Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
6, N
o
. 2
,
A
p
r
il
201
6, p
p
.
74
3
~
75
0
I
S
SN
: 208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i
2.9
060
7
43
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
A Simulation-based Approach to
Optimize th
e Execution Time
and Minimizati
on of Average W
a
iting Ti
me Using Queui
n
g
Model in Cloud Computing Environment
Sou
v
ik
Pal, Pr
asant Kum
a
r
Pattn
a
ik
School of Comp
uter
Engineering
,
K
IIT University
, Bhuban
e
swar, India
Article Info
A
B
STRAC
T
Article histo
r
y:
Received Sep 22, 2015
Rev
i
sed
No
v
28
, 20
15
Accepted Dec 20, 2015
Cloud computin
g is the emerg
i
n
g
domain in academia and I
T
In
dustr
y
.
It is
a
business framework for delivering th
e services and computing power on-
demand basis. C
l
oud users hav
e
to pay
the service prov
iders b
a
sed on
their
usage. For
enter
p
rises, cloud
co
mputing
is th
e
worth
y
o
f
consideration and
they
try
to build busine
ss sy
ste
m
s w
ith lower costs, higher
profits and
quality
-
o
f-service. Consid
ering
cost
op
tmization, service pr
ovider m
a
y
initi
all
y
tr
y to
use less num
ber of CP
U cores and data ce
nt
ers.
For t
h
at
reason, this pap
e
r deals with CloudS
im si
mulation tool which has been
utili
zed
for
eva
l
uating
th
e nu
m
b
er of CPU cores and
ex
ec
ution t
i
m
e
.
Minim
i
zation
of
waiting
tim
e
i
s
also a
consid
erabl
e
issue
.
W
h
en a
larg
e
number of jobs
are r
e
quested
,
th
ey
hav
e
to wa
it
for gett
ing a
llo
c
a
ted
to th
e
servers which in
turn m
a
y
incre
a
se the qu
eue
le
ngth and a
l
so waiting
tim
e
.
This pap
e
r a
l
so dea
l
s with qu
euing m
odel
with m
u
lti-se
rver
and f
i
ni
t
e
capacity
to
redu
ce
the waiting
time and qu
eue length.
Keyword:
C
l
ou
d br
oke
r
C
l
ou
d c
o
m
put
i
n
g
Que
u
i
n
g m
ode
l
Virtu
a
lizatio
n
Waitin
g
tim
e
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
So
uvi
k Pal
,
Sch
ool
o
f
C
o
m
put
e
r
E
ngi
neer
i
ng,
C
a
m
pus 15
, KI
IT Uni
v
ersi
t
y
,
Pat
i
a
,
B
h
uba
ne
swar
, Odi
s
ha-
7
51
0
2
4
,
In
di
a
Em
ail: souvi
kpal22@gm
a
il.com
1.
INTRODUCTION
C
l
ou
d c
o
m
put
i
n
g
i
s
t
h
e
use
of t
h
e
Int
e
rnet
fo
r t
h
e t
a
s
k
s
t
h
e u
s
ers
pe
rf
orm
i
ng
on
t
h
e
i
r com
put
er.
C
l
ou
d com
put
i
n
g
,
al
so k
n
o
w
n
as Int
e
r
n
et
co
m
put
i
ng,
provi
d
es on-dem
and services
w
h
i
c
h are co
nce
r
ne
d wi
t
h
share
d
res
o
urc
e
s, soft
ware
, i
n
f
o
rm
at
i
on, an
d ot
he
r
application specific
services
. C
l
ou
d Ser
v
i
ce Pr
o
v
i
d
e
r
s
(CSPs) prov
ide th
e serv
ice acco
rd
ing
to
the clien
t
s’ requ
i
r
em
en
t with
min
i
m
a
l effo
rt at stip
u
l
ated
time [1
].
Clo
u
d
co
m
p
u
t
in
g
h
a
s also
th
e capab
ility t
o
d
e
li
v
e
r the
serv
ices t
o
th
e u
s
ers b
y
g
r
owing
or sh
ri
nk
ing
in
instantane
o
u
s r
e
qui
rem
e
nts [2
]
[3]
.
Clo
u
d
in
frastr
u
ctu
r
e implies a busines
s
fram
e
work
between cl
oud
users
and
C
SPs
creat
i
ng a
rel
a
t
i
o
ns
hi
p t
h
r
o
ug
h cl
o
u
d
b
r
oke
rage
s
e
rvi
ces
[
4
]
[
5
]
.
Clo
u
d
b
r
ok
ers
[6
], actin
g
as
n
e
go
tiato
r
o
r
med
i
ato
r
, h
e
l
p
o
u
t
th
e u
s
ers to
m
a
in
tain
all
th
e activ
itie
s
suc
h
as
speci
fic data ce
ntre
re
quired, e
x
e
c
ution tim
e of the
res
p
ective re
quests
,
num
b
er of CPU cores,
wai
t
i
ng t
i
m
e of t
h
e re
spect
i
v
e user
re
quest
.
C
l
ou
d b
r
o
k
e
r
al
so hel
p
s t
h
e
users
f
o
r sel
f
-p
ro
vi
si
o
n
i
n
g
of t
h
e
resources.
In
t
h
is m
a
n
u
s
crip
t
,
Clo
u
d
S
im
Si
m
u
la
to
r h
a
s b
e
en
used
t
o
fi
nd
th
e m
i
n
i
m
u
m ex
ecu
tion
time with
respect
t
o
t
h
e
num
ber
of C
P
U co
res. C
l
ou
d b
r
oke
r i
s
t
h
e
r
e to
decide the avera
g
e
waiting tim
e in the syste
m
and
also
in que
u
e using queui
n
g
m
odel.
The
pa
per i
s
o
r
gani
ze
d as
f
o
l
l
o
ws:
Sectio
n
2
d
eal
s with
th
e literatu
re surv
ey
o
f
th
e
related
work
and
obj
ectiv
e o
f
t
h
e stu
d
y
.
In
t
h
e
sect
i
on
3,
we
have
di
sc
usse
d a
b
o
u
t
t
h
e
si
m
u
l
a
t
i
on w
o
rk
fl
o
w
, se
q
u
enc
e
di
ag
ram
of t
h
e
fl
o
w
c
ont
e
n
t
an
d
si
m
u
latio
n
resu
lt u
s
ing
Clo
u
d
S
im
Si
m
u
lato
r V3
.0
. In
th
e
sect
i
on 4
,
we
have
prese
n
t
e
d
que
ui
n
g
m
odel
for
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
74
3 – 7
5
0
74
4
cl
ou
d c
o
m
put
i
ng e
n
vi
r
onm
ent
.
Sect
i
o
n
5
deal
s
wi
t
h
t
h
e
num
eri
cal
resul
t
a
n
al
y
s
i
s
and
g
r
a
phi
cal
rep
r
ese
n
t
a
t
i
on.
At
t
h
e
en
d
o
f
t
h
e
pape
r,
co
ncl
u
si
o
n
sect
i
o
n
b
r
i
n
gs t
o
a cl
o
s
e
o
f
t
h
e
w
o
r
k
.
2.
LITERATURE SURVE
Y
OF THE
RELATED WORK
In
clou
d
co
m
p
u
tin
g
en
v
i
ronmen
t,
m
i
n
i
miz
a
tio
n
of waiting
ti
m
e
in
th
e q
u
e
u
e
an
d
in
th
e syste
m
is
th
e
core
resea
r
ch
area in aca
de
mia and also
i
n
i
n
d
u
s
t
r
y
[
7
]
.
I
t
i
s
a
l
s
o
c
o
n
c
e
r
n
e
d
w
i
t
h
t
h
e
n
u
m
b
e
r
o
f
s
e
r
v
e
r
s
o
r
n
u
m
b
e
r of CPU co
res to
d
e
liv
er th
e serv
ices as p
e
r th
e u
s
er requ
est. Min
i
mizatio
n
o
f
waitin
g
tim
e
l
ead
s to
cust
om
er sat
i
s
fact
i
on a
nd
u
s
age o
f
l
e
ss num
ber o
f
CPU cores e
n
hances the
perform
ance of the cost
opt
i
m
i
zati
on.
Exec
ut
i
on t
i
m
e of a
part
i
c
ul
ar j
o
b re
que
st
i
s
al
so a co
n
s
i
d
era
b
l
e
i
ssue
i
n
cl
ou
d c
o
m
put
i
n
g
envi
ro
nm
ent
.
I
n
t
h
i
s
m
a
nuscri
p
t
,
we
have
m
a
de a
n
e
x
t
e
nsi
v
e su
rvey
t
h
at
i
s
rel
a
t
e
d t
o
wo
r
k
.
B
u
y
y
a
, R
a
n
j
an
, an
d C
a
l
h
ei
ro
s, 2
0
0
9
[
8
]
,
ha
ve
prese
n
t
e
d
C
l
ou
dSi
m
t
ool
ki
t
whi
c
h e
n
abl
e
s t
o
m
odel
an
d sim
u
late i
n
cloud
co
m
p
u
tin
g env
i
ro
nmen
ts. In
Clou
dSim
si
m
u
lat
o
r, th
ey
h
a
v
e
g
i
v
e
n
th
e provisio
n
to
create the number
of
Virtual
Machines
(VM
s
) in a
particul
ar Data Ce
nt
e
r
and t
o
use
di
f
f
ere
n
t
VM
al
l
o
cat
i
o
n
an
d
VM select
io
n
p
o
licies to
m
o
d
e
l. Mo
reov
er, th
is
p
a
p
e
r
allo
ws th
e
u
s
ers to
fi
n
d
t
h
e execu
tio
n ti
m
e
o
f
a jo
b
req
u
est
,
t
o
do
VM
m
i
grat
i
on,
an
d al
s
o
re
g
u
l
a
r scal
i
n
g
of
ap
pl
i
cat
i
ons i
n
cl
ou
d c
o
m
put
i
n
g
en
vi
r
onm
ent
.
Pu,
Li
u
,
et
al
.,
2
0
1
0
[
9
]
,
ha
v
e
p
r
esent
e
d t
h
e
i
r pe
rf
orm
a
nce
anal
y
s
i
s
i
n
pa
ral
l
e
l
pr
og
ressi
on
o
f
C
P
U.
They
ha
ve al
so
foc
u
se
d o
n
m
oni
t
o
ri
n
g
wo
rk
l
o
ad
on
Xe
n V
i
rt
ual
M
achi
n
e
M
oni
t
o
rs,
whi
c
h i
s
co
ncer
ne
d wi
t
h
net
w
or
k i
n
t
e
n
s
i
v
e w
o
r
k
l
o
a
d
. Thei
r pa
per
deal
s wi
t
h
t
h
e ex
pe
ri
m
e
nts fo
r fi
n
d
i
n
g
out
t
h
e pe
rf
or
m
a
nce
m
easurem
ent
s
on
net
w
o
r
k I/
O
w
o
r
k
l
o
a
d
.
Khazaei, Misic, et al., 2011
[10], ha
ve disc
usse
d about
the techniques of re
source provisioning, the
pr
oce
d
u
r
es
of
del
i
v
eri
ng
di
ffe
rent
t
y
pes
of se
r
v
i
ces su
ch as i
n
frast
r
u
ct
u
r
e-
base
d,
pl
at
fo
rm
-based
, an
d
soft
ware
-ba
s
ed. They ha
ve
foc
u
se
d to ca
lculate th
e pe
rform
a
nce
m
easurem
ent
while p
r
ov
ision
i
ng
th
e
resources to
realize the Service Leve
l Ag
reem
en
ts (SLAs). In
th
is
pap
e
r, th
ey
h
a
ve also
d
e
scri
bed
an
an
alytical r
e
p
r
esen
tatio
n
for
ev
alu
a
tion
of
dif
f
e
r
e
n
t
issu
es
su
ch
as r
e
sponse ti
m
e
, ser
v
er
f
a
r
m
s, an
d
num
b
e
r
o
f
tasks
for s
u
ffic
ient accuracy.
R
o
d
r
i
g
o, R
a
n
j
a
n
, et
al
., 20
11
[
11]
, ha
ve de
scr
i
bed ab
o
u
t
t
h
e sim
u
l
a
t
i
on st
rat
e
gy
of C
l
ou
dS
im
i
n
t
h
ei
r
pape
r. C
l
o
u
d
Si
m
t
ool
ki
t
su
pp
ort
s
t
h
e m
odel
i
ng a
n
d si
m
u
l
a
ti
on i
n
cl
ou
d e
n
vi
r
onm
ent
,
w
h
i
c
h i
s
al
so c
onc
erne
d
wi
t
h
res
o
u
r
ce
pr
o
v
i
s
i
oni
ng
, creat
i
on
of VM
s,
m
odel
i
n
g of
dat
a
cent
e
rs, cl
ou
d b
r
o
k
e
r
p
o
l
i
c
i
e
s, di
ffere
nt
SLA
s
et
c. C
l
ou
dSi
m
t
ool
ki
t
al
s
o
wo
rks
i
n
b
o
t
h
si
n
g
l
e
cl
o
u
d
an
d
f
e
derat
i
o
n
of
cl
ou
ds.
T
h
i
s
pa
p
e
r al
so
re
pre
s
e
n
t
s
t
h
e
i
m
p
r
ov
em
en
t o
f
th
e app
licatio
n
Qu
a
lity of Serv
ice
(Q
o
S
) req
u
i
rem
e
n
t
s.
Sp
illn
er, Brito
, et a
l
., 2
0
1
2
[12
]
, h
a
v
e
p
r
esen
ted
an
economicall
y
co
m
p
e
n
sation
co
n
c
ept to
raise th
e
gra
n
ularity and efficacy
of
res
e
rve
d
c
o
m
putation. T
h
is
pa
pe
r ena
b
les
highl
y
virtualized
re
source
broker i
n
the
b
u
s
i
n
ess-orien
t
ed
m
a
rk
et p
l
ace, wh
ich facilitates th
e co
nsumer with
con
f
i
g
urab
le
VMs fo
r resou
r
ce sharing.
Th
is
p
a
p
e
r supp
orts
o
n
-d
em
a
n
d resou
r
ce pro
v
i
si
o
n
i
n
g
wit
h
th
e h
e
l
p
o
f
scalab
ility.
Khazaei, Misi
c, et al., 2012
[13],
ha
ve
discusse
d about the
m
odeli
ng
of cloud cente
rs.
They ha
ve
pr
o
pose
d
a p
e
r
f
o
r
m
a
nce
m
easurem
ent
m
odel
t
o
eval
uat
e
t
h
e
cl
ou
d fa
rm
s and
fo
u
nd
o
u
t
t
h
e sol
u
t
i
on t
o
get
t
h
e
esti
m
a
t
i
o
n
o
f
prob
ab
ility d
i
strib
u
tion
.
Th
ei
r
m
o
d
e
l h
e
lp
s the CPSs to
d
eci
d
e
t
h
e
n
u
m
b
e
r
o
f
serv
ers, i
n
pu
t size,
and num
b
er
of
tasks in the
syste
m
.
Pal and
Pattnaik
,
20
13
[
14], h
a
v
e
p
r
esen
ted
v
i
r
t
u
a
lizatio
n
classif
i
catio
n
i
n
cloud co
m
p
u
ting
envi
ronm
ent. Virtualization
techno
logy
manage
s and coordinates the
accesses from the resourc
e
pool.
Virtu
a
lizatio
n
h
e
lp
s th
e CSPs to
ov
erco
m
e
co
m
p
o
s
ite
wo
rkl
o
ads
,
a
n
d diffe
rent so
ft
wa
re arc
h
itecture.
They
have
di
sc
usse
d
ab
out
t
h
e
vi
rt
u
a
l
i
zat
i
on cl
assi
fi
cat
i
on a
n
d t
h
ei
r w
o
r
k
i
n
g
p
r
i
n
ci
pl
e i
n
t
h
e
pa
per
.
Xi
ao,
So
n
g
, a
n
C
h
e
n
,
20
1
3
[1
5]
, ha
ve
des
c
ri
be
d t
h
e t
ech
ni
q
u
e o
f
al
l
o
c
a
t
i
ng dat
a
ce
nt
er res
o
u
r
ces
t
h
r
o
u
g
h
vi
rt
ua
l
i
zat
i
on t
echn
o
l
ogy
. Th
ey
have i
n
t
r
od
uce
d
t
h
e i
d
ea of “s
kew
n
ess” f
o
r m
easurem
ent
of t
h
e
u
n
e
v
e
n
n
e
ss o
f
resource u
tiliz
atio
n
o
f
a
server
in
m
u
ltid
i
m
en
sion
al way. Th
ey h
a
v
e
also
d
e
v
e
l
o
p
e
d
a way
by
whi
c
h t
h
e
o
v
e
r
al
l
cons
um
pt
i
on
of
ser
v
er
res
o
u
r
ces ca
n
be i
m
prove
d.
K
a
rt
h
i
ck
, Ram
a
raj, and
Sub
r
a
m
an
ian
,
2
014 [16
]
,
h
a
v
e
pro
p
o
s
ed
MQS (Mu
lti Q
u
eu
e
Sch
e
d
u
ling)
alg
o
rith
m
wh
ich
aim
s
to
m
i
n
i
mize th
e co
st o
f
bo
th
on-d
e
m
a
n
d
requ
irem
en
ts and
reserv
ed
p
l
an
s with
the h
e
lp
of gl
obal
sc
he
dul
e
r
. Gl
o
b
al
sche
dul
er i
n
t
e
nds t
o
s
h
are t
h
e p
h
y
s
i
cal
resou
r
ces t
o
i
t
s
m
a
xim
u
m
l
e
vel
.
The
pr
o
pose
d
al
g
o
r
i
t
h
m
uses t
h
e t
echni
que
of cl
ust
e
ri
n
g
t
h
e t
a
s
k
s de
pe
ndi
ng
u
p
o
n
t
h
e
bu
rst
t
i
m
e. Thi
s
pape
r
al
so
reduces t
h
e c
h
ances
of fragm
e
ntation
and
al
so m
i
nim
i
zes the starvation
problem
.
Yan
g
,
K
w
o
n
,
et
al
.,
20
1
4
[
17]
,
ha
ve i
n
t
r
od
uce
d
t
h
e
t
e
chni
que
s
whi
c
h
req
u
i
r
es
co
m
p
il
er co
de
analysis. T
h
is
proce
d
ure m
i
nimizes the trans
f
erred data si
z
e
with t
h
e
help of c
h
anging t
h
e hea
p
objects. They
h
a
v
e
d
i
scu
ssed th
e p
r
o
c
ed
ure o
f
co
st cu
tting
techn
i
qu
es fo
r
d
y
n
a
m
i
c ex
ecu
tio
n
i
n
cloud
. Th
eir resu
lt
sh
ows
that re
duce
d
si
ze affects
both
the tra
n
sfe
r
time an
d e
x
ecut
i
o
n
of
fl
oa
di
n
g
i
n
an e
ffi
ci
e
n
t
m
a
nne
r.
Pal
an
d Pat
t
n
ai
k, 2
0
15 [
1
8]
, h
a
ve pre
s
ent
e
d t
h
e m
i
nim
i
zat
i
on of a
v
era
g
e w
a
i
t
i
ng t
i
m
e
usi
ng J
o
h
n
s
on
sequenci
ng algorithm
.
When
a huge
nu
m
b
er of requ
ests arriv
e
, they h
a
v
e
t
o
wait fo
r allo
catio
n
.
Th
is situatio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A Si
mul
a
tion-base
d A
p
proac
h
to
Optimize
the Execution
Ti
me and Minimi
zation of
…
(Souvik P
a
l)
74
5
i
n
creases
wai
t
i
ng t
i
m
e and q
u
e
ue l
e
n
g
t
h
.
Th
ey
have m
i
ni
mized ave
r
age
waiting tim
e
in
the system
and in the
q
u
e
u
e
b
y
m
ean
s of
q
u
e
u
i
ng
m
o
d
e
l
with
fin
ite cap
acity and
m
u
l
ti-serv
er cap
a
b
ility.
C
a
l
e
ro an
d Ag
uad
o
,
20
1
5
[1
9
]
, have pre
s
ent
e
d a
m
oni
t
o
ri
n
g
archi
t
ect
u
r
e conce
r
ned t
o
t
h
e C
SP a
n
d
cloud use
r
. T
h
is architecture
allows the
use
r
to cust
o
m
ize
th
e m
e
trics. Th
e clou
d
p
r
ov
id
ers can
easily track
t
h
e ser
v
i
ces
u
s
ed
by
t
h
e
u
s
ers. C
SPs
ha
ve
use
d
Ada
p
t
i
v
e di
st
ri
but
e
d
m
oni
t
o
ri
ng
t
echni
q
u
e
w
h
i
c
h i
s
im
ple
m
ented in cloud infra
structure.
2.
1. Ob
jecti
v
e
of
the
St
ud
y
In
t
h
e
p
r
evi
ous
sect
i
o
n
,
we
ha
ve
di
scus
sed
t
h
e C
l
o
u
d
Si
m
sim
u
l
a
t
o
r,
Que
u
i
ng m
odel
,
Vi
r
t
ual
i
zat
i
on,
avera
g
e
waiting tim
e. W
h
e
n
a huge
num
b
er of re
quests
h
a
ve c
o
m
e
, t
h
e user
re
que
st
s
have t
o
wai
t
i
n
t
h
e
que
ue a
nd i
n
t
h
e sy
st
em
. In
t
h
i
s
pa
per,
w
e
have
use
d
C
l
ou
dSi
m
simul
a
t
o
r
ver
s
i
o
n
3.
0 t
o
fi
n
d
o
u
t
t
h
e
m
i
nim
u
m
execut
i
o
n
t
i
m
e
wi
t
h
re
spect
t
o
nu
m
b
er of C
P
Us
.
Aft
e
r
get
t
i
ng
t
h
e m
i
nim
u
m
num
ber
of
C
P
U a
n
d
execut
i
o
n t
i
m
e
, que
ui
n
g
m
odel
has been i
m
pl
em
ent
e
d t
o
reduce the a
v
erage waiting tim
e. In
th
is
m
a
n
u
s
crip
t
,
M/M/c an
d
M/M/c/K q
u
e
u
i
ng
system h
a
s been
u
s
ed
to
com
p
are th
e av
erag
e waiting
time in
th
e q
u
e
u
e
and
also in the syst
e
m
.
3.
SIMULATION WORKFL
OW
In
t
h
i
s
sect
i
o
n,
we
ha
ve
bri
e
f
l
y
di
scusse
d
o
u
r
si
m
u
l
a
ti
on
wo
rk
-fl
ow
as s
h
o
w
n as
fi
g
u
r
e
[
1
]
,
an
d
w
e
are goi
ng
t
o
de
scri
be o
u
r
seq
u
e
n
ce
diagram
stepwise as
follows:
STEP
1
:
Cloud u
s
er
send
s th
e
j
o
b
r
e
q
u
e
st t
o
t
h
e
U
s
er In
terf
ace
STEP
2:
User Interface
a
n
alyzes the
request a
ccording t
o
the
Service
Le
vel Agreem
ent
(SLA).
STEP
3:
User Interface
assigns the tas
k
s t
o
cl
oud
broker.
STEP
4
:
Cloud br
ok
er d
i
v
i
d
e
t
h
e task
s in
t
o
same sized
clou
dlets.
STEP
5
:
Cloud br
ok
er sen
d
s
th
e cloud
lets to
V
i
r
t
u
a
l Mach
ine Man
a
g
e
r
(VMM)
.
STEP
6
:
Each
d
a
ta cen
ter en
tity reg
i
sters
with
th
e Clou
d In
fo
rm
atio
n
Serv
i
ce (CIS)
reg
i
stry.
STEP
7
:
Cloud
b
r
o
k
er
send
s th
e resour
ce req
u
e
st to
Cloud
Inf
o
r
m
atio
n
ser
v
ice
(
C
IS)
.
Th
en
Clou
d bro
k
e
r
co
nsu
lts the C
I
S t
o
o
b
t
ain the list o
f
resou
r
ces wh
ich
can offe
r infra
structure s
e
rvices
that m
a
tches user’s
har
d
ware a
n
d s
o
ft
ware
re
quire
m
e
nts.
STEP
8
:
Th
e
clo
u
d
b
r
ok
er gets in
fo
rm
atio
n
ab
ou
t t
h
e
availab
ility o
f
the d
a
tacen
t
er an
d resou
r
ces fro
m
th
e
CIS.
STEP 9:
Vi
rt
u
a
l
m
achi
n
e
m
a
nage
r (V
MM
)
creates the
virt
ual m
achine.
STEP
1
0
:
Data Cen
t
er en
tity in
vok
es Upd
a
te VM Pro
cessi
ng
fo
r ev
ery
h
o
st th
at
m
a
n
a
g
e
s it as p
r
o
c
essin
g
o
f
t
a
sk
uni
t
s
i
s
ha
ndl
e
d
by
res
p
e
c
t
i
v
e VM
s.
S
o
t
h
ei
r
pr
og
ress
m
u
st
be co
nt
i
n
uo
usl
y
up
dat
e
d
an
d m
oni
t
o
re
d
.
STEP 11
: At th
e ho
st lev
e
l, in
vo
catio
n
o
f
Upd
a
te VM
P
r
oces
si
n
g
t
r
i
g
g
e
rs an U
p
dat
e
C
l
ou
dl
et
Proce
ssi
n
g
m
e
t
hod t
h
at
di
rect
s eve
r
y
Vi
rt
ual
M
achi
n
e
(VM
)
t
o
u
pdat
e
i
t
s
t
a
sk u
n
i
t
st
at
us (
f
i
n
i
s
h
,
sus
p
en
d,
exec
ut
i
n
g
)
with
th
e Data cen
t
er en
tity.
STEP
12: VM
analyze the a
p
proxim
a
t
e execution
ti
m
e
an
d
send
s to
host
machine.
STEP
13: Host
m
achine analyzes
sm
al
lest ti
me to
n
e
x
t
ev
en
t.
STEP
1
4
:
Dat
a
cent
e
r
p
r
ovi
des t
h
e i
n
f
o
rm
at
i
on a
b
out
t
h
e exec
ution time and diff
erent res
o
urces
suc
h
a
s
O
p
er
ating
Syst
e
m
, V
MM u
s
ed
, R
A
M size,
MI
PS, num
b
er
of cloudlets,
num
b
er of
CPU, stora
g
e ca
paci
ty etc.
STEP
15
: Requ
est for ex
ecu
t
i
o
n
of th
e cloud
let is sen
t
t
o
t
h
e
v
i
rtu
a
l m
ach
in
e
b
y
VMM.
STEP
16
: Cloud
let is b
e
i
n
g execu
ted
i
n
t
h
e
VM.
STEP
17
: V
M
send
s th
e ex
ecu
t
ed cl
oudlets t
o
the
VMM.
STEP
18
: After co
m
p
letin
g
the ex
ecu
tion
,
VM
releases
the
reso
u
r
ces fo
r f
u
rt
her use
.
STEP
19: CIS
updates t
h
e re
gistry according to
the
inform
a
tion se
nt
by dat
a
center.
STEP
2
0
: VM
M
fu
rthe
r
pass
es the e
x
ec
uted cloudlets
t
o
cl
ou
d
b
r
o
k
er
.
STEP
21
: Cloud
b
r
ok
er co
m
b
in
es all th
e ex
ec
uted cl
oudlets t
oget
h
er to
form
the task.
STEP 22:
Cloud broker se
nds the
com
p
leted task
to the
Use
r
Interface.
STEP 23: After com
p
letion of the task
, Use
r
Inte
rface ca
n either expi
re the session or make anot
her renewal
req
u
est.
STEP
24: If se
ssion is e
xpire
d, t
h
en User
In
terface se
nds t
h
e e
x
ecute
d tas
k
to the cl
oud
user.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
74
3 – 7
5
0
74
6
Figure 1.
Se
quence Diag
ram
of
the
w
o
r
k
fl
o
w
3.
1.
Si
mul
a
ti
o
n
Res
u
l
t
In th
is sectio
n,
we are
go
ing
t
o
test t
h
e ex
ecu
tio
n tim
e with
resp
ect to th
e
n
u
m
b
e
r
of CPUs. Th
e tests
were c
o
n
d
u
ct
e
d
o
n
a 3
2
-
bi
t
Int
e
l
C
o
re i
5
m
achi
n
e ha
vi
n
g
2
.
6
0
GHz a
nd
3
GB
R
A
M
ru
nni
ng
wi
nd
o
w
s 7
Pro
f
essi
onal
a
n
d
JD
K
1.
6.
T
h
e m
a
i
n
goal
o
f
o
u
r
t
e
st
s i
s
to evaluate t
h
e e
x
ecution tim
e
whe
n
t
h
e
num
bers
of
C
P
U co
res as VM
param
e
t
e
rs whi
c
h
vari
es
fr
om
1 t
o
8. Accor
d
i
n
g t
o
t
h
a
t
vari
at
i
on, h
o
w
t
h
e exec
ut
i
o
n t
i
m
e
is ch
an
g
i
n
g
and
we are
g
o
i
n
g
to
find
in an op
ti
m
i
zed
situ
atio
n
wh
ere ex
ecu
tio
n ti
m
e
is less.
We h
a
ve
use
d
Ecl
i
p
se Jav
a
EE I
D
E f
o
r
Web
De
vel
o
p
e
rs,
Versi
on:
J
u
n
o
Se
r
v
i
ce R
e
l
ease 2 a
n
d
C
l
ou
dSi
m
versi
on
3.
0 f
o
r
si
m
u
l
a
t
i
on
p
u
r
pos
e. I
n
o
u
r
ex
pe
ri
m
e
nt
al
set
up, t
h
i
s
si
m
u
l
a
t
i
o
n w
o
rks
o
n
l
y
wh
e
n
sim
u
lation is pause
d
for 5 sec
and this sim
u
lation creates
a
datacenter B
r
oker
dynam
i
cally and also s
u
bject to
ot
he
r c
onst
r
ai
n
t
s t
h
i
s
Si
m
u
l
a
t
i
on
i
s
do
ne.
The sim
u
lation environm
ent c
onsists of two hosts;
each
hos
t
has been m
o
deled to have
1000 MIPS,
16
GB of RAM
m
e
m
o
ry, 1 TB of stora
g
e and
10 num
b
er
s of
VMs each of which
has
been m
odeled
to ha
ve
500
M
I
PS, 1 GB of
R
A
M, and
10 GB of im
age
size.
A
datacenter is c
r
eated, wh
ich
has t
h
e c
h
aract
eristics
l
i
k
e x8
6 o
f
arc
h
i
t
ect
ure, Li
nu
x as o
p
erat
i
n
g
sy
st
em
, Xen as VM
M
.
Sim
u
l
a
t
i
on uses
VM
Al
l
o
cat
i
on
Si
m
p
l
e
as
VM Allocatio
n Po
licy,
wh
ich
ch
oo
ses, as th
e ho
st for
VM
, t
h
e
h
o
st with less pro
c
essing
ele
m
en
ts in
u
s
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A Si
mul
a
tion-base
d A
p
proac
h
to
Optimize
the Execution
Ti
me and Minimi
zation of
…
(Souvik P
a
l)
74
7
Fi
gu
re
2.
Eval
uat
i
o
n
o
f
E
x
ec
ut
i
o
n
Ti
m
e
In the fig
u
r
e [2
]
,
we want to sho
w
t
h
e exec
ut
i
on t
i
m
e
i
n
our
experi
m
e
nt
al
si
t
u
at
i
on.
Whi
l
e chan
gi
n
g
the Num
b
er of CPUs, t
h
e execu
tio
n ti
m
e
fo
r t
h
e u
s
er-requ
e
st v
a
ries in
a rando
m
o
r
d
e
r. In
th
is case,
we can
say th
at if we
u
s
e
4
nu
m
b
ers
o
f
CPUs, t
h
e ex
ecu
tion
tim
e
will b
e
m
i
n
i
ma
l. Th
erefo
r
e, clo
u
d
u
s
er can
get th
e
serv
ices with
min
i
m
a
l t
i
m
e.
4.
QUEU
IN
G M
O
DEL
F
O
R
CLOU
D CO
MP
UTIN
G
E
NVI
RO
N
M
E
N
T
Qu
eu
ing
system is
a
math
e
m
atical
m
o
d
e
l fo
r an
alysis o
f
waitin
g
lin
e.
W
a
itin
g
lin
es o
r
qu
eu
es h
a
v
e
occurre
d
whe
n
service
dem
a
nd
goes bey
o
nd t
h
e capacit
y
o
f
th
e serv
i
ce p
r
ov
id
er. Th
e qu
eu
ing
mo
d
e
l is
basi
cal
l
y
di
scu
ssed
by
s
p
eci
f
y
i
ng [
2
0]
[
21]
arri
val
a
n
d se
r
v
i
ce p
r
o
cess,
n
u
m
b
er o
f
se
rve
r
s a
nd t
h
e m
a
xim
u
m
capaci
t
y
of t
h
e sy
st
em
. In t
h
i
s
pa
pe
r,
we
conce
n
t
r
at
e
d
o
n
M
/
M
/
c que
u
i
ng m
odel
an
d
M
/
M
/
c/
K que
ui
n
g
m
odel
.
Assum
i
ng t
h
at
t
h
e re
que
st
s com
e
to t
h
e serve
r
at
Poi
sso
n di
st
ri
but
i
o
n rat
e
an
d t
h
e pr
ocess
t
i
m
e
i
s
tak
e
n
as ex
pon
en
tial d
i
st
ribu
tio
n.
It is also
assu
m
e
d
th
at all th
e pro
c
esses are non
-pre em
p
tiv
e. In th
ose
que
ui
n
g
m
odel
,
c den
o
t
e
s t
h
e
num
ber o
f
ser
v
ers an
d K i
n
de
nt
i
f
i
e
s t
h
e pl
ac
es fo
r m
a
xim
u
m
capaci
t
y
. So, (K
-
c) is the
que
u
e
capacity.
W
e
have c
o
nsid
ere
d
five
places
of
maxim
u
m
capacity.
Accord
ing
to
t
h
e Kend
al’s
no
tatio
n
[22
]
, av
erag
e arriv
a
l
rate will b
e
]
[
1
E
, whe
r
e
τ
=
Inte
r-
arri
val
t
i
m
e
and E
[
τ
] is
d
e
no
t
e
d
as th
e
av
erag
e
o
r
m
ean
In
t
e
r-arriv
a
l
ti
m
e
. Serv
ice
rate
will b
e
)
(
1
S
E
,
whe
r
e S is denoted as service
time of the custom
er a
nd E(S) identifies as
avera
g
e service time. To
make a
stab
le system
,
an
eq
u
ilibriu
m
co
nd
itio
n is to
b
e
m
a
in
tain
ed
wh
ere t
h
e
u
tili
zatio
n
factor
1
.
5.
NU
MER
I
C
A
L
AN
AL
YSIS
In
t
h
is section
,
we
h
a
v
e
briefl
y d
i
scu
ssed
numerical an
alysi
s
u
s
ing
qu
eu
i
n
g
m
o
d
e
l.
W
e
hav
e
in
itially
t
a
ken a
v
era
g
e
arri
val
rat
e
an
d
avera
g
e se
rvi
c
e rat
e
as sh
ow
n i
n
t
h
e t
a
bl
e [
1
]
.
Usi
ng M
/
M
/
c que
ui
n
g
m
odel
an
d
M
/
M
/
c/
K que
u
i
ng m
odel
,
we
have
a c
o
m
p
ari
s
on
st
u
d
y
of
w
a
i
t
i
ng t
i
m
e.
Tab
l
e
1
.
In
itial Param
e
ter (Averag
e
ar
riv
a
l
rate an
d Serv
ice
rate [2
3
])
µ
20
40
60
70
120
122
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
74
3 – 7
5
0
74
8
Acco
r
d
i
n
g t
o
que
ui
n
g
sy
st
e
m
,
i
t
has bee
n
den
o
t
e
d av
e
r
age num
b
er of cust
om
er
s in the system
, a
v
era
g
e
n
u
m
b
e
r
o
f
cu
st
o
m
ers in
th
e
qu
eu
e, av
erag
e
waitin
g
tim
e in
th
e system
, an
d
av
erag
e
waitin
g ti
m
e
in
th
e
q
u
e
u
e
as Ls, Lq,
W
s
, and
Wq res
p
ectively. The tables [2-3
] sho
w
t
h
e co
m
p
ar
ison
stud
y u
s
i
n
g
d
i
ff
er
en
t qu
eu
i
ng
m
odel
.
Tabl
e
2. R
e
s
u
l
t
s
wi
t
h
M
/
M
/
c Quei
ng
M
o
del
Lq
Ls
Wq
Ws
λ
= 20
0.
0003
0.
5003
0.
0000
1
0.
0250
λ
= 60
0.
0033
0.
8605
0.
0000
6
0.
0144
λ
= 120
0.
0063
0.
9900
0.
0000
5
0.
0083
Tab
l
e
3
.
Resu
lts with M/M/c/
K
Qu
ei
n
g
Model
Lq
Ls
Wq
Ws
λ
= 20
0.
0002
0.
5001
0.
0000
0
0.
0250
λ
= 60
0.
0021
0.
8575
0.
0000
1
0.
0143
λ
= 120
0.
0036
0.
9837
0.
0000
0
0.
0082
Acco
r
d
i
n
g t
o
t
h
e n
u
m
e
ri
cal
resul
t
s
we ha
ve
di
scusse
d t
h
e
com
p
ari
s
on st
udy
re
ga
rdi
ng
Lq, Ls
,
Wq, a
nd
Ws
.
Th
e
fo
llowing
Fig
u
res
[3
-6
] sh
ow t
h
at th
e av
erag
e
nu
m
b
er of cu
sto
m
ers an
d th
e
av
erage waitin
g tim
e
in
th
e
que
ue a
n
d i
n
t
h
e sy
st
em
can b
e
m
i
nim
i
zed us
i
ng M
/
M
/
c/
K
r
a
t
h
er t
h
an
M
/
M
/
c
m
odel
.
Figure 3.
Graph Analyzing Avera
g
e num
b
er
of
custom
er in t
h
e queue
(L
q
)
Figure 4.
Graph Analyzing Avera
g
e num
b
er
of
custom
er in t
h
e syste
m
(L
s
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A Si
mul
a
tion-base
d A
p
proac
h
to
Optimize
the Execution
Ti
me and Minimi
zation of
…
(Souvik P
a
l)
74
9
Fig
u
re 5
.
Graph
An
alyzing
Av
erag
e waitin
g
ti
m
e
in
the que
u
e
(W
q
)
Fig
u
re 6
.
Graph
An
alyzing
Av
erag
e waitin
g
ti
m
e
in
the system
(W
s
)
6.
CO
NCL
USI
O
N
R
a
pi
d usa
g
e o
f
Int
e
r
n
et
o
v
er
t
h
e gl
obe, C
l
ou
d C
o
m
put
i
n
g has pl
ace
d i
t
s
el
f i
n
every
fi
el
d of IT
industry. T
h
e recent efforts to m
a
ke
cloud
com
puting tec
h
nologies bett
er, which incl
ude
s exec
uting tim
e
.
There
f
ore,
we
have c
o
ncent
r
at
ed o
n
si
m
u
l
a
t
i
on-
base
d ap
p
r
oac
h
es
whi
c
h
hel
p
t
h
e cl
ou
d de
vel
o
pe
rs t
o
t
e
st
perform
a
nce of their service deliver
y po
licies an
d
also
their ex
ecu
tion
time an
d
av
erage waitin
g
ti
m
e
so
th
at
th
e clo
u
d
serv
i
ce p
r
ov
id
ers can
pro
v
i
d
e
b
e
tt
er qu
ality serv
i
ces with
m
i
n
i
m
u
m ex
ecu
tion
ti
m
e
. At th
e en
d
of
o
u
r
w
o
rk
,
w
e
can
co
n
c
l
u
d
e
th
at our
seq
u
e
nce d
i
agr
a
m
an
d
o
u
r
sim
u
lati
o
n
r
e
su
lts m
a
y
h
e
lp to
g
r
ow
i
n
cloud
i
n
fra
st
ruct
ure
i
n
s
u
r
g
e
o
f
fa
st
-
g
r
o
wi
n
g
usage
of
i
n
t
e
r
n
et
am
ong
t
h
e
pe
opl
e
.
REFERE
NC
ES
[1]
F.B. Shaikh and
S Haider, “Security
threats in cloud computing”, “
6
th
International IEEE con
f
er
ence on Int
e
rne
t
Technology and
secured transaction”
, 11-14 D
e
cember 2012, pp.
214-219.
[2]
V. Sarath
y
,
P.
Naray
a
n and R
.
Mikkilinen
i, “Next ge
ner
a
tion
cloud com
puti
ng archit
ecture-
e
nabl
ing real-ti
m
e
d
y
namism for shared distribu
ted
ph
y
s
ical infr
astr
ucture”,
in
the P
r
oceedings
of 19
th IEEE
In
ternational Workshops
on Enabling Technologies: Infra
s
tructures
for Collaborative En
terprises (
W
ETI
C
E’10)
, Larissa, Greece, 2010
, p
p
.
48-53.
[3]
S. Pal and P.K.
Pattnaik
,
“Efficient ar
chit
ectural Framework of Cloud Computing”, in “
International Journal o
f
Cloud
Computin
g and Services S
c
ien
c
e (
I
J-CLOS
ER)
”
, Vol. 1
,
N
o
. 2
,
2012
, pp
. 6
6
-73.
[4]
V. I. Munteanu
,
C. Mindruta
an
d T. Fortis, “Ser
vice brok
ering in cloud govern
ance”,
in
the Pr
o
ceed
ings
of 14th
International Symposium on Sy
mbolic and Numer
i
c
Algorithms
fo
r Scien
tific Computing
, IEEE, 20
12, pp
. 497-504
.
[5]
S. Sundareswaran, A. Squicc
iarini and
D. Lin, “A Brokerage-B
ased
Approach fo
r Cloud Service
Selection”,
in
th
e
Proceed
ings of 2
012 IEEE
Fif
t
h I
n
terna
tional Co
nferenc
e
on Clo
ud Computing
, I
EEE Com
puter
Societ
y
,
2012, p
p
.
558-565.
[6]
Stella G
a
tziu Grivas, Tr
ipath
i
Uttam
Kumar, Holger Wache, “Cloud Broker: Br
in
ging Intellig
ence into the Cloud
”,
IEEE 3rd International Con
f
eren
ce on
Cloud
Co
mputing
, I
E
EE,
2010, pp
.544~545.
[7]
T. S
o
wjan
ya
,
et al.
, “Th
e
Queuin
g Theor
y
in clou
d Computing to Reduce
the Waiting Time”,
Inter
national Journa
l
of Computer S
c
ience and
Engineering Techno
log
y
(
I
JCSET)
, Vol. 1, No. 3
,
2011
,
pp. 110-112
.
[8]
R. Bu
yy
a, R
.
Ranjan
and R
.
N. C
a
lh
eiros, “Modeling and
simula
tion of scalable C
l
oud computing
environments and
the CloudSim
t
oolkit:
Chal
leng
es and opportun
ities”,
in th
e Pr
oceed
ings of In
t
e
rnational Con
f
erence
on
High
Performance Co
mputing
&
Simu
lation (
H
PCS '0
9)
,
21-24 June 2
009
,
pp
. 1-11
.
[9]
P. Xing,
et a
l
.
, “Understanding P
e
rformance Inter
f
erence of
I/O
Workload in Vir
t
ualized
Cloud
Environments”,
in
the Proceedings of
IEEE
3r
d International Conference
on C
l
oud C
o
mputing (
C
LOUD)
,
5-10 July
2
010
, pp
.51-58.
[10]
H. Khazaei, J
.
Misic and V.B. Mi
sic, “Modelling of Cloud C
o
mputi
ng Centers Using M/G/
m Queues”,
in t
h
e
Proceed
ings of 31st IEEE In
t
e
rnational Conf
erenc
e
on
Distributed
Computing Systems Workshops
(
I
CDCSW
)
,
20-
24 June, 2011, p
p
. 87-92
.
[11]
R.N. Calh
eiros,
et al.
, “CloudSim: A Toolkit for Modeling and Simula
tion of Clo
ud Computing Environments and
evalu
a
tion
of R
e
source Provisio
n
ing Algorithms”,
So
ftware:
Pra
c
tice and
E
x
per
i
ence
(
S
PE)
, Jan
u
ar
y
2011, Vol.
41, No. 1, pp. 23
-50.
[12]
J. Spilln
er,
et al.
, “A Highly
-
Virtualising
Cloud
Resource Brok
er
”,
in
the
Proce
e
d
ings of I
EEE
F
ifth
Internat
iona
l
Conferenc
e
on
Utilit
y and
Clou
d Computing (
U
CC)
,
5-8 Nov., 2
012, pp
. 233-23
4.
[13]
H. Khazaei, J.
Misic and V.B
.
Misi
c, "Perfor
m
ance Analy
s
is
of Cloud
Computing Cen
t
ers
Using M/G/m/m+r
Queuing S
y
s
t
em
s
”
,
IEEE Transa
c
tions on
Parallel and Distributed Systems
, May
2012, Vol. 23, No. 5, pp. 936-
943.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
74
3 – 7
5
0
75
0
[14]
S.
Pa
l a
nd P.
K.
Pa
ttna
i
k,
“Cla
ssi
fication of Virtu
a
lization Enviro
nment for Cloud
Computing”, in
Indian Journal o
f
Scien
c
e and Technology (
I
JST)
”
, Vol. 6, Issue 1,
Januar
y
2013
, p
p
. 3965~3971.
[15]
Z. Xiao
, W. Song and Q. Chen
, “D
y
n
amic Res
ource A
llo
catio
n Using Virtual Machines for
Cloud Computing
Environment”,
I
EEE Transactio
n on Parallel
an
d Distributed
Systems
, Vol. 24,
No. 6, June 2013
, pp
. 1107-1117
.
[16]
A.
V.
Karthick,
et al.
, "An Efficient Multi Queu
e Job Scheduling for Cloud Co
m
puting”,
in the
Pr
oceedings
of
World Congress on
Computing
and Communication Technolog
ies (
W
CCCT)
,
IE
EE, Feb
.
27- March 1, 2014
, pp
.
164-166.
[17]
S. Yang,
et al.
, “Techn
i
ques to Minimize State
Transfer Costs fo
r Dy
n
a
mic Execution Offload
i
n
g
in Mobile Cloud
Computing”,
I
E
EE Transactions
Mobile Computing
, Vol. 13
, No.
11, November 2
014, pp
. 2648-2
659.
[18]
S. Pal and P.K.
Pattnaik
,
“Adaptati
on of Johnson Sequencing for
Job Schedu
ling to Minimize the
Average Waitin
g
Time in Cloud
Computing Environment”,
in “
Journal of Engineering Scie
n
ce and Technology (
J
ESTEC)
”
,
Tay
l
or’s Univ
ersity
. (Article in
Press).
[19]
J.M.A. Calero and J.G. A
guado, “Mon
PaaS: An
Adaptive Monitoring Pl
atform as a Service for Cloud Computing
Infrastructur
es and Services”,
IEEE Transactions
Service Computing
, Vol. 8
,
No
.
1, Januar
y
2015
, pp. 65-78.
[20]
L.
Tad
j
,
“
W
a
itin
g in
line
”
,
Po
ten
tial,
IE
EE
, Vol.
14, No. 5, 1996,
pp. 11-13
.
[21]
C. Cheng, J
.
L
i
and Y. W
a
ng, "An energ
y
-s
av
i
ng tas
k
s
c
heduli
ng s
t
rateg
y
bas
e
d on vacat
ion queuing theor
y
in
cloud computing
"
,
Tsinghua Science and
Techno
logy
, Vol. 20
, No
. 1
,
2015, pp. 28-
39.
[22]
D.G. Kendall
. “
S
om
e P
r
oblems
in Theor
y
of Queues
”,
J. Ro
y. Stat. Soc., Series B
, Vol. 13, No.
1, 1951, pp. 151
-
185.
[23]
L. L
i
, "An Opti
m
i
stic Differen
t
iat
e
d Servic
e Jo
b Sche
duling S
y
stem for Clou
d Computing Service Users an
d
Providers, “
Third International
Conference on
Multimedia
and
Ubiquitous En
gineering (
M
UE '09)
”,
4-6 Jun
e
2009, pp
. 295-2
99.
BIOGRAP
HI
ES OF
AUTH
ORS
Souvik Pal, Member of
CSTA/ACM
,
USA,
Member of
IAENG
,
Hong Kong, Member of
IACSIT
,
Singapore, is Assistant Professor at
the D
e
partment
of Computer
Science and
Eng
i
neer
ing,
Elitte
College o
f
Engineering
,
Kolkata and Ph.D. Rese
arch Scholar at
KIIT
University
, Bhubaneswar.
He has published var
i
ous R
e
search
Papers
in
p
eer-r
eviewed Intern
ation
a
l Journals an
d
Conferences. His resear
ch area includes C
l
oud C
o
mputing.
Dr. P
r
asant Kum
a
r P
a
ttnaik
,
S
e
nior Mem
b
er I
EEE (US
A
), F
e
llow IETE
, is P
r
ofessor at the
School of Computer Eng
i
neering
,
KIIT Univ
ersity
, Bhuban
e
swar. He has more th
an a d
e
cade of
tea
c
hing
and r
e
s
earch
exper
i
en
ce.
Dr. P
a
ttna
i
k
has published
numbers of Research
Papers in
peer-rev
i
ewed I
n
ternational Journals and conf
er
ences. His
areas of inter
e
st include Mobile
Computing and
Cloud Computin
g.
Evaluation Warning : The document was created with Spire.PDF for Python.