slightly increase on tracker response time for single download requestdownload request

Less than 10% system throughput decreaseLess than 10% system throughput decrease

ConclusionsConclusionsBT high efficiency has not been leveraged to distribute the majority of copyrighted digital content over the Internet. We propose a security mechanism based on the existing BT infrastructureto enable copyright protection. We have implemented a prototype system and conducted real experiments in PlanetLab. The evaluation results show that our scheme can still achieve comparable content distribution efficiency to the original BT system. – That is, to enable DRM, our proposed scheme causes less than

10% degradation of the system throughput.

We are further optimizing the system performanceWe are exploring DRM mechanism on tracker-less BT We are exploring DRM mechanism on tracker-less BT systemssystems

Thank you !Thank you !

Q & AQ & A