Databases Reference
In-Depth Information
Table 1.
Modeling A Core Cuboid
Q
1
Q
2
Q
3
Q
4
ALL
Bob
x
1
x
2
x
3
8500
Alice
x
4
x
5
10000
Jim
x
6
x
7
6100
Mallory
x
8
x
9
12400
ALL
10000 6000 11000 9000 36000
Table 2.
Modeling the Aggregation Cuboids
⎛
⎝
⎞
⎠
⎛
⎝
⎞
⎠
⎛
⎝
⎞
⎠
111000000
000110000
000001100
000000011
100001010
010100000
001010000
000000101
111111111
x
1
x
2
x
3
x
4
x
5
x
6
x
7
x
8
x
9
8500
10000
6100
12400
10000
6000
11000
9000
36000
×
=
in Table 3. From
M
rref
it can be observed that the system of linear equations
in Table 2 has infinitely many solutions. This means that an adversary cannot
infer the
entire
core cuboid from the given aggregation cuboids. However, the
first row vector of
M
rref
being a unit vector (that is, it has a single 1) indi-
cates that the value of
x
1
must remain the same among all the solutions to the
system of equations. Consequently, the adversary can infer Bob's commission
in
Q
1
.
Table 3.
The Reduced Row Echelon Form
M
rref
⎛
⎝
⎞
⎠
1000 0 000 0
0100
−
1000 0
0010 1 000 0
0001 1 000 0
0000 0 100
−
1
0000 0 010 1
0000 0 001 1
0000 0 000 0
0000 0 000 0
The existence of at least one unit row vectors in
M
rref
is indeed the neces-
sary and sucient condition for any unknown variable to have the same value