Watermarking Techniques for Relational Databases: Agrawal et al. and Sion et al., Lecture notes of Statistics

Download Watermarking Techniques for Relational Databases: Agrawal et al. and Sion et al. and more Lecture notes Statistics in PDF only on Docsity!

Watermarking

Relational Databases

Acknowledgement: Mohamed Shehab from Purdue Univ.

Outline

n Introductory Material

n General Watermarking Model & Attacks

n WM Technique 1 (Agrawal et al.)

n WM Technique 2 (Sion et al.)

n Future Challenges and References

What is Watermarking? (Cont.)

n Robust Watermark: for proof of ownership,

copyrights protection.

n Fragile Watermark: for tamper proofing, data

integrity.

Robust Fragile

Watermark

Why Watermarking?

n Digital Media (Video, Audio, Images, Text) are

easily copied and easily distributed via the web.

n Database outsourcing is a common practice:

¨ Stock market data ¨ Consumer Behavior data (Walmart) ¨ Power Consumption data ¨ Weather data

n Effective means for proof of authorship.

¨ Signature and data are the same object.

n Effective means of tamper proofing.

¨ Integrity information is embedded in the data.

What defines the usability constraints? n Usability constraints are application dependent.

¨ Alterations performed by the watermark

embedding should be unidentifiable by the

human visual system in images/video.

¨ For consumer behavior data: watermarking

should preserve periodicity properties of the

data.

Courtesy of http://maps.google.com^8

What defines the usability constraints? (Cont.)

Outline

n Introductory Material

n General Watermarking Model & Attacks

n WM Technique 1 (Agrawal et al.)

n WM Technique 2 (Sion et al.)

n Future Challenges and References

Watermarking Model

Attacker Channel Watermark Decoder Watermark Encoder Secret Key, Ks Data, D Watermarked Data, DW Attacked Data, D’W Watermark W=(100100100….) Decoded Watermark WD=(100100100….)

Attacker Model

n Attacker has access to only the

watermarked data set.

n The attacker’s goal is to weaken or even

erase the embedded watermark and at the

same time keep the data usable.

“Attacker’s Dilemma”

n Possible Attacks

¨ Tuple deletion

¨ Tuple alteration

¨ Tuple insertion

Outline

n Introductory Material

n General Watermarking Model & Attacks

n WM Technique 1 (Agrawal et al.)

n WM Technique 2 (Sion et al.)

n Future Challenges and References

WM Technique 1: Encoder

Attacker Channel Watermark Decoder Watermark Encoder Secret Key, Ks Data, D Watermarked Data, DW Attacked Data, D’W Watermark W=(100100100….) Decoded Watermark WD=(100100100….) Instead: Watermark is a function of the data and the secret key

WM Technique 1: Encoder

n Assumptions

¨ K, e, m and v are selected by the data owner and

are kept secret.

¨ “K” is the secret key.

¨ “e” least significant bits can be altered in a

number without affecting its usability. Example,

e=3, 101101101.1011 101

¨ “m” used for marker selection and 1/m is fraction

of tuples marked

¨ “v” is the number of attributes used in the

watermarking process.

WM Technique 1: Encoder

n For all tuples r in D

¨ MAC(r.P) = MAC(r.P) = H(K || MAC(K||r.P)

¨ if(MAC(r.P) mod m == 0) // Marker Selection

n i = (MAC(r.P) mod v // Selected Attribute n b = (MAC(r.P) mod e // Selected LSB index n if((MAC(r.P) mod 2 == 0) // MAC is even ¨ Set bit b of r.Ai n Else ¨ Clear bit b of r.Ai

WM Technique 1 : Encoder

MAC mod m 1 4 0 9 (MAC mod v= PKey Attribute 0^ Attribute 1^ ……….^ Attribute v-^1 1234 2345 3390 4455

MAC is MAC(K || MAC(K || r.P)) MAC mod e