A Deep Learning-based approach to VM behavior identification in - - PowerPoint PPT Presentation

▶
a deep learning based approach to vm behavior
SMART_READER_LITE
LIVE PREVIEW

A Deep Learning-based approach to VM behavior identification in - - PowerPoint PPT Presentation

Jan 24, 2024 20 likes •184 views

A Deep Learning-based approach to VM behavior identification in cloud systems Matteo Stefanini, Riccardo Lancellotti, Lorenzo Baraldi, Simone Calderara University of Modena and Reggio Emilia Department of engineering Enzo Ferrari CLOSER

slide-1
SLIDE 1

CLOSER 2019, May., 2-4, Heraklion, Greece 1

A Deep Learning-based approach to VM behavior identification in cloud systems

Matteo Stefanini, Riccardo Lancellotti, Lorenzo Baraldi, Simone Calderara University of Modena and Reggio Emilia Department of engineering “Enzo Ferrari”

slide-2
SLIDE 2

CLOSER 2019, May., 2-4, Heraklion, Greece 2

C l

  • u

d C

  • mp

u t i n g C h a l l e n g e s

  • C

r i t i c a l

  • p

e r a t i

  • n

s i n C l

  • u

d d a t a c e n t e r s

– M

  • n

i t

  • r

i n g (

  • v

e r l

  • a

d e d / u n d e r u t i l i z e d V M s a n d H

  • s

t s )

– M

a n a g e m e n t ( h u g e b i n p a c k i n g p r

  • b

l e m )

  • M

a i n c h a l l e n g e : S c a l a b i l i t y

– V

  • l

u m e

  • f

d a t a f

  • r

m

  • n

i t

  • r

i n g

– S

i z e ( a n d d i m e n s i

  • n

a l i t y )

  • f
  • p

t i m i z a t i

  • n

p r

  • b

l e m

  • C

u r r e n t s

  • l

u t i

  • n

– O

v e r s i m p l i fj c a t i

  • n
  • f

t h e p r

  • b

l e m

slide-3
SLIDE 3

CLOSER 2019, May., 2-4, Heraklion, Greece 3

I d e n t i fj c a t i

  • n
  • f

V M s

  • A

l t e r n a t i v e a p p r

  • a

c h :

– E

x p l

  • i

t s i m i l a r i t y i n V M s : ( c l a s s e s , n

  • t

i n s t a n c e s )

– R

e d u c e d p r

  • b

l e m s i z e ( l e s s d a t a , l e s s V M s )

  • P

r

  • b

l e m : h

  • w

t

  • c

l a s s i f y V M s ?

– F

a s t a n d a c c u r a t e c l a s s i fj c a t i

  • n

VM VM VM VM VM VM VM VM VM VM VM VM VM VM VM VM

CL1 CL2

slide-4
SLIDE 4

CLOSER 2019, May., 2-4, Heraklion, Greece 4

S t a t e

  • f

t h e a r t

  • T

r a d e

  • fg

a c c u r a c y / s p e e d

– F

a s t c l a s s i fj c a t i

  • n

i s n

  • t

a c c u r a t e

– A

c c u r a t e c l a s s i fj c a t i

  • n

t a k e s t i m e

– C

a n n

  • t

b e a p p l i e d t

  • n
  • d

e m a n d V M s i n p u b l i c C l

  • u

d

  • A

d a p t i v e G r a y A r e a T E c h n i q u e ( A G A T E )

– A

d d a c

  • n

fj d e n c e v a l u e t

  • c

l a s s i fj c a t i

  • n

– F

a s t a n d a c c u r a t e c l a s s i fj c a t i

  • n
  • f

s

  • m

e V M s

– S

t i l l u n s a t i s f a c t

  • r

y → P r

  • p
  • s

a l

  • f

a d i fg e r e n t a p p r

  • a

c h

slide-5
SLIDE 5

CLOSER 2019, May., 2-4, Heraklion, Greece 5

D e e p L e a r n i n g mo d e l

  • I

n p u t : t i m e s e r i e s

  • f

W s a m p l e s

  • f

s e v e r a l V M s m e t r i c s

  • O

u t p u t : c l a s s b e l

  • n

g i n g p r

  • b

a b i l i t i e s

  • M

u l t i p l e l a y e r s ( n u m b e r d e p e n d i n g

  • n

t h e i n p u t s i z e )

  • T

w

  • m
  • d

e l s :

– D

e e p C

  • n

v : b a s e d

  • n

c

  • n

v

  • l

u t i v e n e t w

  • r

k s F

  • c

u s

  • n

p a t t e r n s b e t w e e n s a m p l e s

– D

e e p F F T : b a s e d

  • n

F a s t F

  • u

r i e r T r a n s f

  • r

m a t i

  • n

F

  • c

u s

  • n

s p e c t r a l d

  • m

a i n ( n

  • v

e l D e e p L e a r n i n g a p p r

  • a

c h )

slide-6
SLIDE 6

CLOSER 2019, May., 2-4, Heraklion, Greece 6

D e e p L e a r n i n g mo d e l

  • G

e n e r a l s t r u c t u r e :

– I

n p u t l a y e r ( p r e

  • p

r

  • c

e s s i n g

  • f

s a m p l e s )

– P

r

  • c

e s s i n g b l

  • c

k s ( m u l t i p l e l a y e r s )

– F

u l l y c

  • n

n e c t e d l a y e r ( a n d s

  • f

t m a x c l a s s i fj e r )

  • D

e e p C

  • n

v :

– S

t a n d a r d m

  • d

e l

  • D

e e p F F T :

– P

e r f

  • r

m s F F T i n i n p u t l a y e r

slide-7
SLIDE 7

CLOSER 2019, May., 2-4, Heraklion, Greece 7

M

  • d

e l r e p r e s e n t a t i

  • n

Input metrics (channels)

Class probabilities

Block 1 Block 2 Block 3 Block 4 Fully Connected layer (data flattened) Time OR Frequency

Softmax

slide-8
SLIDE 8

CLOSER 2019, May., 2-4, Heraklion, Greece 8

P r

  • c

e s s i n g b l

  • c

k

  • E

a c h p r

  • c

e s s i n g b l

  • c

k c

  • n

t a i n s

– A

c t i v a t i

  • n

f u n c t i

  • n

( R e L U )

– B

a t c h N

  • r

m a l i z a t i

  • n

– 1

  • D

i m e n s i

  • n

a l c

  • n

v

  • l

u t i

  • n
  • E

a c h b l

  • c

k :

– R

e d u c e s b y 2 t h e i n p u t s i z e ( s t r i d e = 2 )

– D

  • u

b l e s t h e n u m b e r

  • f

c h a n n e l s

  • N

u m b e r

  • f

b l

  • c

k s :

slide-9
SLIDE 9

CLOSER 2019, May., 2-4, Heraklion, Greece 9

I mp l e me n t a t i

  • n

d e t a i l s

  • I

m p l e m e n t a t i

  • n

b a s e d

  • n

P y t

  • r

c h

– I

n

  • h
  • u

s e i m p l e m e n t a t i

  • n
  • f

F F T

  • S
  • u

r c e c

  • d

e a v a i l a b l e

– C

  • d

e i n g i t r e p

  • s

i t

  • r

y

– S

e e p a p e r f

  • r

d e t a i l s

  • D

e p l

  • y

m e n t

  • n

C I N E C A d a t a c e n t e r

slide-10
SLIDE 10

CLOSER 2019, May., 2-4, Heraklion, Greece 10

E x p e r i me n t a l s e t u p

  • D

a t a f r

  • m

a r e a l d a t a c e n t e r ( e

  • h

e a l t h a p p )

  • T

w

  • c

l a s s e s

  • f

V M s :

– We

b s e r v e r s

– D

B M S

  • T

r a c e s d i v i d e d i n c h u n k s w i t h d i fg e r e n t w i n d

  • w

– 1

s a m p l e e v e r y 5 m i n

– 4

s a m p l e s ( 2 m i n s ) → 2 5 6 s a m p l e s ( 2 1 h r s )

slide-11
SLIDE 11

CLOSER 2019, May., 2-4, Heraklion, Greece 11

E x p e r i me n t a l s e t u p

  • 1

6 m e t r i c s ( v i r t u a l i z e d H W / g u e s t O S )

slide-12
SLIDE 12

CLOSER 2019, May., 2-4, Heraklion, Greece 12

D e e p L e a r n i n g p e r f

  • r

ma n c e

slide-13
SLIDE 13

CLOSER 2019, May., 2-4, Heraklion, Greece 13

C

  • mp

a r i s

  • n

w i t h A G A T E

slide-14
SLIDE 14

CLOSER 2019, May., 2-4, Heraklion, Greece 14

C

  • n

c l u d i n g r e ma r k s

  • C

h a l l e n g e : s c a l a b i l i t y

  • f

m

  • n

i t

  • r

i n g / m a n a g e m e n t i n C l

  • u

d d a t a c e n t e r s → V M s i d e n t i fj c a t i

  • n
  • C
  • m

p l e x t

  • a

c h i e v e f a s t a n d a c c u r a t e i d e n t i fj c a t i

  • n
  • P

r

  • p
  • s

a l

  • f

a D e e p L e a r n i n g

  • b

a s e d a p p r

  • a

c h

  • O

u t p e r f

  • r

m s s t a t e

  • f

t h e a r t ( A G A T E )

  • S

u i t a b l e a l s

  • f
  • r
  • n
  • d

e m a n d V M s

slide-15
SLIDE 15

CLOSER 2019, May., 2-4, Heraklion, Greece 15

F u t u r e r e s e a r c h d i r e c t i

  • n

s

  • T

h

  • r
  • u

g h e v a l u a t i

  • n

i n c a s e s w i t h l i m i t e d / l

  • w
  • q

u a l i t y d a t a

  • I

d e n t i fj c a t i

  • n
  • f

n e w c l a s s e s :

– A

u t

  • e

n c

  • d

e r s / t r i g g e r s i n N N

– I

n t e g r a t i

  • n

w i t h A G A T E

  • G

e n e r a t i v e A d v e r s a r i a l N e t w

  • r

k f

  • r

w

  • r

k l

  • a

d g e n e r a t i

  • n
slide-16
SLIDE 16

CLOSER 2019, May., 2-4, Heraklion, Greece 16

A Deep Learning-based approach to VM behavior identification in cloud systems

Matteo Stefanini, Riccardo Lancellotti, Lorenzo Baraldi, Simone Calderara University of Modena and Reggio Emilia Department of engineering “Enzo Ferrari”