Vous êtes sur la page 1sur 14

A Retrievable Data Perturbation Method

used in Privacy-preserving in Cloud


Computing

Radha Krishna
1CS07IT011
Guided By- Anjan Kumar
Introduction
• What is Cloud Computing?
• Advantages of Cloud computing.

• Methods used to secure the information:


 Encryption
 K-Anonymity

• Disadvantages of the cryptographic methods.


• How can it be overcome by Privacy preserving data perturbation method in cloud
computing?

Dept. of IT, BMSIT 2019 2


Introduction Cont..d
• A method is proposed to solve the below two problems:
 The data owners data should be protected from the cloud service provider.
 The authorized users should be able to access the data based on their level of access.

• Advantages of Privacy preservation using data perturbation method over other cryptographic
methods.

• Applications of Privacy preservation in cloud computing:


 Govt. agencies share very sensitive information which should not be accessible by the public
or the employees at the lowest level.
 The online software licensing uses privacy preservation method to protect their products.

Dept. of IT, BMSIT 2019 3


Process Flow:
• The data flow from Data Owners -> Cloud servers-> third party auditors.
• The data flow from Data Owners -> Third Part Auditors like Research Institutes for conducting
experiments.

Dept. of IT, BMSIT 2019 4


Retrievable General Additive Perturbation
Method(RGADP)
This method comes in four steps:

• Noise is generated using a Random Generator.


• Perturbation Algorithm
• Retrieval method
• Using the key, original data is retrieved back successfully.

Dept. of IT, BMSIT 2019 5


System Model

• This method involves adding noise to the original data.


• The mean and covariance of the original data should be equal to the perturbed data.

µx = µy and ∑x = ∑y

Example of a hospital patients list

Name Gender Age Medical History Consulting Treatment Cost


Doctor
Leela Female 34 Diabetes Dr. Rita $230,000
Shweta Female 21 Heart Disease Dr.Falsy $30,000
Anushri Female 23 Kidney Transplantation Dr.Radha $34,000
Kiran Male 45 Liver Failure Dr. Rita $23,000
Pragya Female 43 Constipation Dr.Falsy $89,000
Paul Male 22 Asthama Attack Dr.Radha $78,000

Dept. of IT, BMSIT 2019 6


Perturbation Algorithm
• Noise is generated using a random generator.
• Then this noise is added to the original data and is outsourced to cloud.
• Using a suitable key, authorized user is able to retrieve the data back from the perturbed
data.

Perturbation Algorithm:
INPUT : X,S(Confidential and non-confidential data)
Output :Y( Perturbed data)
Generate noise € by using the random generator
µx = µy and ∑x = ∑y
Add the noise € to the input matrix X such that,
Y <- X + €
Outsource the data Y to the cloud.

Dept. of IT, BMSIT 2019 7


Privacy Preserving Scheme based on RGADP

• How privacy of the data owner’s data is preserved using RGADP method.
• Comparision of RGADP with other methods like GADP and EGADP.

Dept. of IT, BMSIT 2019 8


Performance and Security

• Time complexity of perturbation is independent of number of groups.


• Graph shows the running time of perturbation and retrieval process for different number of
groups.
• Security is preserved since without the key, unauthorized users will not be able to access the
data.

Dept. of IT, BMSIT 2019 9


Experiments
• Experiments conducted taking into account all the three methods.
• Table shows the Min, Max and the Average errors between the retrieval data and original
data.

Dept. of IT, BMSIT 2019 10


Conclusion

• Proposed a method to affectively preserve data of the data owner from cloud service
provider.
• Main advantage is that the data can be retrieved back faster as compared to other methods.

• Few shortcomings:
 The datasets should be order sensitive.
 Large number of keys are used, hence key management should be affective.

• Research will be continued to overcome the above drawbacks.

Dept. of IT, BMSIT 2019 11


References
• [1] Mell P, Grance T. The NIST definition of cloud computing [EB/OL].
http://csrc.nist.gov/publications/nistpubs/800-145/SP80 0-145.pdf, Sept.2011.
• [2] RIVEST RL, ADLEMAN L, DERTOUZOS ML. On data banks and privacy homomorphisms
[A].DeMillo RA. Foundations of Secure Computation [C]. NY, Academic, 1978: 169-180.
• [3] HUANG Ruiwei, GUI Xiaolin, YU Si, et al. Privacy-preserving computable encryption
scheme of cloud computing[J]. Chinese journal of computers 2011, 34(12): 2391-2402.
• [4] LI Jin, WANG Qian, WANG Cong, et al. Fuzzy Keyword Search over Encrypted Data in Cloud
Computing[C]//Infocom 2010. San Diego, CA,USA, IEEE press, 2010: 1-5.
• [5] DUNCAN GT, LAMBERT D. The risk of disclosure for microdata[J]. J. Bus. Econom. Statist.
1989,7(2): 201–217.

Dept. of IT, BMSIT 2019 12


Questions

Dept. of IT, BMSIT 2019 13


Dept. of IT, BMSIT 2019 14

Vous aimerez peut-être aussi