T
E
L
KO
M
N
I
KA
T
e
lec
om
m
u
n
icat
ion
,
Com
p
u
t
i
n
g,
E
lec
t
r
on
ics
an
d
Cont
r
ol
Vol.
18
,
No.
3
,
J
une
2020
,
pp.
12
5
2
~
12
5
8
I
S
S
N:
1693
-
6930,
a
c
c
r
e
dit
e
d
F
ir
s
t
G
r
a
de
by
Ke
me
nr
is
tekdikti
,
De
c
r
e
e
No:
21/E
/KP
T
/2018
DO
I
:
10.
12928/
T
E
L
KO
M
NI
KA
.
v18i3.
11915
1252
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
Im
p
r
ov
e
d
q
u
al
ity o
f
se
r
vi
c
e
-
b
ase
d
c
lo
u
d
s
e
r
vi
c
e
r
a
n
k
i
n
g
an
d
r
e
c
o
m
m
e
n
d
at
io
n
m
od
e
l
S
iris
h
a
P
ot
l
u
r
i,
Kat
t
a
S
u
b
b
a
Rao
D
ep
ar
t
men
t
o
f
Co
m
p
u
t
er
Sci
en
ce
a
n
d
E
n
g
i
n
eeri
n
g
,
K
o
n
eru
L
ak
s
h
ma
i
ah
E
d
u
cat
i
o
n
Fo
u
n
d
a
t
i
o
n
,
In
d
i
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Nov
27
,
201
8
R
e
vis
e
d
J
a
n
31
,
2020
Ac
c
e
pted
F
e
b
24
,
2020
O
n
e
o
f
t
h
e
o
n
g
o
i
n
g
t
ech
n
o
l
o
g
i
e
s
w
h
i
c
h
are
u
s
e
d
b
y
l
ar
g
e
n
u
mb
er
o
f
co
mp
a
n
i
e
s
an
d
u
s
ers
i
s
c
l
o
u
d
c
o
mp
u
t
i
n
g
e
n
v
i
ro
n
m
en
t
.
T
h
i
s
co
m
p
u
t
i
n
g
t
ech
n
o
l
o
g
y
h
as
p
ro
v
ed
t
h
a
t
i
t
p
r
o
v
i
d
e
s
cer
t
ai
n
l
y
a
d
i
ffere
n
t
l
ev
e
l
o
f
effi
ci
e
n
cy
,
s
ec
u
ri
t
y
,
p
ri
v
acy
,
fl
e
x
i
b
i
l
i
t
y
a
n
d
a
v
ai
l
ab
i
l
i
t
y
t
o
i
t
s
u
s
ers
.
Cl
o
u
d
co
mp
u
t
i
n
g
d
el
i
v
er
s
o
n
d
ema
n
d
s
er
v
i
c
es
t
o
t
h
e
u
s
e
rs
b
y
u
s
i
n
g
v
ar
i
o
u
s
s
erv
i
ce
-
b
a
s
ed
mo
d
el
s
.
A
l
l
t
h
es
e
mo
d
el
s
w
o
rk
o
n
u
t
i
l
i
t
y
-
b
a
s
ed
co
m
p
u
t
i
n
g
s
u
c
h
t
h
at
u
s
ers
p
ay
fo
r
t
h
e
i
r
u
s
e
d
s
erv
i
ces
.
A
l
o
n
g
w
i
t
h
t
h
e
v
ari
o
u
s
ad
v
a
n
t
a
g
es
o
f
t
h
e
c
l
o
u
d
c
o
mp
u
t
i
n
g
e
n
v
i
ro
n
men
t
,
i
t
h
a
s
i
t
s
o
w
n
l
i
m
i
t
a
t
i
o
n
s
an
d
p
r
o
b
l
ems
s
u
c
h
as
eff
i
ci
e
n
t
re
s
o
u
rce
i
d
en
t
i
f
i
cat
i
o
n
o
r
d
i
s
c
o
v
er
y
,
s
ec
u
ri
t
y
,
t
as
k
s
c
h
ed
u
l
i
n
g
,
co
mp
l
i
a
n
ce
an
d
s
u
s
t
a
i
n
a
b
i
l
i
t
y
.
A
mo
n
g
t
h
es
e
re
s
o
u
rce
i
d
e
n
t
i
fi
ca
t
i
o
n
an
d
s
c
h
ed
u
l
i
n
g
p
l
a
y
s
a
n
i
m
p
o
r
t
an
t
ro
l
e
b
eca
u
s
e
u
s
ers
al
w
a
y
s
s
u
b
mi
t
s
t
h
e
i
r
j
o
b
s
an
d
ex
p
ect
s
res
p
o
n
s
es
i
n
l
eas
t
p
o
s
s
i
b
l
e
t
i
me.
Res
earch
i
s
h
ap
p
en
i
n
g
al
l
aro
u
n
d
t
h
e
w
o
rl
d
t
o
o
p
t
i
m
i
z
e
t
h
e
res
p
o
n
s
e
t
i
me,
mak
e
s
p
a
n
s
o
as
t
o
red
u
ce
t
h
e
b
u
rd
e
n
o
n
t
h
e
cl
o
u
d
re
s
o
u
rce
s
.
In
t
h
i
s
p
ap
er,
Q
o
S
b
a
s
ed
s
erv
i
ce
ran
k
i
n
g
mo
d
el
i
s
p
r
o
p
o
s
e
d
fo
r
cl
o
u
d
c
o
mp
u
t
i
n
g
en
v
i
ro
n
men
t
t
o
fi
n
d
t
h
e
es
s
en
t
i
a
l
t
o
p
ran
k
ed
s
er
v
i
ce
s
.
Pro
p
o
s
ed
mo
d
el
i
s
i
mp
l
emen
t
ed
i
n
t
w
o
p
h
as
e
s
.
In
t
h
e
fi
rs
t
p
h
as
e,
s
i
mi
l
ar
i
t
y
co
mp
u
t
a
t
i
o
n
b
et
w
een
t
h
e
u
s
er
s
an
d
t
h
ei
r
s
erv
i
ces
i
s
c
o
n
s
i
d
ered
.
I
n
t
h
e
s
ec
o
n
d
p
h
as
e,
co
mp
u
t
i
n
g
t
h
e
mi
s
s
i
n
g
v
a
l
u
e
s
b
as
ed
o
n
t
h
e
co
mp
u
t
e
d
s
i
mi
l
ari
t
y
meas
u
re
s
i
s
cal
cu
l
at
e
d
.
T
h
e
effi
ci
e
n
cy
o
f
t
h
e
p
ro
p
o
s
ed
ran
k
i
n
g
i
s
meas
u
re
d
an
d
t
h
e
av
erag
e
p
rec
i
s
i
o
n
co
rrel
a
t
i
o
n
o
f
t
h
e
p
r
o
p
o
s
e
d
ran
k
i
n
g
meas
u
re
i
s
s
h
o
w
i
n
g
b
et
t
er
res
u
l
t
s
t
h
an
t
h
e
ex
i
s
t
i
n
g
meas
u
res
.
K
e
y
w
o
r
d
s
:
C
loud
c
omput
ing
Qua
li
ty
of
s
e
r
vice
S
e
r
vice
r
a
nking
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e
.
C
or
r
e
s
pon
din
g
A
u
th
or
:
S
ir
is
ha
P
otl
ur
i
,
De
pa
r
tm
e
nt
of
C
omput
e
r
S
c
ienc
e
a
nd
E
nginee
r
ing
,
Kone
r
u
L
a
ks
hmaia
h
E
duc
a
ti
on
F
ounda
ti
on,
Gr
e
e
n
F
ields
,
Va
dde
s
wa
r
a
m,
Andhr
a
P
r
a
de
s
h
522
502,
I
ndia
.
E
mail:
s
ir
is
ha
.
ve
gunta@
gmail.
c
om
1.
I
NT
RODU
C
T
I
ON
On
de
mand
s
e
r
vice
s
e
lec
ti
on
is
p
r
ovidi
ng
wide
s
e
lec
ti
on
of
f
a
c
il
it
ies
a
nd
s
a
ti
s
f
ying
the
c
ha
nging
ne
e
ds
of
I
T
indus
tr
ies
.
T
he
r
e
a
r
e
many
a
dva
ntage
s
a
nd
is
s
ue
s
whic
h
a
r
e
a
s
s
oc
iate
d
with
c
loud
c
o
mput
ing
e
nvir
onment
a
nd
ongoing
r
e
s
e
a
r
c
h
is
he
lpi
ng
it
im
pr
ove
in
a
ll
it
s
r
e
ga
r
ds
.
C
loud
c
omput
ing
pr
o
vides
a
ll
the
a
dva
ntage
s
to
the
c
loud
u
s
e
r
s
to
mi
gr
a
te
to
ne
xt
leve
l
of
c
omput
ing.
T
his
c
omput
ing
e
nvir
onme
nt
is
s
ti
ll
in
a
n
e
volvi
ng
pha
s
e
a
nd
de
li
ve
r
s
the
s
e
r
vic
e
s
in
va
r
ious
wa
ys
us
ing
va
r
ious
de
ploym
e
nt
models
.
T
he
r
e
lations
hip
be
twe
e
n
c
l
oud
s
e
r
vice
p
r
ovider
a
nd
c
us
tom
e
r
c
ompl
e
tely
li
e
s
on
the
on
de
mand
a
va
il
a
bil
it
y
of
the
r
e
s
our
c
e
s
,
e
f
f
icie
nt
mana
ge
ment
of
the
r
e
s
our
c
e
s
a
nd
c
ompete
nt
u
s
e
of
the
r
e
s
our
c
e
s
.
Alo
ng
with
the
va
r
ious
a
dva
ntage
s
of
the
c
loud
c
omput
ing
e
n
vir
onment,
it
ha
s
it
s
ow
n
li
mi
tations
a
nd
pr
oblems
s
uc
h
a
s
e
f
f
icie
nt
r
e
s
our
c
e
identi
f
ica
ti
on
o
r
dis
c
ove
r
y,
s
e
c
ur
it
y,
tas
k
s
c
he
duli
ng,
c
ompl
ianc
e
a
nd
s
us
tai
na
bil
it
y.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
ol
I
mpr
ov
e
d
quali
ty
of
s
e
r
v
ice
-
bas
e
d
c
loud
s
e
r
v
ice
r
a
nk
ing
and
r
e
c
omm
e
ndati
on
mode
l
(
Sir
is
ha
P
ot
lur
i
)
1253
Among
thes
e
r
e
s
our
c
e
identif
ica
ti
on
a
nd
s
c
he
duli
ng
plays
a
n
im
po
r
tant
r
ole
be
c
a
us
e
us
e
r
s
a
lwa
ys
s
ubmi
ts
their
jobs
a
nd
e
xpe
c
ts
r
e
s
pons
e
in
lea
s
t
pos
s
ibl
e
ti
me.
R
e
s
e
a
r
c
h
is
ha
ppe
ning
a
ll
a
r
ound
the
wo
r
ld
to
opti
mi
z
e
the
r
e
s
pons
e
ti
me,
make
s
pa
n
s
o
a
s
to
r
e
duc
e
the
b
ur
de
n
on
the
c
loud
r
e
s
our
c
e
s
[1
,
2]
.
2.
RE
S
E
AR
CH
I
S
S
UE
S
AN
D
CHAL
L
E
NG
E
S
I
N
CL
OUD
COM
P
UT
I
NG
2.
1.
P
r
ivacy
an
d
s
e
c
u
r
it
y
Yunc
hua
n
S
un
E
t
a
l.
s
tate
d
that
c
loud
c
omput
ing
e
nvir
onment
make
us
e
of
lar
ge
a
mount
s
o
f
da
tas
e
ts
f
or
s
tor
a
ge
a
nd
pr
oc
e
s
s
ing.
On
pr
e
mi
s
e
da
ta
s
tor
a
ge
,
ha
ndli
ng,
pr
oc
e
s
s
ing
is
c
ompl
e
tely
r
is
k
f
r
e
e
.
B
ut
if
the
da
ta
is
ge
tt
ing
s
a
ve
d
in
othe
r
p
lac
e
s
ther
e
is
a
gr
e
a
ter
r
is
k
invol
ve
d
in
that
.
C
loud
s
e
r
vice
p
r
ovider
s
,
int
e
r
media
r
ies
or
the
c
ompanie
s
invol
ve
d
in
hos
ti
ng
of
c
loud
platf
o
r
ms
may
d
is
c
los
e
the
da
ta
whic
h
is
be
ing
in
a
c
c
e
s
s
to
them.
Due
to
whic
h
da
ta
may
be
tam
pe
r
e
d
or
los
t.
I
n
e
it
he
r
o
f
the
c
a
s
e
s
it
lea
ds
to
s
e
c
ur
it
y
a
nd
pr
ivac
y
is
s
ue
s
[
3]
.
Due
to
thes
e
r
e
a
s
ons
of
s
e
c
ur
i
ty
a
nd
pr
ivac
y,
c
loud
c
omput
ing
is
taking
s
li
ghtl
y
a
ba
c
k
s
te
p
a
pa
r
t
f
r
om
it
s
wide
r
a
dopti
on
[4
,
5]
.
2.
2.
P
e
r
f
or
m
an
c
e
W
e
izhong
Qia
ng
E
t
a
l.
s
tate
d
that
in
c
loud
e
nvi
r
onment,
thi
s
is
s
ue
is
ve
r
y
much
im
por
tant
a
s
it
is
one
of
the
mea
s
ur
ing
tec
hnique
in
c
omput
ing
e
nvi
r
onment
.
P
e
r
f
o
r
manc
e
is
a
major
c
onc
e
r
n
in
c
loud
s
ys
tem
whic
h
a
f
f
e
c
ts
the
de
li
ve
r
y
o
f
va
r
ious
s
e
r
vice
s
,
r
e
v
e
n
ue
ge
ne
r
a
ted
due
to
s
e
r
vice
de
li
ve
r
y
a
nd
the
nu
mber
o
f
c
us
tom
e
r
s
invl
ovled
in
the
p
r
oc
e
s
s
[
6]
.
2.
3.
Re
li
ab
il
it
y
an
d
a
vail
a
b
il
it
y
M
oha
mm
a
d
R
e
z
a
M
e
s
ba
hi
E
t
a
l
.
s
tate
d
that
r
e
li
a
bil
it
y
de
f
ines
the
c
e
r
tain
e
xpe
c
ted
f
unc
ti
oning
of
the
s
ys
tem
unde
r
s
e
lec
ted
ti
m
e
int
e
r
va
l
a
nd
with
g
iven
c
ondit
ions
.
Ava
il
a
bil
i
ty
o
f
a
ny
s
ys
tem
c
a
n
be
de
f
ined
a
s
de
gr
e
e
in
whic
h
the
s
e
t
o
f
r
e
s
our
c
e
s
a
r
e
a
va
il
a
ble
a
nd
thes
e
a
va
il
a
ble
r
e
s
our
c
e
s
a
r
e
us
e
d
whe
ne
ve
r
a
nd
whe
r
e
ve
r
ther
e
is
a
ne
e
d
.
T
he
s
e
two
f
a
c
tor
s
c
omes
togethe
r
a
nd
de
f
ines
the
s
tr
e
ngth
a
nd
po
tenc
y
of
the
s
ys
tem
[7
]
.
2.
4.
S
c
alab
il
it
y
a
n
d
e
las
t
ici
t
y
E
manue
l
C
outi
nho
E
t
a
l.
s
tate
d
that
the
s
c
a
labili
ty
c
a
n
be
de
f
ined
a
s
it
s
a
bil
it
y
to
a
dopt
to
the
c
ha
nge
s
.
W
he
n
ther
e
is
a
ne
e
d
o
f
r
e
s
our
c
e
s
by
va
r
ious
c
loud
us
e
r
s
,
the
s
ys
tem
s
hould
take
c
a
r
e
of
wor
kloads
e
f
f
icie
ntl
y.
E
las
ti
c
it
y
o
f
the
s
ys
tem
c
a
n
be
de
f
ined
a
s
the
f
a
c
tor
up
to
whic
h
it
c
a
n
ha
ndle
the
wor
kloads
e
f
f
icie
ntl
y.
E
las
ti
c
it
y
e
ns
ur
e
s
a
ll
oc
a
ti
on
a
nd
de
a
ll
oc
a
ti
on
of
r
e
s
our
c
e
s
to
ha
ndle
the
ne
e
ds
of
on
de
mand
us
e
r
s
[
8]
.
2.
5.
I
n
t
e
r
op
e
r
ab
il
it
y
an
d
p
or
t
ab
i
li
t
y
B
e
niamino
Di
M
a
r
ti
no
E
t
a
l.
s
tate
d
that
int
e
r
ope
r
a
bil
it
y
of
c
loud
c
omput
ing
e
nvir
onment
c
a
n
be
de
f
ined
a
s
the
powe
r
or
a
bil
it
y
whic
h
is
a
dopted
by
the
s
ys
tem
to
r
un
va
r
ious
s
e
r
vice
s
whic
h
a
r
e
f
r
om
va
r
ious
ve
ndor
s
.
I
ts
a
s
s
oc
iate
d
f
a
c
tor
int
e
r
ope
r
a
bil
it
y
c
a
n
be
de
f
ined
a
s
the
wa
y
in
whic
h
dif
f
e
r
e
nt
plat
f
o
r
ms
a
r
e
e
f
f
e
c
ti
ve
ly
us
e
d
a
mong
d
if
f
e
r
e
nt
c
loud
s
e
r
vice
p
r
ovider
s
.
P
or
tabili
ty
is
a
n
im
po
r
tant
f
a
c
tor
whic
h
de
f
ines
the
a
bil
it
y
to
tr
a
ns
f
e
r
the
r
e
late
d
da
ta
a
ppli
c
a
ti
ons
f
r
om
one
C
S
P
to
a
nother
C
S
P
without
much
de
p
e
nde
nc
y
invol
ve
d
in
it
[
9
]
.
2.
6.
Re
s
ou
r
c
e
m
an
age
m
e
n
t
an
d
s
c
h
e
d
u
li
n
g
Ha
r
s
hil
M
e
hta
E
t
a
l.
s
tate
d
that
r
e
s
our
c
e
mana
ge
ment
c
a
n
be
de
f
ined
a
s
a
f
a
c
tor
whic
h
r
e
pr
e
s
e
nts
the
e
f
f
icie
nt
us
a
ge
a
nd
maintena
nc
e
o
f
the
r
e
s
our
c
e
s
.
Va
r
ious
types
of
r
e
s
our
c
e
s
a
r
e
mappe
d
to
the
s
ubmi
tt
e
d
jobs
s
o
a
s
s
c
he
dule
them
in
a
mo
r
e
e
f
f
icie
nt
m
a
nne
r
.
A
be
tt
e
r
r
e
s
our
c
e
s
c
he
duli
ng
or
pr
ovis
ion
c
a
n
be
a
c
hieve
d
by
us
ing
a
be
tt
e
r
s
c
he
duli
ng
po
li
c
y.
B
y
us
ing
thes
e
s
c
he
duli
ng
a
lgor
it
hms
,
we
c
a
n
opt
im
i
z
e
c
loud
s
e
r
vice
s
to
a
c
hieve
high
leve
l
of
qua
li
ty
o
f
s
e
r
vice
.
T
he
r
e
a
r
e
many
c
a
tegor
ies
of
us
e
r
s
,
tas
ks
a
nd
r
e
s
our
c
e
s
us
ing
whic
h
tas
k
mapping
a
nd
e
xe
c
uti
on
i
s
a
c
hieve
d
[
10]
.
2.
7.
E
n
e
r
gy
c
on
s
u
m
p
t
ion
Na
z
mul
Hos
s
a
in
E
t
a
l.
s
tate
d
that
a
ny
c
loud
da
ta
c
e
nter
or
s
tor
a
ge
c
ons
is
ts
of
inf
r
a
s
tr
uc
tur
e
in
lar
ge
a
nd
volum
inous
a
mount
s
.
C
loud
c
omput
in
g
e
nvir
onment
is
de
li
ve
r
ing
va
r
ious
types
of
s
e
r
vice
s
to
e
nor
mous
us
e
r
s
b
y
us
ing
it
s
mas
s
ive
inf
r
a
s
tr
uc
tur
e
.
C
ooli
ng
inf
r
a
s
tr
uc
tur
e
is
ne
e
de
d
to
c
ompens
a
te
the
he
a
t
whic
h
is
p
r
oduc
e
d
by
thes
e
s
e
r
ve
r
s
.
T
h
is
E
xpe
ns
ive
inf
r
a
s
tr
uc
tur
e
s
hould
be
de
ve
loped,
ope
r
a
ted
a
nd
maintaine
d
ve
r
y
c
a
r
e
f
ull
y
a
s
it
pr
oduc
e
s
gr
e
e
nh
ous
e
ga
s
e
s
whic
h
a
r
e
a
dve
r
s
e
to
e
nvir
onmenta
l
ba
lanc
e
.
P
r
ope
r
c
ooli
ng
s
ys
tem
a
nd
mec
ha
nis
ms
a
r
e
maint
a
ined
by
c
loud
c
e
nter
s
to
maintain
the
c
loud
inf
r
a
s
tr
uc
tur
e
a
s
e
nvir
onment
f
r
iendly
[
11
-
13]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
ol
,
Vol.
18
,
No.
3
,
J
une
2020:
125
2
-
12
5
8
1254
2.
8.
Vir
t
u
ali
z
at
ion
Yuping
Xing
E
t
a
l.
s
tate
d
that
in
dis
tr
ibut
e
d
a
n
d
c
loud
c
omput
ing
e
nvi
r
one
ment
,
the
c
onc
e
pt
of
vir
tualiza
ti
on
plays
a
ve
r
y
im
po
r
tant
r
ole.
T
his
r
e
f
e
r
s
to
the
pr
oc
e
s
s
of
c
r
e
a
ti
ng
vir
tual
mac
hines
f
or
the
va
r
ious
phys
ica
l
r
e
s
our
c
e
s
.
T
his
c
onc
e
pt
is
c
lo
s
e
ly
r
e
late
d
to
tas
k
s
c
he
duli
ng
p
r
oc
e
s
s
whe
r
e
tas
k
que
ue
is
mappe
d
to
va
r
ious
vi
r
tual
mac
hines
f
or
s
c
he
duli
ng.
T
he
main
a
dva
ntage
s
of
v
ir
tualiztion
a
r
e
c
os
t
r
e
duc
ti
on,
highl
y
s
c
a
lable
a
nd
pr
ovide
e
las
ti
c
be
ha
vior
to
the
c
loud
c
omput
ing
e
nvir
onment
[
14
,
15
]
.
2.
9
.
B
an
d
wid
t
h
c
os
t
P
huong
Ha
E
t
a
l
.
s
tate
d
that
a
c
c
or
ding
to
c
loud
c
omput
ing,
ba
ndwidth
r
e
f
e
r
e
s
to
the
r
a
te
in
whic
h
bit
s
a
r
e
tr
a
ns
f
e
r
r
e
d.
T
his
mea
s
ur
e
ment
is
mainly
us
e
f
ul
whe
n
we
a
r
e
c
a
lcula
ti
ng
thr
oughput
of
the
s
ys
tem
.
S
ince
c
loud
c
omput
ing
e
nvir
one
ment
is
pr
ov
idi
ng
high
s
pe
e
d
tr
a
ns
f
e
r
of
da
ta
a
mong
s
ys
tems
it
r
e
qui
r
e
s
high
ba
ndwidth
[
16]
.
3.
QUAL
I
T
Y
OF
S
E
RV
I
CE
I
N
CL
OUD
COM
P
U
T
I
NG
3.
1.
S
L
A
is
s
u
e
s
in
c
lou
d
c
om
p
u
t
in
g
e
n
vironm
e
n
t
S
e
r
vice
leve
l
a
gr
e
e
ment
c
omm
only
include
s
va
r
ious
s
e
gments
li
ke
mea
s
ur
ing
the
pe
r
f
or
manc
e
,
mana
ging
the
p
r
oblems
,
c
us
tom
e
r
s
a
ti
s
f
a
c
ti
on
,
r
e
c
ove
r
y
f
r
om
f
a
il
u
r
e
s
a
nd
f
inally
ter
mi
na
ti
on
of
the
a
gr
e
e
ment.
S
L
A’
s
a
r
e
c
ons
is
tently
c
he
c
ke
d
a
n
d
va
li
da
ted
us
ing
wh
ich
it
c
r
e
a
tes
a
be
tt
e
r
c
omm
unica
ti
on
c
ha
nne
l
be
twe
e
n
c
ons
umer
a
nd
the
p
r
ovider
.
S
L
A
e
s
tablis
he
s
the
qua
l
it
y
of
s
e
r
vice
(
QoS)
a
g
r
e
e
ment
be
twe
e
n
s
e
r
vice
-
b
a
s
e
d
s
y
s
tem
pr
ovider
s
a
nd
us
e
r
s
.
W
it
h
a
viol
a
ti
on
of
S
L
A,
the
pr
ovide
r
mus
t
pa
y
p
e
na
lt
ies
.
C
onti
nuous
p
e
r
f
or
manc
e
c
he
c
k
is
r
e
quir
e
d
in
or
de
r
to
e
ns
ur
e
that
thes
e
C
S
P
s
a
r
e
p
r
ovidi
ng
s
e
r
vice
s
a
s
mentioned
in
S
L
As
.
S
L
A
is
c
ons
is
tently
met,
thes
e
a
gr
e
e
ments
a
r
e
f
r
e
que
ntl
y
de
s
igned
wi
th
s
pe
c
if
ic
li
ne
s
o
f
de
mar
c
a
ti
on
a
nd
the
pa
r
ti
e
s
invol
ve
d
a
r
e
r
e
q
uir
e
d
to
mee
t
r
e
gular
ly
to
c
r
e
a
te
a
n
ope
n
f
o
r
um
f
or
c
omm
unica
ti
on.
C
loud
c
omput
ing
is
us
ing
c
oll
e
c
ti
on
of
r
e
s
our
c
e
s
a
nd
a
ll
thes
e
a
r
e
s
uppor
ted
by
the
i
nter
f
a
c
e
a
nd
inf
r
a
s
tr
uc
tur
e
c
onc
e
pt
of
dis
tr
ibut
e
d
a
nd
c
lou
d
c
omput
ing
e
nvir
onment
.
S
e
r
ive
leve
l
a
gr
e
e
ment
c
a
n
be
view
e
d
a
s
a
doc
ument
be
twe
e
n
c
us
tom
e
r
a
nd
pr
ovider
a
nd
is
maily
ba
s
e
d
on
s
e
r
vice
s
a
nd
not
on
c
us
tom
e
r
s
[
17]
.
3.
2.
QoS
m
e
t
r
ics
an
d
c
las
s
if
icat
ion
As
we
ha
ve
d
is
c
us
s
e
d
e
a
r
l
i
e
r
c
loud
c
omput
ing
e
nv
ir
onment
is
in
de
mand
of
c
onti
nuous
pe
r
f
or
manc
e
mea
s
ur
e
ment.
T
he
two
mec
ha
nis
ms
s
uc
h
a
s
m
e
a
s
ur
e
ment
moni
tor
a
nd
e
xpos
ur
e
of
c
loud
c
o
mput
ing
e
nvir
onment
is
done
by
us
ing
thes
e
metr
ics
.
T
he
main
is
s
ue
a
nd
dr
a
wba
c
k
wi
th
th
is
c
omput
ing
e
nvi
r
one
ment
is
it
s
pe
r
f
or
manc
e
.
T
he
pe
r
f
o
r
a
mce
is
mea
s
ur
e
d
by
us
ing
va
r
ious
f
a
c
tor
s
r
e
late
d
to
us
e
r
e
xpe
r
i
e
nc
e
a
nd
a
bil
it
y.
T
he
main
dr
a
wba
c
k
a
nd
p
r
oblem
of
c
loud
c
omput
ing
is
ha
ving
dif
f
icul
tl
y
in
identif
ying
ba
s
e
or
r
oo
t
c
a
us
e
f
or
s
e
r
vice
pr
oblems
s
uc
h
a
s
int
e
r
r
up
ti
ons
b
e
c
a
us
e
the
s
tr
uc
tur
e
of
the
e
nvir
one
ment
i
s
ve
r
y
c
ompl
e
x.
T
he
dyna
mi
c
e
nvi
r
one
ment
of
the
c
loud
make
s
us
e
of
va
r
ious
metr
ics
to
mea
s
ur
e
the
pe
r
f
o
r
manc
e
.
All
thes
e
a
tt
r
ibut
e
s
a
r
e
li
s
ted
a
s
QoS
metr
ic
to
e
nf
or
c
e
the
pr
omi
s
e
d
leve
l
of
ba
ha
viour
in
c
loud
[
18
,
19]
.
T
he
r
e
a
r
e
many
QoS
metr
ics
us
ing
whic
h
the
pe
r
f
or
manc
e
of
C
S
P
c
a
n
be
mea
s
ur
e
d.
M
a
inl
y
ther
e
a
r
e
thr
e
e
a
s
pe
c
ts
of
QoS
metr
ic
na
mely
pe
r
f
or
manc
e
,
de
pe
nda
bil
it
y
a
nd
c
onf
igur
a
ti
on
.
P
e
r
f
o
r
manc
e
c
a
n
be
mea
s
ur
e
d
by
us
ing
r
e
s
pons
e
ti
me,
pr
oc
e
s
s
ing
ti
me,
s
e
r
vice
thr
ough
put,
da
ta
t
r
a
ns
f
e
r
r
a
te
a
nd
late
nc
y.
De
pe
nda
bil
it
y
c
a
n
be
mea
s
ur
e
d
by
us
ing
a
va
il
a
bil
it
y,
e
las
ti
c
it
y,
r
e
li
a
bil
it
y,
ti
melines
s
,
r
e
s
il
ienc
e
a
nd
s
c
a
labili
ty.
S
im
il
a
r
ly
c
onf
igur
a
ti
on
c
a
n
be
mea
s
ur
e
d
by
us
ing
vi
r
tual
s
y
s
tems
a
nd
loca
ti
on.
T
he
QoS
met
r
ics
c
las
s
if
ica
ti
on
[
20]
is
given
in
T
a
ble
1.
3.
3.
E
xis
t
in
g
QoS
p
r
e
d
ic
t
ion
a
p
p
r
oac
h
e
s
an
d
m
od
e
l
s
f
or
s
e
r
vice
r
an
k
in
g
E
xis
ti
ng
QoS
pr
e
diction
a
pp
r
oa
c
he
s
a
r
e
C
S
tr
us
t,
C
OFI
L
L
,
L
oNM
F
,
OL
M
F
,
S
C
M
F
,
GN
M
F
,
T
a
S
R
a
nd
C
S
M
F
e
tc.
E
xis
ti
ng
r
a
nking
pr
e
diction
a
ppr
oa
c
he
s
a
r
e
C
loud
r
a
nk
QoS
r
a
nking
pr
e
diction
f
r
a
mew
or
k
-
C
loudR
a
nk1
a
nd
C
loudR
a
nk2,
P
S
O
ba
s
e
d
QoS
r
a
nking
pr
e
diction
a
lgor
i
thm
,
T
im
e
a
wa
r
e
tr
us
t
wor
thi
ne
s
s
r
a
nking
p
r
e
diction
a
ppr
oa
c
h
,
M
ult
i
r
a
nk1
a
nd
M
ult
iR
a
nk2
a
nd
R
a
nking
o
r
iente
d
p
r
e
diction
method
e
tc.
T
he
objec
ti
ve
of
the
pr
opos
e
d
me
c
ha
nis
m
to
a
c
hieve
be
tt
e
r
r
e
c
omm
e
nda
ti
on
c
om
pa
r
e
d
to
c
onve
nti
ona
l
a
ppr
oa
c
he
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
ol
I
mpr
ov
e
d
quali
ty
of
s
e
r
v
ice
-
bas
e
d
c
loud
s
e
r
v
ice
r
a
nk
ing
and
r
e
c
omm
e
ndati
on
mode
l
(
Sir
is
ha
P
ot
lur
i
)
1255
T
a
ble
1.
C
las
s
if
ica
ti
on
of
QoS
metr
ic
in
c
loud
c
om
puti
ng
e
nvir
onment
A
s
pe
c
t
M
e
tr
ic
D
e
s
c
r
ip
ti
on
P
e
r
f
or
ma
nc
e
R
e
s
pons
e
T
im
e
E
la
ps
e
d
ti
me
be
twe
e
n
us
e
r
s
ubmi
tt
e
d
a
ta
s
k
a
s
a
s
e
r
vi
c
e
r
e
que
s
t
a
nd
f
in
a
l
r
e
s
pons
e
i
s
r
e
c
e
iv
e
d
P
r
oc
e
s
s
in
g T
im
e
I
t
is
r
e
pr
e
s
e
nt
e
d a
s
a
v
e
r
a
ge
r
e
s
pon
s
e
t
im
e
S
e
r
vi
c
e
T
hr
oughput
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
numbe
r
of
ta
s
k
s
c
ompl
e
te
d
by
c
lo
ud
c
omput
in
g
e
nvi
r
onme
nt
pe
r
uni
t
a
mount
of
t
im
e
.
D
a
ta
T
r
a
ns
f
e
r
R
a
te
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
s
pe
e
d
a
t
w
hi
c
h
da
ta
c
a
n
be
tr
a
ns
f
e
r
r
e
d
qui
c
kl
y
a
nd
s
e
c
ur
e
ly
L
a
te
nc
y
P
a
c
ke
t
L
e
ngt
h
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
de
la
y
be
tw
e
e
n
w
he
n
f
ir
s
t
pa
c
ke
t
bi
t
pa
s
s
e
d
th
e
in
put
c
he
c
k point
a
nd t
he
l
a
s
t
p
a
c
ke
t
bi
t
pa
s
s
e
d t
he
out
put
c
he
c
k poin
t
N
e
twor
k L
a
te
nc
y
T
hr
ough of
a
C
S
P
’
s
ne
twor
k
D
e
pe
nda
bi
li
ty
A
va
il
a
bi
li
ty
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
pe
r
c
e
nt
a
ge
of
ti
me
th
a
t
th
e
r
e
s
our
c
e
is
r
e
a
dy
f
or
th
e
im
me
di
a
te
us
e
E
la
s
ti
c
it
y
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
a
bi
li
ty
by
w
hi
c
h
th
e
C
S
P
c
a
n
e
xpa
nd
a
nd
c
ont
r
a
c
t
hi
s
s
e
r
vi
c
e
s
R
e
li
a
bi
li
ty
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
a
bi
li
ty
to
e
ns
ur
e
c
ont
in
uou
s
pr
oc
e
s
s
of
th
e
pr
ogr
a
m
w
it
hout
l
os
s
T
im
e
li
ne
s
s
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
a
bi
li
ty
to
s
uppl
y
in
f
or
ma
ti
on
in
ti
me
w
he
n
us
e
r
s
a
r
e
in
ne
e
d t
o a
c
c
e
s
s
i
t
R
e
s
il
ie
nc
e
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
a
bi
li
ty
to
r
e
s
is
t
a
nd
r
e
c
ove
r
in
a
n
e
f
f
e
c
ti
ve
a
nd
ti
me
ly
ma
nne
r
,
in
c
lu
di
ng
th
e
pr
e
s
e
r
va
ti
on
a
nd
r
e
s
to
r
a
ti
on
of
it
s
e
s
s
e
nt
ia
l
ba
s
ic
s
tr
uc
tu
r
e
s
a
nd f
unc
ti
ons
S
c
a
la
bi
li
ty
H
or
iz
ont
a
l
I
t
is
r
e
pr
e
s
e
nt
e
d
a
s
th
e
a
bi
li
ty
to
in
c
r
e
a
s
e
c
lo
ud
r
e
s
our
c
e
s
of
th
e
s
a
me
ty
pe
by
us
in
g V
M
s
V
e
r
ti
c
a
l
C
onf
ig
ur
a
ti
on
V
ir
tu
a
l
S
ys
te
ms
I
t
is
r
e
pr
e
s
e
nt
e
d t
o i
de
nt
if
y c
lo
ud vir
tu
a
l
in
f
r
a
s
tr
uc
tu
r
e
L
oc
a
ti
on
B
a
s
e
d on QoS
lo
c
a
ti
on
a
f
f
in
it
y
4.
Q
OS
B
ASE
D
S
E
RV
I
CE
RA
NKI
NG
M
E
CHAN
I
S
M
QoS
is
a
mea
s
ur
e
ment
to
c
he
c
k
f
or
the
gua
r
a
ntee
d
s
e
r
vice
is
a
c
hieve
d
f
r
om
the
c
loud
p
r
ovider
s
or
not
[
21
-
24]
.
T
his
mec
ha
nis
m
c
ontains
two
ph
a
s
e
s
f
or
opti
mal
s
e
r
vice
s
e
lec
ti
on
in
c
loud
c
o
mput
ing
e
nvir
onment.
I
n
the
f
ir
s
t
pha
s
e
,
the
c
omput
a
ti
on
of
s
im
il
a
r
it
y
a
mong
the
us
e
r
s
a
nd
their
s
e
r
vice
s
is
c
ons
ider
e
d.
I
n
the
s
e
c
ond
pha
s
e
,
c
omput
a
ti
on
of
the
mi
s
s
ing
va
lues
b
a
s
e
d
on
the
va
lu
e
s
whic
h
a
r
e
us
ing
c
omput
e
d
s
im
il
a
r
it
y
mea
s
ur
e
s
is
c
a
lcula
ted.
T
he
mec
ha
nis
m
of
pr
e
diction
a
nd
r
a
nking
ba
s
e
d
on
QoS
is
va
li
da
ted
us
in
g
W
S
Dr
e
a
m#1
QoS
da
tas
e
t
a
nd
W
e
b
S
e
r
vice
QoS
da
tas
e
t
us
ing
s
tatis
ti
c
a
l
mea
s
ur
e
s
mea
n
s
qua
r
e
d
e
r
r
or
a
nd
r
oot
mea
n
s
qua
r
e
e
r
r
or
.
QoS
matr
ix
whic
h
c
ontains
QoS
va
lues
of
va
r
ious
s
e
r
vice
s
f
or
va
r
ious
us
e
r
s
a
c
ts
a
s
e
s
s
e
nti
a
l
da
ta
s
our
c
e
f
or
s
e
r
vice
e
va
luation
a
n
d
s
e
lec
ti
on.
B
ut
thi
s
is
a
s
pa
r
s
e
ma
tr
ix
with
mi
s
s
ing
va
lues
.
P
r
opos
e
d
method
is
us
e
f
ul
f
or
QoS
va
lue
pr
e
diction
a
nd
s
ubs
e
que
ntl
y
he
lps
in
opti
mal
r
a
nking
s
e
que
nc
e
of
s
e
r
vice
s
.
C
os
ine
ba
s
e
d
s
im
il
a
r
it
y
mea
s
ur
e
he
lps
f
or
QoS
va
lue
pr
e
diction
a
nd
i
mpr
ove
d
bin
a
r
y
gr
a
vit
a
ti
ona
l
s
e
a
r
c
h
is
us
e
f
ul
f
o
r
s
e
r
vice
r
a
nkingm
e
c
ha
nis
m.
S
im
il
a
r
it
y
c
omput
a
ti
on
a
nd
mi
s
s
ing
QoS
va
lue
pr
e
diction
a
r
e
pe
r
f
or
med
on
U
us
e
r
s
a
nd
S
s
e
r
vice
s
a
nd
the
output
is
opti
mal
s
e
r
vice
r
a
nking
of
the
r
e
s
our
c
e
s
in
c
loud
c
omput
ing
e
nvir
onment
[
25
]
.
5.
P
ROP
OS
E
D
M
E
T
HO
D
Algor
it
hm
s
teps
:
S
tep
1:
C
omput
ing
the
tas
k
r
a
ti
ng
s
im
il
a
r
it
y
r
e
que
s
ted
by
the
a
c
ti
ve
a
nd
no
r
mal
us
e
r
is
given
by
̅
a
n
d
̅
r
e
pr
e
s
e
nts
the
a
ve
r
a
ge
log
li
ke
li
hood
e
s
ti
mator
o
f
a
ll
s
e
r
vice
s
invoked
by
a
c
ti
ve
a
nd
nor
mal
us
e
r
s
to
s
e
r
vice
.
T
he
c
ontextua
l
tas
k
s
im
il
a
r
it
y
r
e
que
s
ted
by
the
a
c
ti
ve
c
loud
a
nd
nor
mal
c
loud
us
e
r
s
is
given
by
:
(
,
)
=
∑
(
,
−
̅
̅
̅
)
∈
(
,
−
̅
̅
̅
)
√
∑
(
,
−
̅
̅
̅
)
2
∑
(
,
−
̅
̅
̅
)
2
∈
∈
∗
(
(
,
)
)
S
tep
2:
I
n
thi
s
s
tep,
the
s
im
il
a
r
it
y
c
omput
a
ti
on
o
f
t
he
us
e
r
s
ba
s
e
d
on
the
QoS
va
lues
is
given
by
t
he
s
im
il
a
r
it
y
be
twe
e
n
the
us
e
r
s
int
e
r
a
c
ti
ng
with
the
s
e
r
vice
is
(
,
)
=
(
(
,
)
=
∑
(
,
−
̅
,
)
∈
(
,
−
̅
,
)
√
∑
(
,
−
̅
,
)
2
∑
(
,
−
̅
,
)
2
∈
∈
S
tep
3:
QoS
mi
s
s
ing
va
lue
pr
e
diction
us
ing
the
we
ight
e
d
s
im
il
a
r
it
y
mea
s
ur
e
s
a
s
S
tep
4:
Nor
malize
d
c
onf
idenc
e
s
c
or
e
f
o
r
us
e
r
s
is
c
omput
e
d
a
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
ol
,
Vol.
18
,
No.
3
,
J
une
2020:
125
2
-
12
5
8
1256
=
=
=
(
(
,
)
.
(
∑
(
,
)
∑
(
,
)
S
tep
5:
Nor
malize
d
c
onf
idenc
e
s
c
or
e
f
o
r
tas
ks
is
c
omput
e
d
a
s
:
ℎ
=
=
=
(
(
,
)
.
(
∑
(
,
)
∑
(
,
)
S
tep
6:
W
e
ight
e
d
s
c
or
e
f
or
us
e
r
s
is
c
omput
e
d
a
s
:
ℎ
=
=
.
(
)
+
(
1
−
)
S
tep
7:
W
e
ight
e
d
s
c
or
e
f
or
tas
ks
is
c
omput
e
d
a
s
:
ℎ
=
=
(
1
−
)
.
(
)
+
(
1
−
)
S
tep
8:
W
e
i
gted
p
r
e
dicte
d
s
c
or
e
f
o
r
Qos
mi
s
s
ing
v
a
lue
is
c
omput
e
d
a
s
=
2
(
.
)
(
+
)
{
[
]
,
[
]
}
S
tep
9:
Apply
P
S
O
a
lgo
r
it
hm
f
or
us
e
r
s
to
tas
ks
r
a
n
king.
6.
RE
S
UL
T
S
A
ND
AN
AL
YSI
S
T
he
e
f
f
icie
nc
y
of
the
pr
opos
e
d
s
c
he
me
is
c
ompa
r
e
d
with
s
ix
e
xis
ti
ng
c
ompeting
methods
na
mely
I
M
E
AN
,
UM
E
AN
,
UPC
C
,
I
P
C
C
,
UI
P
C
C
a
nd
I
B
GSS
QPr
e
d.
I
M
E
AN
:
it
e
m
mea
n:
a
ve
r
a
ge
Qo
S
va
lue
obs
e
r
ve
d
f
r
om
the
us
e
d
s
e
r
vice
a
s
the
pr
e
dicte
d
QoS
va
lue
of
the
us
e
r
f
or
the
unus
e
d
s
e
r
vice
.
Us
e
r
m
e
a
n
(
UM
E
AN
)
–
a
ve
r
a
ge
QoS
va
lue
known
by
the
u
s
e
r
f
or
the
us
e
d
s
e
r
vice
a
s
the
pr
e
dicte
d
QoS
va
lue
of
the
us
e
r
f
or
the
unus
e
d
s
e
r
vice
.
Us
e
r
-
b
a
s
e
d
c
o
ll
a
bor
a
ti
ve
f
il
ter
ing
method
us
ing
P
C
C
(
UPC
C
)
-
T
op
-
K
ne
ighbor
s
of
the
us
e
r
s
a
r
e
f
ound
us
ing
P
C
C
s
im
il
a
r
it
y
mea
s
ur
e
s
.
I
tem
-
ba
s
e
d
c
oll
a
bor
a
ti
ve
f
il
ter
ing
method
us
ing
P
C
C
(
I
P
C
C
)
-
T
op
-
K
ne
ighbor
s
of
the
s
e
r
vice
s
a
r
e
f
ound
us
ing
P
C
C
s
im
il
a
r
it
y
mea
s
ur
e
s
.
Us
e
r
a
nd
i
tem
-
ba
s
e
d
c
oll
a
bor
a
ti
ve
f
il
ter
ing
method
us
ing
P
C
C
(
UI
P
C
C
)
–
c
ombi
ne
s
UPC
C
a
nd
I
P
C
C
;
t
op
-
K
n
e
ighbor
s
of
the
us
e
r
s
a
nd
s
e
r
vice
s
a
r
e
f
ound
us
ing
P
C
C
s
i
mi
lar
it
y
mea
s
ur
e
s
f
or
QoS
p
r
e
diction.
T
he
c
omp
a
r
is
on
of
mea
n
s
qua
r
e
d
e
r
r
or
of
pr
opos
e
d
QoS
mea
s
ur
e
to
t
he
tr
a
dit
ional
QoS
mea
s
ur
e
s
on
the
c
loud
da
tas
e
t
i
s
s
hown
in
F
igur
e
1.
F
igur
e
1.
C
ompar
is
on
of
mea
n
s
qua
r
e
d
e
r
r
or
o
f
p
r
o
pos
e
d
model
to
the
tr
a
dit
ional
QoS
model
s
on
the
tr
a
ini
ng
c
loud
da
tas
e
t
F
igur
e
1
re
pr
e
s
e
nts
the
c
ompar
a
ti
ve
a
na
lys
is
of
pr
opos
e
d
QoS
model
to
the
c
onve
nti
ona
l
QoS
models
on
the
tr
a
ini
ng
c
loud
da
tas
e
t.
I
n
the
f
igur
e
,
the
mea
n
s
qua
r
e
e
r
r
or
r
a
te
of
the
p
r
opos
e
d
mo
de
l
a
nd
e
xis
ti
ng
models
a
r
e
c
omput
e
d
on
the
t
r
a
ini
ng
da
tas
e
t.
F
r
om
the
F
ig
ur
e
1
it
is
noted
that
the
p
r
e
s
e
nt
m
ode
l
ha
s
les
s
e
r
r
or
r
a
te
than
the
c
onve
nti
ona
l
models
.
F
igu
r
e
2
r
e
pr
e
s
e
nts
the
c
ompar
a
ti
ve
a
na
lys
is
of
pr
opo
s
e
d
QoS
model
to
the
c
onve
nti
ona
l
QoS
models
on
the
tr
a
i
ning
c
loud
da
tas
e
t.
I
n
the
f
igur
e
,
the
r
oot
mea
n
s
qua
r
e
e
r
r
o
r
r
a
te
of
the
pr
opos
e
d
model
a
nd
e
xis
ti
ng
models
a
r
e
c
omput
e
d
on
the
tr
a
ini
ng
da
tas
e
t.
F
r
om
the
F
ig
u
r
e
2
i
t
is
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
ol
I
mpr
ov
e
d
quali
ty
of
s
e
r
v
ice
-
bas
e
d
c
loud
s
e
r
v
ice
r
a
nk
ing
and
r
e
c
omm
e
ndati
on
mode
l
(
Sir
is
ha
P
ot
lur
i
)
1257
noted
that
the
pr
e
s
e
nt
model
ha
s
les
s
R
M
S
E
r
a
t
e
than
the
c
onve
nti
ona
l
models
.
T
he
c
ompar
is
on
of
QoS
r
unti
me
o
f
p
r
opos
e
d
QoS
mea
s
ur
e
to
the
e
xis
ti
n
g
QoS
r
unti
me
mea
s
ur
e
on
the
c
loud
da
tas
e
t
is
s
hown
in
F
igur
e
3
.
F
igur
e
4
r
e
pr
e
s
e
nts
the
e
f
f
icie
nc
y
o
f
the
pr
opos
e
d
r
a
nking
mea
s
ur
e
on
c
loud
us
e
r
s
to
s
e
r
vic
e
s
a
nd
it
is
s
howing
that
the
a
ve
r
a
ge
va
lue
of
p
r
e
c
is
ion
c
or
r
e
lation
of
the
pr
opos
e
d
a
lgor
it
hm
is
be
tt
e
r
than
the
pr
e
vious
ly
e
xis
ti
ng
mea
s
ur
e
s
f
or
us
e
r
s
to
s
e
r
vice
s
int
e
r
a
c
ti
on.
F
igur
e
2.
C
ompar
is
on
of
r
oot
mea
n
s
qua
r
e
d
e
r
r
or
(
R
M
S
R
)
of
pr
opos
e
d
model
to
the
tr
a
dit
ional
Qo
S
models
on
the
tr
a
ini
ng
c
loud
da
tas
e
t
F
igur
e
3.
C
ompar
a
ti
ve
a
na
lys
is
of
p
r
opos
e
d
QoS
r
unti
me
to
the
e
xis
ti
ng
QoS
r
unti
me
mea
s
ur
e
s
on
the
input
c
loud
s
e
r
vice
da
tas
e
t
F
igur
e
4
.
C
ontextua
l
r
a
nking
c
ompar
is
on
of
pr
opo
s
e
d
r
a
nking
mea
s
ur
e
to
the
e
xis
ti
ng
r
a
nking
mea
s
ur
e
on
c
loud
we
b
s
e
r
vice
da
tas
e
t
7.
CONC
L
USI
ON
AN
D
F
UT
UR
E
WORK
QoS
ba
s
e
d
s
e
r
vice
r
a
nking
mec
ha
nis
m
us
ing
s
e
r
vice
s
im
il
a
r
it
y
ba
s
e
d
P
S
O
in
c
loud
c
omput
ing
e
nvir
onment
is
pr
opos
e
d
in
thi
s
pa
pe
r
.
T
his
mec
ha
nis
m
c
ontains
two
pha
s
e
s
f
or
opti
mal
s
e
r
vice
s
e
lec
ti
on
in
c
loud
c
omput
ing
e
nvi
r
onm
e
nt.
I
n
the
f
i
r
s
t
pha
s
e
,
the
c
omput
a
ti
on
of
s
im
il
a
r
i
ty
a
mong
the
us
e
r
s
a
nd
their
s
e
r
vice
s
is
c
on
s
ider
e
d.
I
n
the
s
e
c
ond
pha
s
e
,
c
omput
a
ti
on
of
the
mi
s
s
ing
va
lues
ba
s
e
d
on
the
va
lues
whic
h
a
r
e
us
ing
c
omput
e
d
s
im
il
a
r
it
y
mea
s
ur
e
s
is
c
a
lcula
ted.
T
he
mec
ha
nis
m
of
pr
e
diction
a
nd
r
a
nking
ba
s
e
d
o
n
QoS
is
va
li
da
ted
us
ing
W
S
Dr
e
a
m#1
QoS
da
tas
e
t
a
nd
we
b
s
e
r
vice
QoS
da
tas
e
t
us
ing
s
tatis
ti
c
a
l
mea
s
ur
e
s
mea
n
s
qua
r
e
d
e
r
r
or
a
nd
r
oot
mea
n
s
qua
r
e
e
r
r
o
r
.
T
he
e
f
f
i
c
ienc
y
o
f
the
r
a
nking
a
nd
r
e
c
omm
e
nda
ti
on
model
whic
h
is
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
ol
,
Vol.
18
,
No.
3
,
J
une
2020:
125
2
-
12
5
8
1258
pr
opos
e
d
in
thi
s
pa
pe
r
is
mea
s
ur
e
d
a
nd
the
va
lue
o
f
a
ve
r
a
ge
pr
e
c
is
ion
c
or
r
e
lation
is
givi
ng
be
tt
e
r
r
e
s
ult
s
than
the
e
xis
ti
ng
mea
s
ur
e
s
.
T
his
r
a
nking
mec
ha
nis
m
c
a
n
be
us
e
d
in
tas
k
s
c
he
duli
ng
to
e
f
f
icie
ntl
y
s
c
he
dule
the
tas
ks
to
r
e
s
our
c
e
s
.
I
n
the
f
utur
e
wor
k,
a
hyb
r
id
P
S
O
opti
mi
z
a
ti
on
model
is
int
e
gr
a
ted
with
the
p
r
opos
e
d
r
e
c
omm
e
nda
ti
on
a
nd
r
a
nking
model
f
o
r
e
f
f
icie
nt
ta
s
k
s
c
he
duli
ng
in
c
loud
c
omput
ing
e
nvir
onment.
RE
F
E
RE
NC
E
S
[1
]
N
.
K
rat
z
k
e,
P.
Ch
ri
s
t
i
an
Q
u
i
n
t
,
“
U
n
d
ers
t
an
d
i
n
g
cl
o
u
d
-
n
at
i
v
e
ap
p
l
i
cat
i
o
n
s
af
t
er
1
0
y
e
ars
o
f
cl
o
u
d
co
m
p
u
t
i
n
g
-
A
s
y
s
t
ema
t
i
c
map
p
i
n
g
s
t
u
d
y
,
”
Jo
u
r
n
a
l
o
f
S
y
s
t
e
m
s
a
n
d
S
o
f
t
wa
r
e
,
v
o
l
.
1
2
6
,
p
p
.
1
-
1
6
,
A
p
r
i
l
2
0
1
7
.
[2
]
T
.
L
y
n
n
et
a
l
.
,
“
A
Prel
i
mi
n
ary
Sy
s
t
em
at
i
c
Rev
i
e
w
o
f
C
o
mp
u
t
er
Sci
e
n
ce
L
i
t
era
t
u
re
o
n
C
l
o
u
d
Co
m
p
u
t
i
n
g
Re
s
ea
rch
u
s
i
n
g
O
p
en
S
o
u
rce
S
i
mu
l
at
i
o
n
P
l
at
f
o
rms
,
”
P
r
o
ceed
i
n
g
s
o
f
t
h
e
7
th
I
n
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
en
ce
o
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
a
n
d
S
er
v
i
ces
S
ci
e
n
ce
(CLO
S
E
R
2
0
1
7
)
,
v
o
l
.
1
,
p
p
.
5
3
7
-
5
4
5
,
2
0
1
7
.
[3
]
Y
.
Su
n
,
J
.
Z
h
an
g
,
Y
.
X
i
o
n
g
,
G
.
Z
h
u
,
“
D
at
a
Secu
ri
t
y
an
d
Pri
v
a
c
y
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
”
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
D
i
s
t
r
i
b
u
t
e
d
S
e
n
s
o
r
Ne
t
wo
r
ks
,
v
o
l
.
2
0
1
4
,
p
p
.
1
-
9
,
J
u
l
y
2
0
1
4
.
[4
]
A
.
A
l
b
u
g
m
i
,
M.
O
.
A
l
as
s
af
i
,
R.
J
.
W
al
t
ers
,
G
.
W
i
l
l
s
,
“
D
at
a
Secu
ri
t
y
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
”
2
0
1
6
F
i
f
t
h
In
t
er
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
F
u
t
u
r
e
G
en
e
r
a
t
i
o
n
C
o
m
m
u
n
i
c
a
t
i
o
n
Tech
n
o
l
o
g
i
e
s
(F
G
CT)
,
L
u
t
o
n
,
p
p
.
5
5
-
5
9
,
2
0
1
6
.
[5
]
I.
A
h
med
,
“
A
b
ri
ef
rev
i
e
w
:
s
ec
u
ri
t
y
i
s
s
u
es
i
n
cl
o
u
d
co
mp
u
t
i
n
g
an
d
t
h
ei
r
s
o
l
u
t
i
o
n
s
,
”
TE
LKO
M
N
IKA
Tel
eco
m
m
u
n
i
ca
t
i
o
n
Co
m
p
u
t
i
n
g
El
ect
r
o
n
i
c
s
Co
n
t
r
o
l
,
v
o
l
.
1
7
,
n
o
.
6
,
p
p
.
2
8
1
2
-
2
8
1
7
,
D
ecem
b
er
2
0
1
9
.
[6
]
W
.
Q
i
an
g
,
“
Perfo
rman
ce
a
n
d
s
ec
u
ri
t
y
i
n
cl
o
u
d
co
mp
u
t
i
n
g
,
”
Th
e
J
o
u
r
n
a
l
o
f
S
u
p
er
Co
m
p
u
t
i
n
g
,
v
o
l
.
7
5
,
no.
1
,
p
p
.
1
-
3
,
J
an
u
ary
2
0
1
9
.
[7
]
M.
R.
Mes
b
ah
i
,
A
.
M.
Rah
ma
n
i
,
M.
H
o
s
s
e
i
n
za
d
eh
,
“
Rel
i
ab
i
l
i
t
y
an
d
h
i
g
h
av
a
i
l
a
b
i
l
i
t
y
i
n
cl
o
u
d
co
m
p
u
t
i
n
g
en
v
i
ro
n
men
t
s
:
a
refere
n
ce
r
o
ad
ma
p
,
”
H
u
m
a
n
-
ce
n
t
r
i
c
Co
m
p
u
t
i
n
g
a
n
d
In
f
o
r
m
a
t
i
o
n
S
c
i
en
ce
s
,
v
o
l
.
8
,
n
o
.
2
0
,
D
ecemb
er
2
0
1
8
.
[8
]
E
.
Co
u
t
i
n
h
o
,
F.
R.
C
So
u
s
a,
P.
A
.
L
.
Reg
o
,
J
.
So
u
za,
“
E
l
as
t
i
ci
t
y
i
n
cl
o
u
d
co
mp
u
t
i
n
g
:
a
s
u
r
v
ey
,
”
A
n
n
a
l
s
o
f
Tel
eco
m
m
u
n
i
ca
t
i
o
n
s
-
A
n
n
a
l
e
s
D
es
Te
l
eco
m
m
u
n
i
ca
t
i
o
n
s
,
v
o
l
.
7
0
,
n
o
.
7
-
8
,
p
p
.
2
8
9
-
3
0
9
,
A
u
g
u
s
t
2
0
1
5
.
[9
]
B.
D
.
Mart
i
n
o
,
G
.
Cret
el
l
a,
A
.
E
s
p
o
s
i
t
o
,
“
Cl
o
u
d
Po
rt
a
b
i
l
i
t
y
an
d
In
t
er
o
p
era
b
i
l
i
t
y
,
”
E
n
cycl
o
p
e
d
i
a
o
f
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
May
2
0
1
6
.
[1
0
]
H
.
Meh
t
a,
V
.
K
.
Pras
ad
,
M.
Bh
av
s
ar,
“
E
ffi
c
i
en
t
Res
o
u
r
ce
Sch
ed
u
l
i
n
g
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
”
In
t
er
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
A
d
va
n
ced
R
es
e
a
r
c
h
i
n
C
o
m
p
u
t
er
S
c
i
en
c
e
,
v
o
l
.
8
,
n
o
.
3
,
p
p
.
8
0
9
-
8
1
5
,
A
p
r
i
l
2
0
1
7
.
[1
1
]
N
.
H
o
s
s
ai
n
,
Md
.
A
.
H
o
s
s
a
i
n
,
A
.
K
.
M
.
F.
Is
l
am,
“
Res
earch
o
n
E
n
erg
y
E
ffi
c
i
en
c
y
i
n
Cl
o
u
d
C
o
mp
u
t
i
n
g
,
”
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
S
c
i
en
t
i
f
i
c
a
n
d
E
n
g
i
n
eer
i
n
g
R
e
s
e
a
r
c
h
,
v
o
l
.
7
,
n
o
.
8
,
p
p
.
3
5
8
-
3
6
7
,
A
u
g
u
s
t
2
0
1
6
.
[1
2
]
R.
A
.
H
as
an
e
t
al
.
,
“
H
SO
:
A
H
y
b
ri
d
Sw
arm
O
p
t
i
m
i
zat
i
o
n
A
l
g
o
r
i
t
h
m
fo
r
Red
u
ci
n
g
E
n
erg
y
Co
n
s
u
m
p
t
i
o
n
i
n
t
h
e
Cl
o
u
d
l
et
s
,
”
TE
LKO
M
NIKA
Tel
ec
o
m
m
u
n
i
c
a
t
i
o
n
Co
m
p
u
t
i
n
g
E
l
ect
r
o
n
i
c
s
Co
n
t
r
o
l
,
v
o
l
.
1
6
,
n
o
.
5,
p
p
.
2
1
4
4
-
2
1
5
4
,
O
ct
o
b
er
2
0
1
8
.
[1
3
]
A.
Is
mai
l
,
N
.
A
.
J
ama
l
u
d
i
n
,
S.
Z
amb
ri
,
“
A
Rev
i
ew
o
f
E
n
er
g
y
-
a
w
are
Cl
o
u
d
Co
m
p
u
t
i
n
g
Su
r
v
ey
s
,
”
TE
LK
O
M
N
IKA
Tel
eco
m
m
u
n
i
ca
t
i
o
n
Co
m
p
u
t
i
n
g
E
l
ect
r
o
n
i
c
s
Co
n
t
r
o
l
,
v
o
l
.
1
6
,
n
o
.
6
,
p
p
.
2
7
4
0
-
2
7
4
6
,
D
ecem
b
er
2
0
1
8
.
[1
4
]
Y
.
X
i
n
g
,
Y
.
Z
h
an
,
“
V
i
r
t
u
a
l
i
za
t
i
o
n
a
n
d
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
”
F
u
t
u
r
e
W
i
r
el
e
s
s
Net
w
o
r
k
s
a
n
d
I
n
f
o
r
m
a
t
i
o
n
S
ys
t
e
m
s
,
v
o
l
.
1
4
3
,
p
p
.
3
0
5
-
3
1
2
,
2
0
1
2
.
[1
5
]
T
as
k
een
Z
ai
d
i
,
Ramp
rat
ap
,
“
V
i
rt
u
al
Mach
i
n
e
A
l
l
o
c
at
i
o
n
Po
l
i
cy
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
E
n
v
i
ro
n
men
t
u
s
i
n
g
Cl
o
u
d
Si
m,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
E
l
ect
r
i
c
a
l
a
n
d
C
o
m
p
u
t
er
E
n
g
i
n
eer
i
n
g
(IJE
C
E
)
,
v
o
l
.
8
,
n
o
.
1
,
p
p
.
3
4
4
–
3
5
4
,
Feb
ru
ar
y
2
0
1
8
.
[1
6
]
H
.
Ph
u
o
n
g
,
X
.
L
i
s
o
n
g
,
“
A
v
ai
l
ab
l
e
b
an
d
w
i
d
t
h
es
t
i
mat
i
o
n
i
n
p
u
b
l
i
c
cl
o
u
d
s
,
”
IE
E
E
INF
O
CO
M
2
0
1
8
-
I
E
E
E
Co
n
f
er
e
n
ce
o
n
C
o
m
p
u
t
er
Co
m
m
u
n
i
ca
t
i
o
n
s
W
o
r
k
s
h
o
p
s
(INF
O
C
O
M
W
K
S
H
P
S
)
,
H
o
n
o
l
u
l
u
,
H
I,
p
p
.
2
3
8
-
2
4
3
,
2
0
1
8
.
[1
7
]
C.
Z
h
an
g
,
Y
.
W
a
n
g
,
Y
.
L
v
,
H
.
W
u
,
H
.
G
u
o
,
“
A
n
E
n
erg
y
an
d
SL
A
-
A
w
are
Res
o
u
rce
Ma
n
ag
eme
n
t
S
t
rat
e
g
y
i
n
C
l
o
u
d
D
at
a
Cen
t
ers
,
”
H
i
n
d
a
w
i
S
c
i
en
t
i
f
i
c
P
r
o
g
r
a
m
m
i
n
g
,
v
o
l
.
2
0
1
9
,
p
p
.
1
6
,
N
o
v
em
b
er
2
0
1
9
.
[1
8
]
Sw
ami
D
as
M.
,
G
o
v
ar
d
h
a
n
A
.
,
V
i
j
a
y
a
L
ak
s
h
m
i
D
.
(
2
0
1
9
)
“
W
eb
Ser
v
i
ce
s
Cl
a
s
s
i
fi
ca
t
i
o
n
A
cro
s
s
Cl
o
u
d
-
Ba
s
ed
A
p
p
l
i
cat
i
o
n
s
,
”
S
o
f
t
Co
m
p
u
t
i
n
g
:
T
h
eo
r
i
e
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
7
4
2
,
p
p
.
2
4
5
-
2
6
0
,
A
u
g
u
s
t
2
0
1
8
.
[1
9
]
S.
H
ei
d
ar
i
,
R.
Bu
y
y
a,
“
Q
u
al
i
t
y
o
f
Serv
i
ce
(Q
o
S)
-
d
ri
v
e
n
r
es
o
u
rce
p
r
o
v
i
s
i
o
n
i
n
g
fo
r
l
arg
e
-
s
cal
e
g
rap
h
p
ro
ce
s
s
i
n
g
i
n
cl
o
u
d
c
o
mp
u
t
i
n
g
e
n
v
i
ro
n
me
n
t
s
:
G
rap
h
Pr
o
ces
s
i
n
g
-
as
-
a
-
Serv
i
ce
(G
PaaS),
”
Fu
t
u
re
G
en
erat
i
o
n
Co
m
p
u
t
er
Sy
s
t
e
ms
,
V
o
l
.
9
6
,
p
p
.
4
9
0
-
5
0
1
,
2
0
1
9
[2
0
]
S.
Q
i
p
i
n
g
,
W
.
X
i
ao
c
h
ao
,
N
.
G
u
i
h
u
a,
C.
D
o
n
g
l
i
n
,
“
QoS
-
aw
are
cl
o
u
d
s
e
r
v
i
ce
co
m
p
o
s
i
t
i
o
n
:
A
s
y
s
t
em
at
i
c
map
p
i
n
g
s
t
u
d
y
fro
m
t
h
e
p
ers
p
ect
i
v
e
o
f
co
m
p
u
t
at
i
o
n
al
i
n
t
el
l
i
g
en
ce,
”
E
x
p
er
t
S
ys
t
em
s
wi
t
h
A
p
p
l
i
ca
t
i
o
n
s
,
v
o
l
.
1
3
8
,
D
ec
emb
er
2
0
1
9
.
[2
1
]
M.
H
.
G
h
ah
raman
i
,
M.
Z
h
o
u
,
T
.
H
.
Ch
i
,
“
T
o
w
ar
d
Cl
o
u
d
Co
mp
u
t
i
n
g
Q
o
S
A
rch
i
t
ec
t
u
re
:
A
n
a
l
y
s
i
s
o
f
Cl
o
u
d
Sy
s
t
ems
an
d
Cl
o
u
d
Serv
i
ces
,
”
i
n
IE
E
E
/
C
A
A
J
o
u
r
n
a
l
o
f
A
u
t
o
m
a
t
i
ca
S
i
n
i
ca
,
v
o
l
.
4
,
n
o
.
1
,
p
p
.
6
-
1
8
,
J
a
n
u
ar
y
2
0
1
7
.
[2
2
]
S.
Po
t
l
u
r
i
,
K
.
S.
Rao
,
“
Q
u
al
i
t
y
o
f
Serv
i
ce
b
as
e
d
T
as
k
Sch
ed
u
l
i
n
g
A
l
g
o
r
i
t
h
ms
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
”
In
t
er
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
E
l
ect
r
i
c
a
l
a
n
d
C
o
m
p
u
t
er
E
n
g
i
n
ee
r
i
n
g
(IJ
E
C
E
)
,
v
o
l
.
7
,
n
o
.
2
,
p
p
.
1
0
8
8
-
1
0
9
5
,
A
p
ri
l
2
0
1
7
.
[2
3
]
Po
t
l
u
ri
S.
,
Rao
K
.
S.
,
“
Si
mu
l
at
i
o
n
o
f
Q
o
S
-
Ba
s
ed
T
a
s
k
Sch
ed
u
l
i
n
g
Po
l
i
c
y
fo
r
D
ep
e
n
d
e
n
t
an
d
In
d
e
p
en
d
e
n
t
T
as
k
s
i
n
a
Cl
o
u
d
E
n
v
i
ro
n
men
t
,
”
S
m
a
r
t
In
t
e
l
l
i
g
e
n
t
C
o
m
p
u
t
i
n
g
a
n
d
A
p
p
l
i
ca
t
i
o
n
s
,
v
o
l
.
1
5
9
,
p
p
.
5
1
5
-
5
2
5
,
Sep
t
em
b
er
2
0
1
9
.
[2
4
]
Sh
efal
i
V
ars
h
n
e
y
,
Raj
i
n
d
er
San
d
h
u
,
P.
K.
G
u
p
t
a,
Q
o
S
Bas
ed
Re
s
o
u
rce
Pro
v
i
s
i
o
n
i
n
g
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
E
n
v
i
r
o
n
me
n
t
:
A
T
ec
h
n
i
cal
S
u
rv
e
y
,
In
t
er
n
at
i
o
n
al
C
o
n
feren
ce
o
n
A
d
v
an
ce
s
i
n
Co
m
p
u
t
i
n
g
an
d
D
a
t
a
Sci
e
n
ces
ICA
CD
S
2
0
1
9
:
A
d
v
a
n
ce
s
i
n
C
o
m
p
u
t
i
n
g
a
n
d
D
a
t
a
S
c
i
en
ces
,
p
p
.
7
1
1
-
7
2
3
,
2
0
1
9
[2
5
]
N
.
So
mu
,
G
au
t
h
ama
R.
M.
R.
,
A
.
K
av
eri
,
A
k
s
h
ay
R.
K
.
,
K
.
K
ri
t
h
i
v
a
s
an
,
Sh
a
n
k
ar
S.
V
.
S.
,
“
IBG
SS:
A
n
Imp
ro
v
e
d
Bi
n
ar
y
G
rav
i
t
at
i
o
n
al
Searc
h
A
l
g
o
ri
t
h
m
b
as
e
d
s
earc
h
s
t
rat
eg
y
fo
r
Q
o
S
an
d
ran
k
i
n
g
p
red
i
ct
i
o
n
i
n
cl
o
u
d
en
v
i
ro
n
men
t
s
,
”
A
p
p
l
i
e
d
S
o
f
t
C
o
m
p
u
t
i
n
g
,
v
o
l
.
8
8
,
March
2
0
2
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.