8/2/2019 Ppt 1st Review_2
1/17
8/2/2019 Ppt 1st Review_2
2/17
Under The Guidance OfMr.J.Britto Dennis M-TechLecturer - IT
By
P.Karthikeyan
A.KarthikJ.Ashok kumarA.Arul vimal Antony
Final ITDSEC.
8/2/2019 Ppt 1st Review_2
3/17
A data distributor has given sensitive data toa set of supposedly trusted agents but some of thedata is leaked and found in an unauthorized place.Our goal is to detect when the distributors sensitivedata has been leaked by agents, and to identify theagent who leaked data. We propose data allocation
strategies and implement algorithms with fakeobjects.
8/2/2019 Ppt 1st Review_2
4/17
Data sharing is more important in business.
sometimes sensitive data must be handed over to
trusted third parties. A data distributor give their data to agents. The data
may be leaked by agents. The distributor must assess the likelihood that the
leaked data came from one or more agents, as opposedto having been independently gathered by othermeans.
8/2/2019 Ppt 1st Review_2
5/17
WATERMARKING:-
Unique code is embedded in each distributedcopy.
If that copy is later discovered in the hands of
an unauthorized party, the leaker can beidentified . A watermarking is mainly used for
embedding or hiding the information.
8/2/2019 Ppt 1st Review_2
6/17
Some modification of the original data.
Watermarks can sometimes be destroyed if
the data recipient is malicious.
If the watermarked copy is found at some
unauthorized site ,then distributor can claim
his ownership.
8/2/2019 Ppt 1st Review_2
7/17
Data allocation strategies:
Two algorithms are present in this allocation
strategies.
Sample data request
Explicit data request
Fake objects:
Does not correspond to real entities but appearrealistic to the agents.
Fake objects are created by data distributor and it
consist of set of objects, subset of fake objects and
conditions.
8/2/2019 Ppt 1st Review_2
8/17
Complex to identify the fake objects .
Difficult to remove fake objects from distributed
objects by agents.
Avoid sending more data.
8/2/2019 Ppt 1st Review_2
9/17
OPTI
MIZA
TION
AGENT 2
AGENT 3
DATA
DISTR
IBUT
OR
AGENT 1
ALGORITHM
IDENTIFYING THE
PROBABILITY OF
LEAKAGE
DETECT THE
GUILTY AGENT
LEAKAGEAGENT 4
SAMPLE DATA REQUESTEXPLICIT DATA REQUEST
ORIGINAL
DATA+FAKE
OBJECTS
LEAKAGE
DATA OBJECT
1
2
3
4
5
6
8/2/2019 Ppt 1st Review_2
10/17
8/2/2019 Ppt 1st Review_2
11/17
Data Distributor Module
Optimization Module
Fake Object Module
Data Allocation Module
8/2/2019 Ppt 1st Review_2
12/17
A data distributor has given sensitive data to
trusted agents. Some of the data is leaked and found in an
unauthorized place. The distributor mustassess the likelihood that the leaked data came
from one or more agents, as opposed to havingbeen independently gathered by other means.
8/2/2019 Ppt 1st Review_2
13/17
The Optimization Module is the distributors data
allocation to agents has one constraint and one
objective. The distributors constraint is to satisfy agents
requests, by providing them with the number ofobjects .
Objective is to be able to detect an agent wholeaks any portion of his data.
8/2/2019 Ppt 1st Review_2
14/17
The distributor may be able to add fake objects
to the distributed data in order to improve hiseffectiveness in detecting guilty agents. Use of fake objects is inspired by the use of
trace records in mailing lists. Function : CreateFakeObject(Ri,Fi,condi).
8/2/2019 Ppt 1st Review_2
15/17
How can the distributor intelligently give
data to agents in order to improve the chancesof detecting a guilty agent. Allocation based on sample data request and
explicit data request.
8/2/2019 Ppt 1st Review_2
16/17
We are not using access control policies
because it allows only authorized users toaccess the sensitive data. However , these
polices are restrictive and may make it
impossible to satisfyagents requests.
8/2/2019 Ppt 1st Review_2
17/17
P. Papadimitriou and H. Garcia-Molina, Dataleakage detection, IEEE Transactions onKnowledge and Data Engineering.
S. Czerwinski, R. Fromm, and T. Hodes. Digitalmusic distribution and audiowatermarking.
L. Sweeney. Achieving k-anonymity privacy
protection using generalization and suppression.