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.
1
8
,
No.
1
,
F
e
br
ua
r
y
2020
,
pp.
530
~
537
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
.
v1
8
i
1
.
12169
530
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
A
r
e
vi
e
w
on
s
e
r
v
e
r
le
ss
a
r
c
h
ite
c
t
u
r
e
s
-
f
u
n
c
t
io
n
as a
s
e
r
vi
c
e
(
FaaS
)
i
n
c
lo
u
d
c
o
m
p
u
t
in
g
Ar
ok
ia
P
au
l
Raj
an
R
D
ep
ar
t
men
t
o
f
Co
m
p
u
t
er
Sci
en
ce,
CH
RIS
T
(D
eeme
d
t
o
b
e
U
n
i
v
er
s
i
t
y
),
Ben
g
al
u
ru
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
De
c
30
,
201
8
R
e
vis
e
d
Nov
5
,
20
19
Ac
c
e
pted
Nov
30
,
20
19
E
merg
e
n
ce
o
f
c
l
o
u
d
co
m
p
u
t
i
n
g
a
s
t
h
e
i
n
ev
i
t
a
b
l
e
I
T
co
mp
u
t
i
n
g
p
ara
d
i
g
m,
t
h
e
p
erce
p
t
i
o
n
o
f
t
h
e
c
o
mp
u
t
e
referen
ce
mo
d
el
a
n
d
b
u
i
l
d
i
n
g
o
f
s
erv
i
ces
h
a
s
ev
o
l
v
e
d
i
n
t
o
n
e
w
d
i
men
s
i
o
n
s
.
Serv
erl
e
s
s
co
m
p
u
t
i
n
g
i
s
a
n
ex
ecu
t
i
o
n
mo
d
el
i
n
w
h
i
ch
t
h
e
cl
o
u
d
s
erv
i
c
e
p
ro
v
i
d
er
d
y
n
am
i
cal
l
y
man
a
g
es
t
h
e
al
l
o
cat
i
o
n
o
f
co
mp
u
t
e
res
o
u
r
ces
o
f
t
h
e
s
erv
er.
T
h
e
co
n
s
u
mer
i
s
b
i
l
l
ed
fo
r
t
h
e
ac
t
u
a
l
v
o
l
u
me
o
f
res
o
u
rces
co
n
s
u
med
b
y
t
h
em,
i
n
s
t
ead
p
a
y
i
n
g
f
o
r
t
h
e
p
re
-
pur
c
h
as
e
d
u
n
i
t
s
o
f
co
mp
u
t
e
ca
p
aci
t
y
.
T
h
i
s
m
o
d
e
l
ev
o
l
v
ed
as
a
w
ay
t
o
ac
h
i
e
v
e
o
p
t
i
mu
m
c
o
s
t
,
mi
n
i
m
u
m
c
o
n
f
i
g
u
rat
i
o
n
o
v
erh
ea
d
s
,
a
n
d
i
n
crea
s
es
t
h
e
a
p
p
l
i
ca
t
i
o
n
's
ab
i
l
i
t
y
t
o
s
cal
e
i
n
t
h
e
cl
o
u
d
.
T
h
e
p
r
o
s
p
ect
i
v
e
o
f
t
h
e
s
erv
er
l
es
s
co
mp
u
t
e
mo
d
e
l
i
s
w
e
l
l
co
n
ce
i
v
e
d
b
y
t
h
e
ma
j
o
r
c
l
o
u
d
s
erv
i
ce
p
ro
v
i
d
ers
a
n
d
ref
l
ect
ed
i
n
t
h
e
ad
o
p
t
i
o
n
o
f
s
erv
er
l
es
s
c
o
mp
u
t
i
n
g
p
ara
d
i
g
m.
T
h
i
s
rev
i
ew
p
ap
er
p
res
en
t
s
a
co
mp
re
h
en
s
i
v
e
s
t
u
d
y
o
n
s
er
v
er
l
es
s
co
mp
u
t
i
n
g
arch
i
t
ect
u
re
an
d
al
s
o
ex
t
e
n
d
s
an
ex
p
er
i
men
t
at
i
o
n
o
f
t
h
e
w
o
rk
i
n
g
p
ri
n
ci
p
l
e
o
f
s
er
v
erl
e
s
s
c
o
mp
u
t
i
n
g
referen
ce
mo
d
e
l
ad
ap
t
ed
b
y
A
W
S
L
amb
d
a.
T
h
e
v
ari
o
u
s
res
earch
av
en
u
es
i
n
s
erv
er
l
es
s
co
mp
u
t
i
n
g
are
i
d
e
n
t
i
fi
e
d
an
d
p
res
e
n
t
e
d
.
K
e
y
w
o
r
d
s
:
AW
S
l
a
mbda
C
loud
c
omput
ing
F
a
a
S
Google
c
loud
f
unc
ti
on
M
icr
os
of
t
A
z
ur
e
f
unc
ti
on
S
e
r
ve
r
les
s
c
omput
ing
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
:
Ar
okia
P
a
ul
R
a
jan
R
,
De
pa
r
tm
e
nt
of
C
omput
e
r
S
c
ienc
e
,
C
HR
I
S
T
(
De
e
med
to
be
Unive
r
s
it
y)
,
B
e
nga
lur
u
,
I
ndia.
E
mail:
a
r
okia
.
r
a
jan@
c
hr
is
tuni
ve
r
s
it
y.
in
1.
I
NT
RODU
C
T
I
ON
C
loud
c
omput
ing
is
the
on
-
de
mand
c
ons
umpt
ion
of
c
omput
e
powe
r
,
s
tor
a
ge
,
da
taba
s
e
,
a
ppli
c
a
ti
ons
,
a
nd
a
ny
I
T
r
e
s
our
c
e
s
thr
ough
the
I
nter
ne
t
f
o
ll
owi
ng
pa
y
-
as
-
you
-
go
pr
icing
model
[
1
]
.
T
he
mos
t
ba
s
ic
wa
y
to
de
f
ine
wha
t
the
'C
loud'
is
that
it
is
a
c
omput
e
r
lo
c
a
ted
s
omew
he
r
e
e
ls
e
that
is
a
c
c
e
s
s
e
d
via
the
I
n
t
e
r
ne
t
a
nd
uti
li
z
e
d
in
s
ome
wa
y.
W
e
b
s
e
r
vice
s
is
a
ls
o
a
nother
na
me
f
or
wha
t
pe
ople
c
a
ll
the
c
loud.
T
he
c
loud
is
c
ompr
is
e
d
of
s
e
r
ve
r
c
omput
e
r
s
loca
ted
in
di
f
f
e
r
e
nt
loca
ti
ons
a
r
ound
the
wor
ld
[
2
]
.
W
he
n
we
us
e
a
c
loud
s
e
r
vice
li
ke
Ama
z
on
we
b
s
e
r
vice
s
(
AW
S
)
or
Google
C
loud
a
r
c
hit
e
c
tur
e
or
M
icr
os
of
t
Az
ur
e
,
we
a
r
e
a
c
tually
uti
li
z
ing
the
c
omput
e
r
s
be
longi
ng
to
thes
e
c
loud
s
e
r
vice
pr
o
vider
s
[
C
S
P
s
]
.
T
he
pr
inciple
of
c
loud
c
omput
ing
is
r
e
maining
a
s
it
is
,
but
,
the
ne
e
d
f
or
s
ha
r
e
d
wor
king
pr
inciple
,
e
nha
nc
e
ments
in
f
a
s
t
r
e
s
pons
e
of
the
s
e
r
vice
s
,
a
gil
it
y
of
r
e
s
our
c
e
pr
ovis
ioni
ng,
a
nd
mi
nim
ize
d
man
a
ge
ment
hur
dles
ha
s
be
e
n
the
ta
r
ge
ts
of
hype
r
s
c
a
le
C
S
P
s
[
3
]
.
I
ns
pe
c
ti
ng
the
leve
l
of
c
a
pit
a
l
inves
tm
e
nt
a
nd
man
a
ge
ment
invol
ve
d
to
r
e
a
c
h
the
a
bove
objec
ti
ve
s
,
ther
e
a
r
e
many
r
e
s
e
a
r
c
he
s
pr
opos
e
d
s
e
v
e
r
a
l
r
e
f
e
r
e
nc
e
models
.
Hype
r
s
c
a
le
da
ta
c
e
nter
s
will
gr
ow
f
r
om
338
in
numbe
r
a
t
the
e
nd
o
f
2016
to
628
by
2021
[
4]
.
T
he
y
will
r
e
pr
e
s
e
nt
53
pe
r
c
e
nt
o
f
a
ll
ins
talled
da
ta
c
e
nter
s
e
r
ve
r
s
by
2021.
T
his
kind
of
dr
a
s
ti
c
incr
e
a
s
e
in
the
c
omp
utational
r
e
quir
e
ments
,
the
r
e
is
a
ne
e
d
f
or
t
r
a
ns
f
or
mi
ng
th
e
‘
tr
a
dit
ional
da
ta
c
e
nter
s
’
int
o
‘
hype
r
s
c
a
le
da
tac
e
nter
s
’
s
ophis
ti
c
a
ted
wit
h
high
leve
ls
o
f
a
bs
tr
a
c
ti
on
a
nd
vir
tualiza
ti
on
.
R
e
c
e
nt
tec
hnica
l
a
dva
nc
e
ment
in
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
o
l
A
r
e
v
iew
on
s
e
r
v
e
r
le
s
s
ar
c
hit
e
c
tur
e
s
-
function
as
a
s
e
r
v
ice
(
F
aaS)
in
c
loud
c
omputing
(
A
r
ok
ia
P
aul
R
ajan
R
)
531
the
vir
tualiza
ti
on
tec
hnologi
e
s
whic
h
a
r
e
ve
r
y
p
r
omi
s
ing
to
a
c
hieve
hype
r
s
c
a
le
da
ta
c
e
ntr
e
s
e
a
s
il
y.
F
ig
ur
e
1
r
e
pr
e
s
e
nts
the
e
volut
ion
s
tage
s
of
c
loud
c
o
mput
ing.
I
n
the
in
i
ti
a
l
s
tage
s
vir
tualiza
ti
on
us
e
d
a
s
the
mea
n
f
or
s
of
twa
r
e
a
nd
s
e
r
vice
c
ons
oli
d
a
ti
on
by
whic
h
a
tt
a
ined
the
maximum
ut
il
iz
a
ti
on
of
the
r
e
s
our
c
e
s
a
nd
e
a
s
y
mana
g
e
ment.
Dur
ing
the
i
nit
ial
pha
s
e
,
ther
e
wa
s
a
c
omm
on
s
ha
r
ing
of
ha
r
d
wa
r
e
.
I
n
the
ne
xt
pha
s
e
,
ther
e
wa
s
a
pool
of
vir
tual
mac
hines
(
VM
s
)
c
r
e
a
ted
on
a
s
e
r
ve
r
a
nd
e
a
c
h
V
M
c
a
r
r
y
a
c
opy
of
a
n
ope
r
a
ti
ng
s
ys
tem.
L
a
ter
,
it
a
dva
nc
e
d
int
o
t
he
c
onc
e
pt
of
c
ontaine
r
s
whe
r
e
it
include
d,
OS
leve
l
vir
tualiza
ti
on
[
5]
.
T
he
c
ontaine
r
s
a
r
e
the
platf
or
m
s
uf
f
icie
nt
e
nough
t
o
hold
the
r
e
s
our
c
e
s
ne
e
de
d
f
or
r
unning
a
s
pe
c
if
ic
a
ppli
c
a
ti
on.
I
t
a
c
hieve
d
higher
a
bs
tr
a
c
ti
o
n
of
r
e
s
our
c
e
s
c
ompar
ing
with
VM
s
.
I
n
c
ontaine
r
iza
ti
on,
r
e
s
our
c
e
pr
ovis
ioni
ng
is
much
f
a
s
ter
than
VM
s
.
Apa
r
t
f
r
om
the
e
f
f
icie
nc
ies
a
nd
f
a
s
ter
r
a
te
of
pr
ovis
ioni
ng
of
r
e
s
our
c
e
s
thr
ough
c
ontaine
r
iza
ti
on,
f
ur
ther
e
nha
nc
e
ments
a
r
e
c
ons
tr
a
ined
with
the
ba
s
ic
inf
r
a
s
tr
uc
tur
a
l
e
leme
nts
c
a
ll
e
d
s
e
r
ve
r
s
.
S
e
r
ve
r
les
s
c
omput
ing
is
a
model
of
pooli
ng
a
nd
uti
li
z
ing
th
e
r
e
s
our
c
e
s
whic
h
include
s
OS,
r
unti
me
e
nvir
on
ments
a
nd
ha
r
dwa
r
e
[
6]
.
F
igur
e
2
pr
e
s
e
nts
the
e
volut
ion
o
f
s
e
r
ve
r
les
s
c
omput
ing
f
r
om
c
ontaine
r
iza
ti
on.
S
e
r
ve
r
les
s
c
omput
ing
or
f
unc
ti
on
-
as
-
a
-
s
e
r
vice
(
F
a
a
S
)
is
de
f
ined
a
s
a
s
of
twa
r
e
a
r
c
hit
e
c
tur
e
whe
r
e
a
n
a
ppli
c
a
ti
on
is
de
c
ompos
e
d
int
o
‘
tr
igger
s
’
(
e
ve
nts
)
a
nd
‘
a
c
ti
ons
’
(
f
unc
ti
ons
)
,
a
nd
ther
e
is
a
platf
o
r
m
that
pr
ovides
a
s
e
a
ml
e
s
s
hos
ti
ng
a
nd
e
xe
c
uti
on
e
nvir
onment
[
7
]
.
T
he
a
ppli
c
a
ti
on
de
ve
loper
’
s
c
onc
e
r
n
only
f
or
li
ght
we
ight
e
d
a
nd
s
tate
les
s
f
unc
ti
ons
that
c
a
n
be
e
xe
c
uted
th
r
oug
h
a
n
AP
I
ba
s
e
d
on
the
on
-
de
mand
pr
inciple
.
T
he
a
p
pl
ica
ti
on
c
ons
umes
the
r
e
s
our
c
e
s
to
the
point
of
e
xe
c
uti
on
a
nd
late
r
the
r
e
s
our
c
e
s
a
r
e
r
e
lea
s
e
d.
T
he
pr
ice
model
include
s
only
the
a
mount
of
ti
me
in
whic
h
the
r
e
s
our
c
e
s
we
r
e
in
us
e
a
nd
the
a
ppli
c
a
ti
on
de
ve
loper
ne
e
d
not
to
pa
y
f
or
r
e
s
our
c
e
s
unti
l
they
a
r
e
e
x
e
c
uted,
thus
it
is
r
e
f
e
r
r
e
d
to
a
s
‘
s
e
r
ve
r
les
s
’
.
I
n
s
e
r
ve
r
les
s
c
omput
ing,
the
r
e
s
pons
ibi
li
ti
e
s
of
the
c
loud
s
e
r
vice
pr
ovider
include
the
mana
ge
ment
of
the
da
ta
c
e
ntr
e
,
s
e
r
ve
r
a
nd
the
r
unti
me
e
nvi
r
on
ment.
A
c
ontr
a
s
t
to
the
other
c
loud
models
,
th
e
mor
e
r
e
s
po
ns
ibi
li
ty
is
ve
s
ted
on
the
s
houlder
of
c
loud
s
e
r
vice
pr
ovider
a
nd
the
de
ve
loper
is
r
e
li
e
ve
d
with
the
mana
ge
ment
a
nd
maintena
nc
e
c
ompl
ica
ti
ons
a
ny
f
ur
ther
[
8
]
.
T
he
r
e
s
t
of
the
pa
pe
r
is
s
tr
uc
tu
r
e
d
a
s
f
oll
ows
:
S
e
c
t
ion
I
I
p
r
e
s
e
nts
the
de
tailed
s
tudy
of
r
e
late
d
wo
r
ks
on
the
c
onc
e
ptualiza
ti
on
of
s
e
r
ve
r
les
s
c
omput
ing.
S
e
c
ti
on
I
I
I
pr
e
s
e
nts
a
f
e
w
e
nter
p
r
is
e
us
e
c
a
s
e
s
f
it
ti
ng
to
the
s
e
r
ve
r
les
s
c
omput
ing.
S
e
c
ti
on
I
V
s
hows
the
c
ompar
is
on
of
f
e
a
tur
e
s
by
the
top
F
a
a
S
s
e
r
vice
pr
ovider
s
.
S
e
c
ti
on
V
pr
e
s
e
nts
a
de
mons
tr
a
ti
o
n
to
unde
r
s
tand
the
wor
king
p
r
inciple
of
s
e
r
ve
r
les
s
c
omput
ing
us
i
ng
AW
S
L
a
mbda.
S
e
c
ti
on
VI
pr
e
s
e
nts
a
f
e
w
tec
hnica
l
di
f
f
ic
ult
ies
a
nd
r
e
s
e
a
r
c
h
ga
ps
in
F
a
a
S
.
S
e
c
ti
on
VI
I
c
onc
l
ude
s
by
r
e
ins
tating
the
s
igni
f
ica
nc
e
of
the
s
e
r
ve
r
les
s
c
omput
ing
pa
r
a
digm
.
F
ig
ur
e
1.
E
vo
lut
ion
o
f
s
ha
r
ing
r
e
s
our
c
e
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
o
l
,
Vol.
1
8
,
No
.
1
,
F
e
br
ua
r
y
2020
:
530
-
537
532
F
ig
ur
e
2.
S
ha
r
ing
r
e
s
our
c
e
s
in
s
e
r
ve
r
les
s
c
omput
in
g
2.
RE
L
AT
E
D
WORKS
B
or
n
f
r
o
m
a
ne
e
d
to
make
platf
o
r
m
a
s
a
s
e
r
vice
(
P
a
a
S
)
mor
e
a
c
c
e
s
s
ibl
e
,
f
ine
-
gr
a
ined,
a
nd
a
f
f
or
da
ble,
s
e
r
ve
r
les
s
c
omput
ing
ha
s
ga
r
ne
r
e
d
int
e
r
e
s
t
f
r
om
both
indus
tr
y
a
nd
a
c
a
de
mi
a
.
T
he
wor
k
[
9]
a
im
s
to
give
a
n
unde
r
s
tanding
of
thes
e
e
a
r
ly
da
ys
of
s
e
r
ve
r
les
s
c
omput
ing:
wha
t
it
is
,
whe
r
e
it
c
omes
f
r
om
,
wha
t
is
t
he
c
ur
r
e
nt
s
tatus
of
s
e
r
ve
r
les
s
tec
hnology,
a
nd
wha
t
a
r
e
it
s
ma
in
obs
tac
les
a
nd
oppo
r
tuni
ti
e
s
.
T
.
L
ynn,
e
t
a
l.
,
[
10]
pr
ovides
a
r
e
view
a
nd
mul
ti
-
leve
l
f
e
a
tur
e
a
na
lys
is
of
s
e
ve
n
e
nter
pr
is
e
s
e
r
ve
r
les
s
c
omput
ing
plat
f
or
ms
.
I
t
r
e
vie
ws
e
xtant
r
e
s
e
a
r
c
h
on
thes
e
plat
f
or
ms
a
nd
identi
f
ies
the
e
mer
ge
nc
e
of
AW
S
L
a
mbda
a
s
a
n
a
c
tual
ba
s
e
pla
tf
or
m
f
or
r
e
s
e
a
r
c
h
on
e
nter
pr
is
e
s
e
r
ve
r
les
s
c
loud
c
omput
ing.
A
nove
l
de
s
ign
of
pe
r
f
or
manc
e
-
or
iente
d
s
e
r
ve
r
les
s
c
omput
ing
platf
or
m
de
ployed
in
M
icr
os
of
t
Az
u
r
e
,
a
nd
uti
li
z
ing
W
indows
c
ontaine
r
s
a
s
f
unc
ti
on
e
xe
c
uti
on
e
nvir
onments
[
11]
.
T
he
r
e
a
r
e
metr
ics
pr
o
pos
e
d
to
ev
a
luate
the
e
xe
c
uti
on
pe
r
f
o
r
manc
e
of
s
e
r
ve
r
les
s
platf
or
ms
a
nd
c
onduc
t
tes
ts
with
the
p
r
opos
e
d
pr
otot
ype
.
T
he
mea
s
ur
e
ments
s
howe
d
s
igni
f
ica
nt
im
pr
ove
me
nt
in
a
c
hieving
gr
e
a
ter
th
r
oughput
than
other
plat
f
or
ms
a
t
mos
t
c
onc
ur
r
e
nc
y
leve
ls
.
T
he
other
platf
or
m
a
lt
e
r
na
ti
ve
s
[
12]
to
AW
S
L
a
mbda,
no
dis
c
r
e
te
a
c
a
de
mi
c
r
e
s
e
a
r
c
he
s
us
ing
Az
ur
e
F
unc
ti
ons
,
Google
C
loud
F
unc
ti
ons
,
I
B
M
B
luemix
Ope
nW
his
k,
I
r
on.
io
I
r
o
nwor
ke
r
,
W
e
btas
k,
Ga
lac
ti
c
F
og
Ge
s
tal
L
a
s
e
r
we
r
e
identif
ie
d.
J
.
S
hor
t
,
e
t
a
l.
,
[
13
]
p
r
e
s
e
nt
ed
thr
e
e
de
mons
tr
a
tor
s
f
or
I
B
M
B
luemix
Ope
nW
his
k.
T
he
y
de
mons
tr
a
te
e
ve
nt
-
ba
s
e
d
pr
ogr
a
mm
ing
tr
igge
r
e
d
by
we
a
ther
f
o
r
e
c
a
s
t
da
ta,
Apple
W
a
tchO
S
2
a
ppli
c
a
ti
on
da
ta,
a
n
d
s
pe
e
c
h
utt
e
r
a
nc
e
s
.
I
t
a
ls
o
de
mons
tr
a
ted
a
c
ha
tbot
us
ing
I
B
M
B
luemix
Ope
nW
his
k
that
c
a
ll
s
on
I
B
M
W
a
t
s
on
s
e
r
vice
s
i
nc
ludi
ng
ne
ws
,
jokes
,
da
tes
,
we
a
ther
,
mus
ic
tut
o
r
a
nd
a
n
a
lar
m
s
e
r
vice
.
[
14
]
c
onduc
ted
a
s
ur
ve
y
on
the
e
xis
ti
ng
s
e
r
ve
r
les
s
platf
or
ms
f
r
om
indus
tr
y,
a
c
a
de
mi
a
,
a
nd
ope
n
s
our
c
e
pr
ojec
ts
,
ke
y
c
ha
r
a
c
ter
is
ti
c
s
a
nd
us
e
c
a
s
e
s
,
a
nd
de
s
c
r
ibe
tec
hnica
l
c
ha
ll
e
nge
s
a
nd
ope
n
pr
ob
lem
s
.
T
his
wor
k
pr
e
s
e
nted
a
ha
nds
-
on
e
xpe
r
ienc
e
of
us
ing
the
s
e
r
ve
r
les
s
tec
hnologi
e
s
a
va
il
a
ble
f
r
om
dif
f
e
r
e
nt
c
loud
pr
ovider
s
s
uc
h
a
s
I
B
M
,
Ama
z
on,
Google
a
nd
M
icr
os
of
t.
Z
.
Al
-
Ali
[
15]
de
s
igned
S
e
r
ve
r
les
s
OS,
c
ompr
is
e
d
of
thr
e
e
ke
y
c
omponents
:
(
a
)
a
ne
w
de
s
e
gr
e
ga
ti
on
model,
whic
h
leve
r
a
ge
s
de
s
e
gr
e
ga
ti
on
f
or
a
bs
tr
a
c
ti
on,
but
e
na
bles
r
e
s
our
c
e
s
to
move
f
lui
d
ly
be
twe
e
n
s
e
r
ve
r
s
f
or
pe
r
f
or
manc
e
,
(
b)
a
c
loud
or
c
he
s
tr
a
ti
on
laye
r
wh
ich
mana
ge
s
f
ine
-
gr
a
ined
r
e
s
our
c
e
a
ll
oc
a
ti
on
a
nd
p
l
a
c
e
ment
thr
oughout
the
a
ppli
c
a
ti
on's
li
f
e
ti
me
via
loca
l
a
nd
global
de
c
is
ion
making
,
a
nd
(
c
)
a
n
is
olation
c
a
pa
b
il
it
y
that
e
nf
or
c
e
s
da
ta
a
nd
r
e
s
our
c
e
is
olation.
[
16]
pr
opos
e
d
a
n
e
f
f
icie
nt
r
e
s
our
c
e
mana
ge
ment
s
ys
tem
f
or
s
e
r
ve
r
les
s
c
loud
c
omput
ing
f
r
a
mew
or
ks
with
the
goa
l
to
e
n
ha
nc
e
r
e
s
our
c
e
with
a
f
oc
us
on
memo
r
y
a
ll
oc
a
ti
on
a
mong
c
ontaine
r
s
.
T
he
de
s
ign
a
dde
d
a
laye
r
top
of
a
n
ope
n
-
s
our
c
e
s
e
r
ve
r
les
s
platf
or
m,
Ope
nL
a
mbda.
I
t
is
ba
s
e
d
upon
a
ppli
c
a
ti
on
wor
kloads
,
a
nd
s
e
r
ve
r
les
s
f
unc
ti
on’
s
memor
y
ne
e
ds
e
ve
nts
a
r
e
tr
igger
e
d.
T
he
memor
y
l
im
it
s
a
ls
o
lea
d
to
va
r
iations
in
the
number
of
c
ontaine
r
s
s
pa
wne
d
on
Ope
nL
a
mbda.
3.
E
NT
E
RP
RI
S
E
USE
CA
S
E
S
F
OR
S
E
RV
E
RL
E
S
S
COM
P
UT
I
NG
P
r
ope
r
inves
ti
ga
ti
on
of
the
na
tu
r
e
a
nd
ne
e
d
of
r
e
c
e
n
t
us
e
c
a
s
e
s
of
the
e
nter
pr
is
e
s
,
s
e
r
ve
r
les
s
c
omput
ing
be
c
omes
inevitable
move
t
o
f
ulf
il
l
the
r
e
quir
e
ment
s
.
T
he
s
c
e
na
r
ios
include
e
ve
nt
pr
oc
e
s
s
ing
with
big
da
ta,
API
or
c
he
s
tr
a
ti
on
a
mong
the
ve
ndo
r
s
,
c
ons
oli
da
ti
on
of
AP
I
s
to
mi
nim
ize
AP
I
c
a
ll
s
,
pr
oc
e
s
s
moni
to
r
ing,
a
nd
e
xe
c
uti
on
c
ontr
ol
f
or
tr
a
c
king
the
is
s
ue
s
.
T
he
f
o
ll
owing
a
r
e
s
ome
of
the
be
s
t
f
it
us
e
c
a
s
e
s
f
or
s
e
r
ve
r
les
s
c
omput
ing
[
17]
:
−
Us
e
c
a
s
e
1:
e
ve
nt
-
tr
igger
e
d
c
omput
ing
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
o
l
A
r
e
v
iew
on
s
e
r
v
e
r
le
s
s
ar
c
hit
e
c
tur
e
s
-
function
as
a
s
e
r
v
ice
(
F
aaS)
in
c
loud
c
omputing
(
A
r
ok
ia
P
aul
R
ajan
R
)
533
I
n
mul
ti
media
p
r
oc
e
s
s
ing
bus
ines
s
a
ppli
c
a
ti
ons
,
huge
volum
e
s
of
f
il
e
s
a
r
e
f
r
e
que
ntl
y
up
loade
d
to
O
bjec
t
S
tor
a
ge
S
e
r
vice
s
[
OSS
]
f
or
p
r
oc
e
s
s
ing.
T
he
r
e
q
uir
e
ments
may
be
s
uc
h
a
s
tr
a
ns
c
oding,
wa
ter
ma
r
king,
f
e
tching
the
da
ta.
T
his
bus
ines
s
s
c
e
na
r
io
invol
ve
s
a
va
r
iety
of
de
vice
s
li
ke
de
s
ktop
c
omput
e
r
s
or
P
D
As
or
mobi
le
phone
s
a
c
c
e
s
s
ing
dif
f
e
r
e
nt
f
il
e
types
of
upl
oa
ding
mul
ti
media
c
ontent
s
uc
h
a
s
im
a
ge
s
,
vi
de
os
,
a
nd
text
f
il
e
s
.
E
ve
nt
-
tr
igger
e
d
c
omput
ing
will
be
a
s
olu
ti
on
f
or
a
ddr
e
s
s
ing
many
tec
hnica
l
dif
f
iculti
e
s
by
e
ve
nt
-
tr
igger
e
d
c
omput
ing
.
−
Us
e
c
a
s
e
2:
li
ve
video
br
oa
dc
a
s
ti
ng
I
n
l
ive
video
br
oa
dc
a
s
ti
ng
s
c
e
na
r
ios
,
the
br
oa
dc
a
s
ti
ng
s
ynthes
izing
node
r
e
c
e
ives
a
udio
a
nd
video
s
tr
e
a
ms
f
r
om
the
hos
ts
.
T
he
c
oll
e
c
ted
da
ta
c
a
n
be
s
ynthes
ize
d
ba
s
e
d
on
the
f
unc
ti
on
c
omput
ing.
F
in
a
ll
y,
the
s
ynthes
ize
d
video
s
tr
e
a
m
ne
e
ds
to
be
pus
he
d
to
C
o
ntent
De
li
ve
r
y
Ne
twor
k
[
C
DN
]
.
−
Us
e
c
a
s
e
3:
I
oT
da
ta
p
r
oc
e
s
s
ing
I
oT
f
r
a
mew
or
k
ne
e
ds
a
n
e
f
f
icie
nt
f
unc
ti
on
c
omput
ing
de
s
ign
that
c
a
n
r
e
c
e
ive
s
tatus
da
ta
f
r
o
m
a
va
r
ie
ty
of
c
onne
c
ted
s
mar
t
de
vice
s
.
Als
o,
it
ne
e
ds
a
n
e
f
f
i
c
ient
e
ve
nt
-
ba
s
e
d
c
omput
ing
a
r
c
hit
e
c
tur
e
to
tr
a
ns
mi
t
the
pr
oc
e
s
s
e
d
da
ta
to
other
de
vice
s
or
s
tor
ing
int
o
t
he
da
taba
s
e
[
18]
.
−
Us
e
c
a
s
e
4:
s
ha
r
e
d
de
li
ve
r
y
s
ys
tem
A
global
g
r
oup
o
f
r
e
s
taur
a
nts
or
a
pr
oduc
t
-
ba
s
e
d
c
ompany
may
ne
e
d
a
n
e
ve
nt
-
ba
s
e
d
noti
f
ica
ti
on
s
ys
tem
to
the
ne
a
r
e
s
t
de
li
ve
r
y
pe
r
s
onne
l
t
o
pick
up
f
r
o
m
the
ne
a
r
e
s
t
s
e
ll
e
r
f
or
the
p
r
oduc
t
de
li
ve
r
y.
T
h
ough
e
ve
nt
-
ba
s
e
d
c
omput
ing
is
a
ppli
c
a
ble
in
many
s
uc
h
us
e
c
a
s
e
s
,
but
i
t
is
not
a
one
-
s
ize
-
f
it
s
-
a
ll
s
olut
ion.
I
f
the
r
e
que
s
ts
a
r
e
not
ha
ving
s
igni
f
ica
nt
f
luctua
ti
on
s
in
the
us
e
c
a
s
e
4,
then
f
un
c
ti
on
c
omput
ing
may
be
a
wr
ong
c
hoice
of
s
olut
ion
de
s
ign
[
19
]
.
4.
COM
P
AR
I
S
ON
OF
T
O
P
S
E
RV
E
RL
E
S
S
COM
P
UT
I
NG
CL
OUD
S
E
RV
I
CE
P
ROVI
DE
RS
I
n
a
s
hor
t
pe
r
iod,
the
s
e
r
ve
r
les
s
tec
hnology
ga
ined
a
lot
of
mom
e
ntum
in
the
indus
tr
y.
T
a
ble
1
pr
e
s
e
nts
the
c
ompar
is
on
of
va
r
iou
s
f
e
a
tur
e
s
pr
ov
ided
by
the
pionee
r
s
of
F
a
a
S
pr
ovide
r
s
[
7,
20
-
23]
.
T
a
ble
1.
C
ompar
is
on
of
f
e
a
tur
e
s
A
W
S
L
a
mbda
G
oogl
e
C
lo
ud F
unc
ti
on
M
ic
r
os
of
t
A
z
ur
e
F
unc
ti
on
I
nt
r
oduc
ti
on
2015
2016
2016
S
c
a
la
bi
li
ty
A
ut
oma
ti
c
A
ut
oma
ti
c
A
ut
oma
ti
c
M
a
x f
unc
ti
ons
U
nl
im
it
e
d
20 pe
r
pr
oj
e
c
t
D
e
pe
nds
on t
he
t
r
ig
ge
r
& a
va
il
a
bl
e
r
e
s
our
c
e
s
S
uppor
te
d l
a
ngua
ge
s
J
a
va
s
c
r
ip
t,
J
a
va
, P
yt
hon,
N
ode
J
S
J
a
va
s
c
r
ip
t
C
#, F
#, N
ode
J
S
, P
yt
hon, P
H
P
, B
a
s
h
C
onc
ur
r
e
nt
e
xe
c
ut
io
n
100 pa
r
a
ll
e
l
e
xe
c
ut
io
ns
pe
r
a
c
c
ount
U
nl
im
it
e
d
B
a
s
e
d on App
s
e
r
vi
c
e
D
e
pl
oyme
nt
Z
I
P
upl
oa
ds
Z
I
P
upl
oa
ds
, C
lo
ud s
to
r
a
ge
G
it
i
nt
e
gr
a
te
d, R
E
S
T
A
P
I
M
e
mor
y a
ll
oc
a
ti
on
P
e
r
f
unc
ti
on
N
ot
s
pe
c
if
ic
P
e
r
A
pp s
e
r
vi
c
e
L
ic
e
ns
in
g
C
lo
s
e
d s
our
c
e
O
pe
n s
our
c
e
O
pe
n s
our
c
e
P
r
ic
in
g mode
l
P
a
y a
s
c
ode
e
x
e
c
ut
e
s
P
a
y a
s
c
ode
e
x
e
c
ut
e
s
P
a
y a
s
c
ode
e
x
e
c
ut
e
s
E
ve
nt
dr
iv
e
n
a
r
c
hi
te
c
tu
r
e
S
3, S
N
S
, D
yna
mo DB
,
K
in
e
s
is
, C
lo
ud W
a
tc
h
C
lo
ud P
ub, C
lo
ud s
to
r
a
ge
obj
e
c
ts
A
z
ur
e
a
nd
th
ir
d
-
pa
r
ty
s
e
r
vi
c
e
s
5.
DE
M
ONS
T
RA
T
I
ON
OF
S
E
RV
E
RL
E
S
S
COM
P
UT
I
NG
USI
NG
AWS
L
AM
DA
T
he
objec
ti
ve
of
the
e
xpe
r
i
ment
c
a
r
r
ied
out
in
thi
s
pa
pe
r
is
to
de
mons
tr
a
te
the
method
of
c
onf
igur
ing
AW
S
L
a
mbda
f
o
r
r
e
s
ponding
the
noti
f
ica
ti
ons
f
r
o
m
the
Auto
S
c
a
li
ng
Gr
oup.
T
a
ble
2
pr
e
s
e
nts
the
gl
os
s
a
r
y
of
AW
S
s
e
r
vice
s
us
e
d
to
a
c
c
ompl
is
h
the
objec
ti
ve
of
the
e
xpe
r
im
e
nt
[
24
]
.
F
igur
e
3
pr
e
s
e
nts
the
de
mo
ns
tr
a
ti
on
s
c
e
na
r
io
whic
h
a
im
s
to
c
r
e
a
te
a
n
e
ve
nt
-
dr
iven
c
om
puti
ng
f
unc
ti
on
.
T
a
ble
2.
AW
S
s
e
r
vice
s
S
.N
o.
S
e
r
vi
c
e
P
ur
pos
e
1
A
ut
o S
c
a
li
ng G
r
oup
L
ogi
c
a
l
gr
oupi
ng
of
E
C
2
in
s
ta
nc
e
s
w
it
h
s
a
m
e
f
e
a
tu
r
e
s
us
e
d
f
or
th
e
s
c
a
le
ma
n
a
ge
me
nt
.
2
S
im
pl
e
N
ot
if
ic
a
ti
on S
e
r
vi
c
e
[
S
N
S
]
S
e
r
vi
c
e
us
e
d
f
or
de
li
ve
r
y
of
me
s
s
a
g
e
s
in
bul
k,
e
s
pe
c
i
a
ll
y
f
or
th
e
mobi
le
us
e
r
s
.
3
I
A
M
I
de
nt
it
y
a
nd
A
c
c
e
s
s
M
a
na
g
e
me
nt
is
a
s
e
r
vi
c
e
w
hi
c
h
pr
ovi
de
s
a
s
e
c
ur
e
w
a
y of
a
s
s
e
s
s
in
g A
W
S
r
e
s
our
c
e
s
.
4
A
W
S
L
a
mbda
E
ve
nt
-
dr
iv
e
n,
s
e
r
ve
r
le
s
s
c
omput
in
g
pl
a
tf
or
m
s
e
r
vi
c
e
th
a
t
r
uns
c
ode
in
r
e
s
pons
e
to
e
ve
nt
s
a
nd
a
ut
oma
ti
c
a
ll
y
ma
na
ge
s
th
e
c
omput
in
g
r
e
s
our
c
e
s
r
e
qui
r
e
d by tha
t
c
ode
[
20]
.
5
C
lo
ud W
a
tc
h
I
t
is
a
s
e
r
vi
c
e
us
e
d
to
r
e
c
e
iv
e
a
nd
moni
to
r
lo
g
f
il
e
s
,
s
e
t
a
l
a
r
ms
,
a
nd
a
ut
oma
ti
c
a
ll
y r
e
s
pond to c
ha
nge
s
i
n A
W
S
r
e
s
our
c
e
s
[
20]
.
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
o
l
,
Vol.
1
8
,
No
.
1
,
F
e
br
ua
r
y
2020
:
530
-
537
534
F
igur
e
3.
S
c
e
na
r
io
of
the
de
mons
tr
a
ti
on
5.
1.
De
m
on
s
t
r
at
ion
S
e
t
u
p
S
ome
of
the
AW
S
s
e
r
vice
s
c
a
n
a
utom
a
ti
c
a
ll
y
ge
ne
r
a
te
noti
f
ica
ti
ons
whe
n
a
n
e
ve
nt
oc
c
ur
s
.
S
uc
h
noti
f
ica
ti
ons
c
a
n
be
us
e
d
a
s
a
tr
igger
to
a
utom
a
te
a
c
ti
ons
without
r
e
quir
ing
human
int
e
r
ve
nti
on.
I
n
i
s
s
e
c
ti
on,
a
n
il
lus
tr
a
ti
on
ha
s
be
e
n
pr
e
s
e
nted
to
unde
r
s
t
a
nd
how
a
s
e
r
v
er
les
s
a
r
c
hit
e
c
tur
e
is
wor
king
[
25
]
.
T
he
s
c
e
na
r
io
is
to
us
e
AW
S
L
a
mbda
f
unc
ti
on
that
will
a
utom
a
ti
c
a
ll
y
s
na
ps
hot
a
nd
a
tt
a
c
h
a
ne
w
AW
S
E
C
2
ins
tanc
e
l
a
unc
he
d
by
the
a
uto
s
c
a
li
ng
gr
oup.
I
n
thi
s
il
lus
tr
a
ti
on
,
a
n
a
ut
o
s
c
a
li
ng
gr
oup
ha
s
b
e
e
n
a
lr
e
a
dy
c
onf
igur
e
d.
T
a
ble
2
s
hows
the
va
r
iety
of
AW
S
s
e
r
vice
s
incor
por
a
ted
to
a
c
c
ompl
is
h
the
objec
ti
ve
of
the
de
m
ons
tr
a
ti
on.
T
he
f
oll
owing
s
teps
a
r
e
c
a
r
r
ied
out
in
or
de
r
to
a
c
hi
e
ve
the
objec
ti
ve
of
thi
s
il
lus
tr
a
ti
on:
−
L
ogin
int
o
AW
S
a
c
c
ount
a
nd
ope
n
the
c
ons
ole
−
C
r
e
a
te
a
n
S
NS
topi
c
a
.
C
li
c
k
‘
Ge
t
s
tar
ted’
b
.
C
r
e
a
te
‘
T
opic’
with
the
topi
c
na
me.
c
.
C
li
c
k
‘
c
r
e
a
te
topi
c
’
−
C
onf
igur
e
Auto
S
c
a
li
ng
to
s
e
nd
E
ve
nt
a
.
On
the
‘
S
e
r
vice
s
’
,
c
li
c
k
E
C
2
b.
S
e
lec
t
‘
Auto
S
c
a
li
ng
Gr
oups
’
c
.
C
li
c
k
‘
Noti
f
ica
ti
ons
’
tab
d.
C
li
c
k
‘
C
r
e
a
te
noti
f
ica
ti
on’
e
.
C
onf
ir
m
‘
S
c
a
leE
ve
nt
’
is
s
e
lec
ted
in
‘
S
e
nd
a
noti
f
ica
ti
on
to’
f
.
I
n
‘
W
he
ne
ve
r
I
ns
tanc
e
s
’
s
e
lec
t
‘
launc
h’
.
g.
S
a
ve
.
−
C
r
e
a
te
a
n
I
AM
r
ole
f
or
the
L
a
mbda
f
unc
ti
on
a
.
I
n
‘
S
e
r
vice
s
’
,
c
li
c
k
‘
I
AM
’
b.
C
li
c
k
‘
R
oles
’
.
c
.
C
li
c
k
‘
C
r
e
a
te
R
ole’
d.
S
e
lec
t
‘
AW
S
S
e
r
v
ice
’
e
.
S
e
lec
t
us
e
c
a
s
e
a
s
‘
L
a
mbda’
.
f
.
Give
the
pe
r
mi
s
s
ion
a
s
‘
Ama
z
onE
C
2F
ull
Ac
c
e
s
s
’
.
g.
Give
a
R
ole
Na
me.
h.
C
li
c
k
‘
C
r
e
a
te
r
ole’
.
−
C
r
e
a
te
a
L
a
mbda
F
unc
ti
on
a
.
S
e
lec
t
‘
L
a
mbda’
f
r
om
‘
S
e
r
vice
s
’
.
b.
C
r
e
a
te
a
‘
F
unc
ti
on’
c
.
C
onf
igur
e
the
Na
me
,
R
unti
me
a
nd
R
ole.
d.
Add
the
F
unc
ti
on
c
ode
us
ing
P
ython
.
e
.
Add
‘
T
r
igge
r
s
’
on
‘
S
c
a
le
E
ve
nt’
−
Auto
S
c
a
li
ng
Gr
oup
S
c
a
li
ng
out
to
tr
igger
the
L
a
m
bda
f
unc
ti
on.
a
.
S
e
lec
t
‘
E
C
2’
in
‘
S
e
r
vice
s
’
.
b.
E
d
it
S
c
a
li
ng
Gr
oup
wi
th
de
s
ir
e
d
ins
tanc
e
s
a
s
2.
c
.
C
r
e
a
te
the
‘
S
na
ps
hots
’
whic
h
is
c
r
e
a
ted
by
the
L
a
mbda
f
unc
ti
on.
F
igur
e
4
s
hows
the
P
ython
c
ode
whic
h
c
r
e
a
tes
tr
igger
on
the
e
ve
nt
o
f
s
c
a
li
ng
out.
F
ig
u
r
e
5
s
hows
the
c
r
e
a
ti
on
of
a
t
r
igger
(
e
ve
nt)
f
or
the
L
a
mbda
f
un
c
ti
on.
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
o
l
A
r
e
v
iew
on
s
e
r
v
e
r
le
s
s
ar
c
hit
e
c
tur
e
s
-
function
as
a
s
e
r
v
ice
(
F
aaS)
in
c
loud
c
omputing
(
A
r
ok
ia
P
aul
R
ajan
R
)
535
F
igur
e
4
.
C
r
e
a
ti
ng
AW
S
L
a
mbda
us
ing
P
ython
in
t
he
c
ode
window
F
igur
e
5
.
C
r
e
a
ti
ng
t
r
igger
f
or
the
L
a
mbda
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
o
l
,
Vol.
1
8
,
No
.
1
,
F
e
br
ua
r
y
2020
:
530
-
537
536
5.
2.
Re
s
u
lt
s
On
s
uc
c
e
s
s
f
ul
c
ompl
e
ti
on
of
the
s
teps
1
to
6
will
c
r
e
a
te
two
s
na
ps
hots
that
we
r
e
c
r
e
a
ted
by
L
a
mbda
f
unc
ti
on.
I
t
is
the
r
e
pr
e
s
e
ntation
of
s
uc
c
e
s
s
f
ul
e
xe
c
uti
on
of
L
a
mbda
f
unc
ti
on
of
the
a
uto
s
c
a
li
ng
gr
oup.
I
f
the
s
na
ps
hots
we
r
e
not
c
r
e
a
ted,
then
the
L
a
mbda
f
unc
ti
on
e
it
he
r
ha
d
a
f
a
il
u
r
e
or
wa
s
not
t
r
igger
e
d.
6.
RE
S
E
AR
CH
AV
E
NU
E
S
I
N
S
E
RV
E
RL
E
S
S
COM
P
UT
I
NG
E
xtending
the
c
omm
e
r
c
ial
us
e
c
a
s
e
s
that
a
r
e
pr
e
s
e
nted
in
S
e
c
ti
on
I
I
would
be
a
be
tt
e
r
wa
y
to
e
xplor
e
ne
w
c
ha
ll
e
nge
s
put
f
or
th
by
thi
s
ne
w
c
omput
in
g
a
r
c
hit
e
c
tur
e
.
C
a
r
e
f
u
l
inves
ti
ga
ti
on
o
f
the
li
ter
a
tur
e
in
the
f
or
m
of
r
e
s
e
a
r
c
he
s
a
nd
white
pa
pe
r
s
s
ugge
s
ts
that
ther
e
a
r
e
a
va
r
iety
o
f
una
tt
e
nde
d
c
ha
ll
e
nge
s
hidden
de
s
pit
e
of
it
s
pr
omi
s
ing
p
r
e
dictions
.
T
he
f
ol
lo
wing
a
r
e
the
majo
r
c
a
tegor
ies
of
c
ha
ll
e
nge
s
identif
ied
b
y
[
26
,
27]
:
−
Ha
r
dwa
r
e
-
leve
l
c
ha
ll
e
nge
s
:
v
ir
tualiza
ti
on
of
s
e
r
ve
r
s
,
dis
tr
ibut
e
d
s
tor
a
ge
a
nd
their
leve
ls
,
int
e
r
ope
r
a
bil
i
ty
of
s
uppor
ti
ng
he
ter
oge
ne
ous
ha
r
dwa
r
e
of
ve
ndor
s
,
c
old
s
tar
ts
,
opti
mi
z
a
ti
on
of
r
e
s
our
c
e
s
,
a
nd
de
s
igni
ng
f
a
ult
tol
e
r
a
nt
s
ys
tem
a
r
c
hit
e
c
t
ur
e
s
.
−
De
ve
loper
-
leve
l
c
ha
ll
e
nge
s
:
l
a
c
k
of
tr
a
c
king
a
nd
de
bugging
tool
s
,
de
c
lar
a
ti
ve
de
ploy
ment,
the
e
xpe
r
ti
z
a
ti
on
of
a
pr
og
r
a
mm
ing
c
a
li
be
r
to
c
a
ter
r
e
f
a
c
tor
ing
of
e
xis
ti
ng
s
ys
tems
,
a
bil
it
y
to
int
e
gr
a
te
a
nd
c
ompos
e
,
mana
ging
a
nd
maintaining
,
de
s
igni
n
g
s
tat
e
f
ul
a
nd
s
tate
les
s
f
unc
ti
ons
,
t
r
a
ns
a
c
ti
on
a
nd
c
onc
ur
r
e
nc
y
mana
ge
ment,
opti
mi
z
e
d
c
ode
gr
a
nular
it
y,
r
e
c
ove
r
y
s
ys
tem
de
s
ign,
a
nd
a
da
pti
ng
De
vOps
pr
inciples
.
−
M
a
na
ge
ment
-
leve
l
c
ha
ll
e
nge
s
:
f
ixi
ng
the
r
e
s
our
c
e
li
mi
ts
,
r
e
s
our
c
e
pr
ovis
ioni
ng
a
nd
load
ba
lan
c
ing,
dyna
mi
c
s
c
he
duli
ng,
launc
h
ove
r
he
a
ds
,
lega
c
y
s
ys
tem
mi
gr
a
ti
on
,
pr
e
dicta
ble
s
c
a
labili
ty,
a
nd
s
e
c
ur
it
y
mec
ha
nis
ms
.
−
B
us
ines
s
-
l
e
ve
l
c
ha
ll
e
nge
s
:
c
os
t
e
s
ti
mation,
pr
icing
model,
mana
ging
hybr
i
d
c
loud
,
a
nd
non
-
c
loud
s
ys
tems
uti
li
z
ing
the
s
e
r
ve
r
les
s
a
r
c
hit
e
c
tur
e
s
.
7.
CONC
L
USI
ON
S
e
r
ve
r
les
s
c
omput
ing
is
in
the
s
tage
o
f
c
onc
e
ptual
iza
ti
on
by
the
r
e
s
e
a
r
c
he
r
s
a
nd
e
xpe
r
im
e
ntation
by
the
indus
tr
y.
I
t
is
pr
e
dicte
d
that
the
e
volut
ion
o
f
thi
s
ne
w
c
omput
ing
pa
r
a
digm
in
the
C
loud
will
de
f
ini
tely
lea
d
to
a
s
im
pler
,
c
he
a
pe
r
a
nd
mo
r
e
e
f
f
icie
nt
r
e
s
our
c
e
mana
ge
ment.
I
t
is
to
be
a
c
knowle
dge
d
that
thes
e
pr
omi
s
e
s
a
r
e
ba
s
e
d
on
howe
ve
r
s
uc
h
pr
opos
it
ions
a
r
e
ba
s
e
d
on
a
s
pe
c
if
ic
us
e
c
a
s
e
with
the
s
mall
s
c
a
le
of
de
plo
yment
by
the
indus
tr
y.
T
he
r
e
view
pr
e
s
e
nted
in
s
e
c
ti
on
2
a
nd
s
e
c
ti
on
3
pr
ovides
a
f
utur
is
t
ic
dim
e
ns
ion
f
or
s
e
r
ve
r
les
s
a
r
c
hit
e
c
tur
e
s
a
s
a
ne
w
e
r
a
of
c
omput
a
ti
on
whic
h
c
a
n
be
a
da
pted
f
or
a
b
r
oa
de
r
us
e
c
a
s
e
.
S
e
c
ti
on
2
a
ls
o
im
pli
e
s
the
pos
s
ibi
li
ti
e
s
of
e
xplor
ing
mor
e
r
e
s
e
a
r
c
h
a
ve
nue
s
f
or
the
a
c
a
de
mi
c
a
nd
r
e
s
e
a
r
c
h
c
omm
uni
ty
in
the
a
r
e
na
of
s
e
r
ve
r
le
s
s
c
omput
ing.
T
he
pr
inciple
of
s
e
r
ve
r
les
s
c
omput
ing
whe
n
de
ployed
in
non
-
c
loud
s
ys
tems
lea
d
to
a
ne
w
c
omput
ing
tec
hnology
known
‘
de
vice
les
s
e
dge
c
omput
ing’
.
S
e
r
ve
r
les
s
a
nd
de
vice
les
s
c
omput
ing
a
r
e
the
ne
w
bu
z
z
wor
ds
in
the
indus
tr
y
,
whic
h
pa
ve
s
a
wa
y
to
n
e
w
r
e
s
e
a
r
c
h
oppor
tuni
ti
e
s
in
the
C
loud
a
s
we
ll
a
s
non
-
c
loud
s
ys
tems
.
S
ince
s
e
r
ve
r
les
s
c
omput
ing
is
in
it
s
in
f
a
nt
s
tage
,
th
e
r
e
a
r
e
a
va
r
iety
of
tec
hnica
l
dif
f
icul
ti
e
s
a
nd
c
ha
ll
e
nge
s
that
a
r
e
una
ddr
e
s
s
e
d,
s
uc
h
a
s
s
moot
h
s
c
a
li
ng
with
a
to
ler
a
nc
e
of
ne
twor
k
ha
s
s
les
a
nd
s
e
c
ur
e
d
r
e
s
our
c
e
pr
ov
is
ioni
ng.
Als
o,
s
e
c
ti
on
6
identif
ied
a
f
e
w
r
e
s
e
a
r
c
h
dir
e
c
ti
ons
.
RE
F
E
RE
NC
E
S
[1
]
B
.
So
s
i
n
s
k
y
,
“
Cl
o
u
d
C
o
mp
u
t
i
n
g
Bi
b
l
e
,”
1s
t
ed
,
W
i
l
ey
Pu
b
l
i
s
h
i
n
g
,
2
0
1
1
.
[2
]
R
A
.
P
.
Raj
a
n
,
“
Serv
i
ce
Req
u
es
t
Sch
e
d
u
l
i
n
g
b
a
s
ed
o
n
Q
u
an
t
i
f
i
cat
i
o
n
Pri
n
ci
p
l
e
u
s
i
n
g
C
o
n
j
o
i
n
t
A
n
al
y
s
i
s
a
n
d
Z
-
s
c
o
re
i
n
Cl
o
u
d
,”
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
E
l
ec
t
r
i
ca
l
a
n
d
Co
m
p
u
t
e
r
E
n
g
i
n
eer
i
n
g
,
v
o
l
.
8
,
pp.
1
2
3
8
-
1
2
4
6
,
2
0
1
8
.
[3
]
L.
A
.
Barro
s
o
,
et
a
l
.
,
“
T
h
e
D
a
t
acen
t
er
as
a
C
o
mp
u
t
er
:
A
n
In
t
r
o
d
u
ct
i
o
n
t
o
t
h
e
D
e
s
i
g
n
o
f
W
are
h
o
u
s
e
-
s
cal
e
Mach
i
n
e
s
,”
S
yn
t
h
e
s
i
s
Lect
u
r
es
o
n
Co
m
p
u
t
e
r
A
r
ch
i
t
ec
t
u
r
e
,
v
o
l
.
8
,
p
p
.
1
-
1
5
4
,
2
0
1
3
.
[4
]
Ci
s
co
,
“
C
i
s
c
o
g
l
o
b
a
l
cl
o
u
d
i
n
d
ex
:
Fo
reca
s
t
an
d
met
h
o
d
o
l
o
g
y
,
2
0
1
5
-
2
0
2
0
,”
CA
:
Ci
s
co
Pu
b
l
i
c
,
2
0
1
8
.
[5
]
Mat
t
So
u
co
u
p
,
“
I
n
t
r
o
d
u
ct
i
o
n
t
o
Serv
er
l
es
s
Co
mp
u
t
i
n
g
,
”
T
e
l
eri
k
,
[O
n
l
i
n
e],
A
v
ai
l
ab
l
e:
h
t
t
p
s
:
/
/
w
w
w
.
t
el
er
i
k
.
c
o
m/
b
l
o
g
s
/
i
n
t
ro
d
u
c
t
i
o
n
-
to
-
s
erv
er
l
es
s
-
co
mp
u
t
i
n
g
,
2
0
1
8
.
[6
]
A
b
ram
s
,
H
.
,
“
T
h
e
E
v
o
l
u
t
i
o
n
o
f
Ser
v
erl
e
s
s
Co
m
p
u
t
i
n
g
,
”
[O
n
l
i
n
e],
A
v
ai
l
ab
l
e:
h
t
t
p
s
:
/
/
w
w
w
.
ca.
co
m
/
u
s
/
m
o
d
ern
-
s
o
ft
w
are
-
fac
t
o
r
y
/
c
o
n
t
en
t
/
t
h
e
-
e
v
o
l
u
t
i
o
n
-
of
-
s
er
v
erl
e
s
s
-
c
o
mp
u
t
i
n
g
.
h
t
ml
,
2
0
1
7
.
[7
]
A
mazo
n
,
“
Bu
i
l
d
i
n
g
A
p
p
l
i
ca
t
i
o
n
s
w
i
t
h
Serv
erl
e
s
s
A
rch
i
t
ec
t
u
re
s
,”
[O
n
l
i
n
e],
A
v
ai
l
a
b
l
e
:
h
t
t
p
s
:
/
/
a
w
s
.
amazo
n
.
co
m
/
l
am
b
d
a
/
s
er
v
erl
e
s
s
-
arc
h
i
t
ect
u
res
-
l
earn
-
m
o
re/
.
[8
]
Iv
an
D
w
y
er
,
“Serv
er
l
es
s
Co
m
p
u
t
i
n
g
D
e
v
el
o
p
er
E
mp
o
w
ermen
t
Reach
e
s
N
ew
H
ei
g
h
t
s
,
”
I
r
o
n
,
[
O
n
l
i
n
e],
A
v
ai
l
ab
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
i
ro
n
.
i
o
/
d
o
c
s
/
W
h
i
t
e
p
ap
er
_
Serv
er
l
es
s
_
F
i
n
a
l
_
V
2
.
p
d
f
.
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
o
l
A
r
e
v
iew
on
s
e
r
v
e
r
le
s
s
ar
c
hit
e
c
tur
e
s
-
function
as
a
s
e
r
v
ice
(
F
aaS)
in
c
loud
c
omputing
(
A
r
ok
ia
P
aul
R
ajan
R
)
537
[9
]
E
.
v
an
E
y
k
,
et
al
.
,
“
Serv
er
l
es
s
i
s
Mo
re
:
Fr
o
m
PaaS
t
o
Pre
s
en
t
Cl
o
u
d
Co
m
p
u
t
i
n
g
,”
I
E
E
E
In
t
er
n
et
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
2
,
p
p
.
8
-
17
,
2
0
1
8
.
[1
0
]
T
.
L
y
n
n
,
et
al
.
,
“
A
Prel
i
mi
n
ary
Rev
i
e
w
o
f
E
n
t
er
p
r
i
s
e
Serv
erl
e
s
s
Cl
o
u
d
Co
mp
u
t
i
n
g
(Fu
n
ct
i
o
n
-
as
-
a
-
Ser
v
i
ce)
Pl
at
f
o
rms
,”
2
0
1
7
IE
E
E
I
n
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
en
ce
o
n
C
l
o
u
d
Co
m
p
u
t
i
n
g
Tech
n
o
l
o
g
y
a
n
d
S
ci
e
n
ce
(Cl
o
u
d
Co
m
)
,
2
0
1
7
,
p
p
.
1
6
2
-
1
6
9
.
[1
1
]
G
.
McG
rat
h
,
P.
R.
Bren
n
e
,
“
Ser
v
erl
e
s
s
C
o
mp
u
t
i
n
g
:
D
e
s
i
g
n
,
Im
p
l
eme
n
t
a
t
i
o
n
,
a
n
d
Perf
o
rman
ce
,”
2
0
1
7
IE
E
E
3
7
th
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
D
i
s
t
r
i
b
u
t
ed
C
o
m
p
u
t
i
n
g
S
y
s
t
em
s
W
o
r
ks
h
o
p
s
,
p
p
.
4
0
5
-
4
1
0
,
2
0
1
7
.
[1
2
]
P.
Cas
t
ro
,
et
a
l
.
,
“
Serv
erl
e
s
s
Pr
o
g
ramm
i
n
g
(Fu
n
ct
i
o
n
as
a
Serv
i
ce)
,”
2
0
1
7
IE
E
E
3
7
th
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
en
ce
o
n
D
i
s
t
r
i
b
u
t
e
d
Co
m
p
u
t
i
n
g
S
ys
t
e
m
s
,
p
p
.
2
6
5
8
-
2
6
5
9
,
2
0
1
7
.
[1
3
]
J
.
Sh
o
rt
,
et
al
.
,
“C
l
o
u
d
E
v
e
n
t
Pro
g
rammi
n
g
Para
d
i
g
ms
:
A
p
p
l
i
cat
i
o
n
s
an
d
A
n
al
y
s
i
s
,
”
i
n
P
r
o
ceed
i
n
g
s
o
f
t
h
e
9
th
I
E
E
E
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
Cl
o
u
d
C
o
m
p
u
t
i
n
g
(CLO
U
D
)
,
p
p
.
4
0
0
-
4
0
6
,
2
0
1
7
.
[1
4
]
I.
Bal
d
i
n
i
,
et
al
.
,
“
Serv
erl
es
s
Co
m
p
u
t
i
n
g
:
Cu
rren
t
T
ren
d
s
an
d
O
p
e
n
Pro
b
l
ems
,”
R
es
e
a
r
ch
A
d
v
a
n
c
es
i
n
Cl
o
u
d
Co
m
p
u
t
i
n
g
,
S
p
r
i
n
g
e
r
,
p
p
.
1
-
20
,
2
0
1
7
.
[1
5
]
Z
.
A
l
-
A
l
i
,
“
Mak
i
n
g
Serv
er
l
es
s
C
o
mp
u
t
i
n
g
Mo
re
Ser
v
e
rl
es
s
,”
2
0
1
8
IE
E
E
1
1
th
In
t
er
n
a
t
i
o
n
a
l
Co
n
f
e
r
en
ce
o
n
C
l
o
u
d
Co
m
p
u
t
i
n
g
(CLO
U
D
)
,
p
p
.
4
5
6
-
4
5
9
,
2
0
1
8
.
[1
6
]
A
.
Sah
a
an
d
S.
J
i
n
d
al
,
“
E
MA
RS:
E
ff
i
ci
e
n
t
Ma
n
ag
eme
n
t
an
d
A
l
l
o
ca
t
i
o
n
o
f
Res
o
u
rce
s
i
n
Serv
er
l
es
s
,”
2
0
1
8
IE
E
E
1
1
th
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
Cl
o
u
d
C
o
m
p
u
t
i
n
g
(CLO
U
D
)
,
p
p
.
8
2
7
-
8
3
0
,
2
0
1
8
.
[1
7
]
E
ri
c
J
o
n
as
,
et
al
.
,
“
Cl
o
u
d
Pro
g
rammi
n
g
S
i
mp
l
i
f
i
ed
:
A
B
erk
el
e
y
V
i
ew
o
n
Ser
v
erl
e
s
s
C
o
mp
u
t
i
n
g
,
”
Tech
n
i
c
a
l
R
e
p
o
r
t
No
.
U
CB
/
E
E
CS
-
2
0
1
9
-
3
,
U
n
i
v
er
s
i
t
y
o
f
Ca
l
i
f
o
rn
i
a
at
Ber
k
el
e
y
,
2
0
1
9
.
[1
8
]
A
d
h
i
t
y
a
B
h
aw
i
y
u
g
a,
e
t
al
.
,
“
A
rch
i
t
ec
t
u
ra
l
d
e
s
i
g
n
o
f
I
o
T
-
cl
o
u
d
co
m
p
u
t
i
n
g
i
n
t
eg
r
at
i
o
n
p
l
at
f
o
rm,
”
TE
LK
O
M
NI
KA
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
a
n
d
Co
n
t
r
o
l
,
v
o
l
.
1
7
,
n
o
.
3,
p
p
.
1
3
9
9
-
1
4
0
8
,
2
0
1
7
.
[1
9
]
Fan
an
d
L
.
L
i
u
,
"
A
Su
rv
e
y
o
f
C
h
al
l
en
g
i
n
g
I
s
s
u
es
a
n
d
A
p
p
r
o
ach
e
s
i
n
Mo
b
i
l
e
Cl
o
u
d
C
o
mp
u
t
i
n
g
,
"
2
0
1
6
1
7
th
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
P
a
r
a
l
l
el
a
n
d
D
i
s
t
r
i
b
u
t
e
d
Co
m
p
u
t
i
n
g
,
A
p
p
l
i
ca
t
i
o
n
s
a
n
d
Tec
h
n
o
l
o
g
i
es
(P
D
C
A
T)
,
G
u
an
g
z
h
o
u
,
p
p
.
8
7
-
90
,
2
0
1
6
.
[2
0
]
K
.
K
ri
t
i
k
o
s
an
d
P.
Sk
rzy
p
e
k
,
“
Si
mu
l
at
i
o
n
-
as
-
a
-
Ser
v
i
ce
w
i
t
h
Serv
er
l
es
s
Co
mp
u
t
i
n
g
,
”
2
0
1
9
IE
E
E
W
o
r
l
d
Co
n
g
r
es
s
o
n
S
er
v
i
ces
,
It
a
l
y
,
p
p
.
2
0
0
-
2
0
5
,
2
0
1
9
.
[2
1
]
L
eo
n
a
Z
h
an
g
,
“
4
U
s
e
Cas
es
o
f
Serv
er
l
es
s
A
rch
i
t
ec
t
u
re
,
”
[O
n
l
i
n
e],
A
v
ai
l
ab
l
e:
h
t
t
p
s
:
/
/
d
z
o
n
e.
co
m
/
art
i
cl
es
/
4
-
use
-
ca
s
es
-
of
-
s
er
v
erl
e
s
s
-
arc
h
i
t
ect
u
re
,
2
0
1
8
.
[2
2
]
Mi
cro
s
o
f
t
,
“
A
zu
re
Fu
n
ct
i
o
n
s
D
o
cu
me
n
t
at
i
o
n
,
”
[O
n
l
i
n
e
],
A
v
ai
l
ab
l
e:
h
t
t
p
s
:
/
/
d
o
cs
.
mi
cr
o
s
o
ft
.
c
o
m/
e
n
-
u
s
/
az
u
re/
az
u
re
-
fu
n
c
t
i
o
n
s
/
.
[2
3
]
G
o
o
g
l
e
, “
G
o
o
g
l
e
Cl
o
u
d
F
u
n
c
t
i
o
n
s
D
o
c
u
men
t
at
i
o
n
,”
[O
n
l
i
n
e],
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
c
l
o
u
d
.
g
o
o
g
l
e.
co
m/
f
u
n
c
t
i
o
n
s
/
.
[2
4
]
A
mazo
n
W
e
b
Ser
v
i
ce
s
,
“
A
W
S
L
amb
d
a
D
ev
e
l
o
p
er
G
u
i
d
e
,”
[
O
n
l
i
n
e],
A
v
ai
l
a
b
l
e
:
h
t
t
p
:
/
/
d
o
cs
.
aw
s
.
amazo
n
.
co
m/
l
amb
d
a/
l
at
e
s
t
/
d
g
/
l
amb
d
a
-
d
g
.
p
d
f
.
[2
5
]
S.
H
en
d
r
i
ck
s
o
n
,
et
al
.
,
“Serv
erl
e
s
s
co
m
p
u
t
at
i
o
n
w
i
t
h
o
p
e
n
l
am
b
d
a,
”
H
o
t
c
l
o
u
d
’
1
6
,
2
0
1
6
U
S
E
NIX
A
n
n
u
a
l
Tech
n
i
c
a
l
Co
n
f
er
e
n
ce
,
2
0
1
6
.
[2
6
]
D
at
a
Ce
n
t
er
Fr
o
n
t
i
er,
“
D
a
t
a
Cen
t
er
D
e
v
el
o
p
er
s
:
Mee
t
i
n
g
t
h
e
C
h
al
l
en
g
e
s
o
f
T
o
d
a
y
’s
Re
q
u
i
reme
n
t
s
,
”
[O
n
l
i
n
e],
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
d
at
ace
n
t
erfr
o
n
t
i
e
r.
co
m/
d
at
a
-
ce
n
t
er
-
d
e
v
el
o
p
er
s
-
meet
i
n
g
-
ch
a
l
l
e
n
g
e
s
/
[2
7
]
D
.
G
an
n
o
n
,
R.
Barg
a
an
d
N
.
Su
n
d
ares
a
n
,
“Cl
o
u
d
-
n
at
i
v
e
ap
p
l
i
cat
i
o
n
s
,
”
I
E
E
E
Cl
o
u
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
4
,
n
o
.
5
,
p
p
.
1
6
–
21
,
2
0
1
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.