CS573 Data Privacy and Security
description
Transcript of CS573 Data Privacy and Security
![Page 1: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/1.jpg)
CS573 Data Privacy and Security
Secure data outsourcing – Combining encryption and fragmentation
![Page 2: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/2.jpg)
Combining Fragmentation and Encryption for outsourcing
• Breaking sensitive associations between attributes
![Page 3: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/3.jpg)
Confidentiality Constraints
• Example: constraint {DoB, Zip, Illness}– Okay to release {DoB, Zip}, {Zip}, …
![Page 4: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/4.jpg)
Constraint Example
![Page 5: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/5.jpg)
Basic ideas
• Singleton constraints– encryption
• Association constraints– encryption of any one attribute– Fragmentation of the attributes
![Page 6: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/6.jpg)
Example
![Page 7: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/7.jpg)
![Page 8: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/8.jpg)
![Page 9: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/9.jpg)
Fragmentation
• Classical distributed database design problem (vertical fragmentation)
• Total number of possible fragmentations given N attributes?
• What would be an optimal fragmentation?
![Page 10: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/10.jpg)
Optimal fragmentation
• Correctly enforce constraints• Maximal visibility• Minimal fragmentation– Maximal attribute affinity
• Problem is NP hard
![Page 11: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/11.jpg)
Algorithm without confidentiality constraints
• Without confidentiality constraints - Hierarchical clustering
• With confidentiality constraints?– {s}, {n,d}, {n,z}, {n, i}, {n, p}, {d, z, i}, {d, z, p}
![Page 12: CS573 Data Privacy and Security](https://reader036.fdocuments.net/reader036/viewer/2022062520/568164d8550346895dd71a79/html5/thumbnails/12.jpg)
References
• Combining fragmentation and encryption to protect privacy in data storage, TISSEC, 2010
• Fragmentation design for efficient query execution over sensitive distributed databases, ICDCS, 2009