Breaking the Kubernetes Kill Chain: Host Path Mount
Sect4 1
1. SECTION 4.1
THE VECTOR SPACE R3
Here the fundamental concepts of vectors, linear independence, and vector spaces are introduced in
the context of the familiar 2-dimensional coordinate plane R2 and 3-space R3. The concept of a
subspace of a vector space is illustrated, the proper nontrivial subspaces of R3 being simply lines
and planes through the origin.
1. a − b = (2, 5, −4) − (1, −2, −3) = (1, 7, −1) = 51
2a + b = 2(2,5, −4) + (1, −2, −3) = (4,10, −8) + (1, −2, −3) = (5,8, −11)
3a − 4b = 3(2,5, −4) − 4(1, −2, −3) = (6,15, −12) − (4, −8, −12) = (2, 23, 0)
2. a − b = (−1, 0, 2) − (3, 4, −5) = (−4, −4, 7) = 81 = 9
2a + b = 2(−1, 0, 2) + (3, 4, −5) = (−2, 0, 4) + (3, 4, −5) = (1, 4, −1)
3a − 4b = 3(−1, 0, 2) − 4(3, 4, −5) = (−3, 0, 6) − (12,16, −20) = (−15, −16, 26)
3. a − b = (2i − 3j + 5k ) − (5i + 3 j − 7k ) = −3i − 6 j + 12k = 189 = 3 21
2a + b = 2(2i − 3j + 5k ) + (5i + 3 j − 7k )
= (4i − 6 j + 10k ) + (5i + 3j − 7k ) = 9i − 3j + 3k
3a − 4b = 3(2i − 3j + 5k ) − 4(5i + 3 j − 7k )
= (6i − 9 j + 15k ) − (20i + 12 j − 28k ) = − 14i − 21j + 43k
4. a − b = (2i − j) − ( j − 3k ) = 2i − 2 j + 3k = 17
2a + b = 2(2i − j) + ( j − 3k ) = (4i − 2 j) + ( j − 3k ) = 4i − j − 3k
3a − 4b = 3(2i − j) − 4( j − 3k ) = (6i − 3 j) − (4 j − 12k ) = 6i − 7 j + 12k
5. v = 3
2 u, so the vectors u and v are linearly dependent.
6. au + bv = a(0, 2) + b(3, 0) = (3b, 2a ) = 0 implies a = b = 0, so the vectors u and v
are linearly independent.
7. au + bv = a(2, 2) + b(2, −2) = (2a + 2b, 2a − 2b) = 0 implies a = b = 0, so the vectors
u and v are linearly independent.
8. v = − u, so the vectors u and v are linearly dependent.
In each of Problems 9-14, we set up and solve (as in Example 2 of this section) the system
2. u v1 a w1
au + bv = 1 b = w = w
u 2 v2 2
to find the coefficient values a and b such that w = au + bv,
1 −1 a 1
9. −2 3 b = 0 ⇒ a = 3, b = 2 so w = 3u + 2 v
3 2 a 0
10. 4 3 b = −1 ⇒ a = 2, b = −3 so w = 2u − 3v
5 2 a 1
11. 7 3 b = 1 ⇒ a = 1, b = −2 so w = u − 2 v
4 −2 a 2
12. 1 −1 b = −2 ⇒ a = 3, b = 5 so w = 3u + 5 v
7 3 a 5
13. 5 4 b = −2 ⇒ a = 2, b = −2 so w = 2u − 3v
5 −6 a 5
14. −2 4 b = 6 ⇒ a = 7, b = 5 so w = 7u + 5 v
In Problems 15-18, we calculate the determinant u v w so as to determine (using Theorem
4) whether the three vectors u, v, and w are linearly dependent (det = 0) or linearly
independent (det ≠ 0).
3 5 8
15. −1 4 3 = 0 so the three vectors are linearly dependent.
2 −6 −4
5 2 4
16. −2 −3 5 = 0 so the three vectors are linearly dependent.
4 5 −7
1 3 1
17. −1 0 −2 = − 5 ≠ 0 so the three vectors are linearly independent.
2 1 2
3. 1 4 3
18. 1 3 −2 = 9 ≠ 0 so the three vectors are linearly independent.
0 1 −4
In Problems 19-24, we attempt to solve the homogeneous system Ax = 0 by reducing the
coefficient matrix A = [u v w ] to echelon form E. If we find that the system has only the
trivial solution a = b = c = 0, this means that the vectors u, v, and w are linearly independent.
Otherwise, a nontrivial solution x = [a b c ] ≠ 0 provides us with a nontrivial linear
T
combination au + bv + cw ≠ 0 that shows the three vectors are linearly dependent.
2 −3 0 1 0 −3
19. 0 1 −2 → 0 1 −2 = E
A =
1 −1 −1
0 0 0
The nontrivial solution a = 3, b = 2, c = 1 gives 3u + 2v + w = 0, so the three vectors
are linearly dependent.
5 2 4 1 0 2
20. 5 3 1 → 0 1 −3 = E
A =
4 1 5
0 0 0
The nontrivial solution a = –2, b = 3, c = 1 gives –2u + 3v + w = 0, so the three
vectors are linearly dependent.
1 −2 3 1 0 11
21. 1 −1 7 → 0 1 4 = E
A =
−2 6 2
0 0 0
The nontrivial solution a = 11, b = 4, c = –1 gives 11u + 4v – w = 0, so the three
vectors are linearly dependent.
1 5 0 1 0 0
22. 1 1 1 → 0 1 0 = E
A =
0 3 2
0 0 1
The system Ax = 0 has only the trivial solution a = b = c = 0, so the vectors u, v, and
w are linearly independent.
2 5 2 1 0 0
23. 0 4 −1 → 0 1 0 = E
A =
3 −2 1
0 0 1
4. The system Ax = 0 has only the trivial solution a = b = c = 0, so the vectors u, v, and
w are linearly independent.
1 4 −3 1 0 0
24. 4 2 3 → 0 1 0 = E
A =
5 5 −1
0 0 1
The system Ax = 0 has only the trivial solution a = b = c = 0, so the vectors u, v, and
w are linearly independent.
In Problems 25-28, we solve the nonhomogeneous system Ax = t by reducing the augmented
coefficient matrix A = [u v w t ] to echelon form E. The solution vector
x = [a b c ] appears as the final column of E, and provides us with the desired linear
T
combination t = au + bv + cw.
1 3 1 2 1 0 0 2
25. −2 0 −1 −7 → 0 1 0 −1 = E
A =
2 1 2 9
0 0 1 3
Thus a = 2, b = –1, c = 3 so t = 2u – v + 3w.
5 1 5 5 1 0 0 1
26. 2 5 −3 30 → 0 1 0 5 = E
A =
−2 −3 4 −21
0 0 1 −1
Thus a = 1, b = 5, c = –1 so t = u + 5v – w.
1 −1 4 0 1 0 0 2
27. 4 −2 4 0 → 0 1 0 6 = E
A =
3 2 1 19
0 0 1 1
Thus a = 2, b = 6, c = 1 so t = 2u + 6v + w.
2 4 1 7 1 0 0 1
28. 5 1 1 7 → 0 1 0 1 = E
A =
3 −1 5 7
0 0 1 1
Thus a = 1, b = 1, c = 1 so t = u + v + w.
29. Given vectors (0, y , z ) and (0, v, w) in V, we see that their sum (0, y + v, z + w) and the
scalar multiple c(0, y, z ) = (0, cy, cz ) both have first component 0, and therefore are
elements of V.
5. 30. If ( x, y , z ) and (u , v, w) are in V, then
( x + v ) + ( y + u ) + ( z + w) = ( x + y + z ) + (u + v + w) = 0 + 0 = 0,
so their sum ( x + u , y + v, z + w) is in V. Similarly,
cx + cy + cz = c( x + y + x) = c(0) = 0,
so the scalar multiple (cx, cy, cz ) is in V.
31. If ( x, y , z ) and (u , v, w) are in V, then
2( x + u ) = (2 x) + (2u ) = (3 y ) + (3v) = 3( y + v),
so their sum ( x + u, y + v, z + w) is in V. Similarly,
2(cx) = c(2 x) = c(3 y ) = 3(cy ),
so the scalar multiple (cx, cy, cz ) is in V.
32. If ( x, y , z ) and (u , v, w) are in V, then
z + w = (2 x + 3 y ) + (2u + 3v) = 2( x + u ) + 3( y + v),
so their sum ( x + u , y + v, z + w) is in V. Similarly,
cz = c(2 x + 3 y ) = 2(cx) + 3(cy ),
so the scalar multiple (cx, cy, cz ) is in V.
33. (0,1, 0) is in V but the sum (0,1, 0) + (0,1, 0) = (0, 2, 0) is not in V; thus V is not
closed under addition. Alternatively, 2(0,1, 0) = (0, 2, 0) is not in V, so V is not
closed under multiplication by scalars.
34. (1,1,1) is in V, but
2(1,1,1) = (1,1,1) + (1,1,1) = (2, 2, 2)
is not, so V is closed neither under addition of vectors nor under multiplication by
scalars.
6. 35. Evidently V is closed under addition of vectors. However, (0, 0,1) is in V but
(−1)(0, 0,1) = (0, 0, −1) is not, so V is not closed under multiplication by scalars.
36. (1,1,1) is in V, but
2(1,1,1) = (1,1,1) + (1,1,1) = (2, 2, 2)
is not, so V is closed neither under addition of vectors nor under multiplication by
scalars.
37. Pick a fixed element u in the (nonempty) vector space V. Then, with c = 0, the scalar
multiple cu = 0u = 0 must be in V. Thus V necessarily contains the zero vector 0.
38. Suppose u and v are vectors in the subspace V of R3 and a and b are scalars. Then
au and bv are in V because V is closed under multiplication by scalars. But then it
follows that the linear combination au + bv is in V because V is closed under addition
of vectors.
39. It suffices to show that every vector v in V is a scalar multiple of the given nonzero
vector u in V. If u and v were linearly independent, then — as illustrated in Example
2 of this section — every vector in R2 could be expressed as a linear combination of u
and v. In this case it would follow that V is all of R2 (since, by Problem 38, V is closed
under taking linear combinations). But we are given that V is a proper subspace of R2,
so we must conclude that u and v are linearly dependent vectors. Since u ≠ 0, it
follows that the arbitrary vector v in V is a scalar multiple of u, and thus V is
precisely the set of all scalar multiples of u. In geometric language, the subspace V is
then the straight line through the origin determined by the nonzero vector u.
40. Since the vectors u, v, w are linearly dependent , there exist scalars p, q, r not all zero
such that pu + qv + rw = 0. If r = 0, then p and q are scalars not both zero such that
pu + qv = 0. But this contradicts the given fact that u and v are linearly independent.
Hence r ≠ 0, so we can solve for
p q
w = − u − v = au + b v ,
r r
thereby expressing w as a linear combination of u and v.
41. If the vectors u and v are in the intersection V of the subspaces V1 and V2, then
their sum u + v is in V1 because both vectors are in V1, and u + v is in V2 because
both are in V2. Therefore u + v is in V, and thus V is closed under addition of
vectors. Similarly, the intersection V is closed under multiplication by scalars, and is
therefore itself a subspace.