Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
12
,
No.
3
,
Decem
ber
201
8
, p
p.
1
159
~
117
0
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v1
2
.i
3
.pp
1159
-
117
0
1159
Journ
al h
om
e
page
:
http:
//
ia
e
s
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
QoS Ma
nagemen
t in WS
N
-
MC
N Conve
rgence Net
work Usi
ng
Priorit
y Based T
raffic M
odels
An
it
a
Sw
ain
,
Arun Kum
ar
Ray
Kali
nga
Instit
u
te of
Industri
al Te
chnol
og
y
,
(Dee
m
ed
to
b
e
Unive
rsit
y
)
,
Bhub
ane
s
war,
Ind
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ma
r
1,
2018
Re
vised
Ma
y
24
, 2
018
Accepte
d
Aug
3
0
, 201
8
Te
chno
logi
c
al advance
m
ent
s
an
d
inc
r
ea
sed
dev
e
lopments
lead
to
inc
r
ea
se
th
e
deve
lopment
of
hand
-
hel
d
d
evi
c
es
which
enc
our
age
th
e
growth
of
Mac
h
ine
-
to
-
Mac
hin
e
(M
2M)
comm
unic
at
ion
.
In
thi
s
cont
ex
t
d
evi
c
es
be
long
to
het
ero
g
ene
ous
n
et
work
r
equi
re
a
comm
on
platfor
m
to
m
eet
th
e
ch
al
l
enge
s
of
dat
a
ce
n
tric
wire
le
ss
services
and
appl
i
cations.
To
sati
sf
y
th
e
n
ee
d
s
of
M2M
comm
unic
at
ion,
Mobile
C
el
lu
l
ar
Ne
twork
(M
CN)
and
W
ire
l
ess
Sensor
Network
(W
SN
)
conve
rg
ed
to
a
single
p
la
tform
.
In
th
is
pape
r
t
wo
priorit
y
m
odel
s
has
been
proposed
wit
h
non
pre
empti
ve
prior
ity
and
pre
emptiv
e
priori
t
y
to
an
aly
z
e
th
e
data
tra
ff
ic
at
MA
C
lay
e
r
of
the
conve
rg
ed
n
e
twork.
The
per
form
anc
e
m
a
tri
c
es
ar
e
deter
m
ine
d
to
m
ai
nt
a
in
QoS
in
te
rm
s
of
red
u
ce
d
dropping
and
bl
ocki
ng
proba
b
il
i
t
y
,
wai
ti
ng
ti
m
e
in
the
queu
e
etc.
Fina
lly
,
th
e
proposed
m
odel
s a
re
compar
ed
in
te
rm
s of
QoS
f
a
ct
ors.
Ke
yw
or
d
s
:
Bl
ock
in
g Pr
obabili
ty
Conver
ge
nce
Droppi
ng Pro
ba
bili
ty
MCN
WSN
Copyright
©
201
8
Instit
ut
e
o
f Ad
van
ce
d
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
An
it
a S
wain
,
Kali
ng
a
Insti
tu
te
o
f
Industrial
Tech
no
l
og
y,
(D
eem
ed
to
be Un
i
ver
sit
y),
Bhuba
nes
war
,
In
dia
,
Em
a
il
:
anita
.sw
ai
n070
3@g
m
ai
l.co
m
1.
INTROD
U
CTION
Netw
orkin
g
te
chnolo
gies
al
l
ow
a
la
r
ge
nu
m
ber
of
div
e
rsity
of
dev
ic
es
s
uch
as
m
ob
il
e
ph
ones,
la
pt
ops,
TVs,
perso
nal
com
pu
te
rs,
s
pe
aker
s
,
li
gh
ts
a
nd
el
ect
ronic
ap
pliances
to
be
c
onnected
i
n
a
s
ea
m
le
ss
m
ann
er.
T
his
par
a
dig
m
create
s
a
po
s
sibil
it
y
of
m
achine
-
to
-
m
achine
(M
2M)
c
omm
un
i
cat
i
on
an
d
e
xc
hange
of
in
f
orm
at
ion
.
M2M
com
m
un
ic
at
ion
s
is
cha
racteri
zed
by
l
ow
powe
r,
l
ow
cost,
a
nd
le
ss
hu
m
an
in
vo
l
ve
m
ent
[1
,
2]
.
At
this
po
i
nt,
the
WSN
is
go
i
ng
t
o
s
at
isfy
the
cha
r
act
erist
ic
s
of
M2M
c
omm
un
ic
at
ion
us
in
g
st
at
ic
an
d
m
ob
il
e
se
ns
or
nodes
.
Howe
ve
r
t
he
de
plo
yi
ng
st
rategy
of
WSN
de
pends
upon
t
he
c
omm
un
ic
at
ion
sta
nd
a
r
d.
Zig
bee
i
s
one
of
the
m
os
t
ap
pro
pr
ia
te
c
omm
un
ic
at
ion
sta
nd
a
r
d
for
WSN
[
3]
.
Howe
ver
the
m
ob
il
e
se
nsor
nodes
a
re
m
or
e
fl
exible
to
real
tim
e
ap
plica
ti
on
s
t
ha
n
the
sta
ti
c
se
nsor
no
des
[4,
5]
.
I
n
a
n
integ
rat
ed
netw
ork
of
MC
N
an
d
WSN,
t
he
fron
t
sensi
ng
pa
rt
is
WSNs
w
hich
can
be
fle
xib
ly
dep
l
oyed
to
detect
diff
e
ren
t
ty
pe
s
of
s
ens
or
y
data
a
nd
th
e
MC
N
act
as
a
bac
kgr
ound
ne
twork
f
or
dat
a
tra
ns
m
issi
on
.
T
hese
dy
nam
i
c
be
ha
vior
of
WSN
e
nc
oura
ges
to
create
a
platf
orm
fo
r
WSN
a
nd
MC
N
to
c
onve
r
ge
a
nd
e
sta
blish
a
c
omm
on
netw
ork.
B
ecause
MC
N
ha
s
th
e
adv
a
ntage
s
of
l
arg
e
co
ver
a
ge
and
po
werfu
l
use
r
te
rm
inals,
WSN
a
nd
MC
N
c
onverge
nce
is
ind
is
pe
ns
ab
le
f
or
su
pp
or
ti
ng
M
2M
com
m
un
ic
ation
s
[6]
.
T
he
c
onve
rg
e
nce
of
W
CN
a
nd
MS
N
can
al
s
o
benefit
each
ot
her
:
(I)
F
or
WSN,
t
he
MC
N
ca
n
pro
vid
e
opti
m
iz
at
ion
to
prolo
ng
WSN
li
fe
tim
e,
prov
i
de
qual
it
y
of
ser
vice
(Qo
S)
a
nd
i
m
pr
ove
sye
te
m
per
fo
rm
ance
;
(I
I
)
F
or
MC
N,
WSN
ca
n
e
xten
d
the
i
ntell
igent
a
pp
li
cat
ion
ra
ng
e
of
M
CN,
i.e.
WSN
can
pro
vi
de
real
-
ti
m
e
measur
em
ent
res
ults
to
MC
N
use
rs.
M
2M
co
m
m
un
ic
at
ion
is
base
d
on
ub
i
qu
it
ou
s
te
chnolo
gy
of
WSN
a
nd
MC
N,
w
hich
use
s
the
cel
l
ular
s
yst
e
m
as
the
ba
ckbo
ne
of
th
e
net
work.
T
he
tw
o
heter
og
e
ne
ou
s
netw
orks
WSN
an
d
MC
N,
c
om
e
tog
et
her
f
or
the
c
onverge
nc
e
to
s
upport
th
e
data
ce
ntric
s
erv
ic
e
s
and ap
plica
ti
on of M2M c
om
m
un
ic
at
ion
.
Conver
ge
nce
of
WSN
a
nd
MC
N
is
a
n
e
m
erg
in
g
resea
rch
fiel
d
in
w
irel
ess
com
m
un
ic
at
ion
[
7].
The
li
te
ratur
e
exp
l
or
es
fe
w
researc
hes
on
the
issues
of
WSN
an
d
M
CN
co
nver
gence
ne
tw
ork
with
it
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1
159
–
1
170
1160
app
li
cat
io
n
ori
ented
im
ple
m
e
ntati
on
[8
-
11]
.
The
idea
of
c
onve
r
ging
WSN
and
MC
N
an
d
it
s
com
par
ison
wit
h
the
inte
gr
at
e
d
netw
ork
of
W
SN
an
d
MC
N
is
gi
ven
in
[12
]
.
H
oweve
r
t
he
integ
rati
on
of
WSN
a
nd
MC
N
has
been
al
rea
dy
im
ple
m
ented
thr
ough
the
ga
te
way.
I
n
the
integr
at
ed
W
SN
a
nd
MC
N
,
the
a
rch
it
ect
ur
e
i
s
hierar
c
hical
an
d
the
gateway
is
j
us
t
a
data
channel
to
e
xc
hange
in
f
or
m
a
ti
on
bet
ween
t
he
tw
o
in
dep
e
nd
e
nt
protoc
ol
sta
ck
s.
I
n
WSN
an
d
MC
N
c
onve
rg
e
nce
netw
ork
t
he
ga
te
wa
y
play
s
an
im
portant
ro
le
f
or
t
h
e
conve
rg
e
nce
of
tw
o
netw
ork
s.
T
her
e
fore,
ga
te
way
sel
ect
ion
in
t
he
c
onve
rg
e
nce
netw
ork
is
com
ing
out
as
a
researc
h
issue
s
w
hic
h
nee
d
to
be
a
ddres
sed
[
13
]
.
T
he
opti
m
al
gate
way
sel
ect
ion
m
echan
ism
sel
ect
s
the
m
os
t
appr
opriat
e g
at
eway f
or integ
r
at
ion
of
WSN
a
nd MC
N [
14, 15].
Fi
gure
1.
Data
flo
w
in
a
WSN
-
MC
N conve
rg
e
nce
ne
twork
The
flat
arc
hitec
ture
of
c
onve
rg
e
nce
netw
ork
as
c
om
par
ed
to
hierar
c
hic
al
of
inte
gr
at
e
d
WSN
a
nd
MC
N
netw
ork
is
go
i
ng
to
be
help
fu
l
for
higher
re
sea
rch
[
6]
.
The
a
dv
a
nta
ges
of
t
he
flat
arch
it
ect
ure
in
cl
ud
es
the
sens
or
no
de
s
to
over
hea
r
the
direct
c
on
t
ro
l
sig
nalli
ng
f
ro
m
the
base
s
ta
ti
on
of
MC
N
wh
e
re
as
MC
N
can
directl
y
con
t
rol
the
eff
ic
ie
ncy
of
WSN.
For
r
eal
converge
nc
e,
the
data
cha
nn
el
s
betwee
n
tw
o
protoc
ols
s
ta
cks
need
to
be
im
plem
ented
f
or
inf
or
m
at
ion
e
xc
hange.
F
or
s
m
oo
th
ex
cha
nge
of
c
ontr
ol
s
ign
al
li
ng
s
om
e
cr
os
s
-
MAC
shou
l
d
be
desig
ned.
Co
ns
ide
rin
g
the
he
avy
traff
ic
ge
ner
at
e
d
from
WSN
as
well
a
s
MC
N
in
co
nverg
e
nce
netw
ork,
the
MAC
la
ye
r
of
co
nver
ge
nce
netw
ork
re
qu
i
r
es
new
re
sour
ce
al
locat
io
n
s
chem
e
to
ac
hieve
t
he
Qu
al
it
y
of
Service
(
QoS).
A
cell
ular
assist
ed
Q
ualit
y
of
Se
rv
ic
e
(
Q
oS)
res
ource
al
locat
io
n
al
gorithm
has
be
e
n
pro
po
se
d
t
o
re
du
ce
colli
sio
n
by
ta
king
t
he
m
ob
il
e
te
r
m
ina
l
as
gateway
w
hich
a
ct
s
a
s
se
r
vice
sc
heduler
for
high
pr
i
or
it
y ser
vice
[16
-
19]
.
WSN
ge
ne
rates
a
huge
am
ount
of
data
on
ti
m
el
y
basis
wh
ic
h
nee
d
to
be
t
ran
sm
it
te
d
by
t
he
gate
way
MS
to
th
e
base
sta
ti
on
.
But
in
the
c
onve
rg
e
nc
e
netw
ork
MC
N
com
pone
nt
is
m
ai
nly
resp
on
si
ble
f
or
rea
lt
i
m
e
data
tra
ns
m
issio
n
f
or
voic
e
ca
ll
s
an
d
vid
e
os
r
at
her
tha
n
t
he
da
ta
transm
issi
on
i
n
WSN
com
pone
nt.
So
to
ha
nd
le
the d
at
a
tra
ff
ic
in
c
onve
rg
e
nce
net
wor
k
t
her
e
is
a n
ee
d
f
or
da
ta
traf
fic
c
on
t
r
olli
ng
pr
ocess
to
m
ai
ntain
the
Q
oS
for
MC
N
as
w
el
l as
W
S
N.
In
t
h
is
pa
pe
r,
we
ha
ve
f
ocu
s
ed
on
t
he
data
flo
w
in
t
he
MAC
la
ye
r
of
W
SN
a
nd
MC
N
conve
rg
e
nc
e
netw
ork
for
Q
oS
m
anag
em
e
nt
[
20]
.
W
e
ha
ve
c
onsidere
d
pr
eem
ptive
an
d
non
preem
pt
ive
que
uing
m
od
el
by
giv
in
g
em
ph
as
is
on
the
pri
ori
ty
of
d
at
a p
ac
ke
t
to
e
nter
i
n
to
the queu
e
. W
e
ha
ve
pro
po
s
ed
a
pri
ori
ty
an
d
a
no
n
pr
i
or
it
y
base
d
qu
e
uing
m
od
el
f
or
WSN
a
nd
MC
N
c
onve
r
gen
ce
net
wor
k
w
her
e
in
ca
se
of
pr
io
rity
,
MC
N
is
ta
ken
as
highe
r
an
d
WSN
is
of
l
ow
e
r
pri
ori
ty
as
sh
own
i
n
Figure
1.
QoS
par
am
et
ers
in
MAC
la
ye
r
suc
h
as
pro
bab
il
it
y
o
f
dro
pp
i
ng
data
pack
et
s
,
pr
ob
a
bili
ty
of
data
pack
et
i
n
se
rvi
ces
of
WSN
and
MC
N
has
be
e
n
analy
zed m
at
he
m
at
ic
ally with
nu
m
erical
r
e
su
lt
s.
The
rest
of
t
he
pa
pe
r
is
or
ga
nized
as
f
ollo
ws:
Sect
i
on
2
descr
i
bes
the
pro
po
se
d
syst
em
m
od
el
for
pr
i
or
it
y
and
non
pr
i
or
it
y
ser
vices.
I
n
Sect
i
on
3,
we
stu
di
ed
the
pe
rform
ance
factors
of
MC
N
a
nd
W
S
N
conve
rg
e
nce.
F
inall
y, we
c
onc
lud
e
our
work i
n
Sect
io
n 4.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
QoS
M
anage
m
ent in W
SN
-
M
CN
Con
ver
ge
nc
e Net
work
Us
ing
Priorit
y B
ase
d
T
ra
ff
ic
... (
Anita
Swa
i
n
)
1161
2.
S
YS
TE
M MO
DEL DES
C
RI
PTION
W
e
c
onsider
a
two
pr
io
rity
syst
e
m
m
od
el
fo
r
WSN
-
MC
N
Conver
ge
nce
netw
ork,
wh
e
r
e
MC
N
data
pa
ckets
c
onsid
ered
bein
g
the
highest
pri
ori
ty
and
WSN
da
ta
pack
et
s
c
onsidere
d
to
be
the
lo
west
pri
or
it
y
for
serv
ic
e
at
m
ob
i
le
syst
em
(MS).
T
he
pac
ket
s
a
rr
ival
f
or
both
MC
N a
nd
WSN
f
ollow
a
P
oisson
proces
s
at
a
rate
λ
h
an
d
λ
l
re
sp
e
ct
ively
.
Data
pa
ckets
of
both
MC
N
an
d
WSN
get
se
rv
e
d
by
the
m
ob
il
e
sta
ti
on
of
m
ob
il
e
cel
lular
netw
ork
a
nd
t
he
se
rv
ic
e
ti
m
e
is
an
ex
pone
ntial
ly
distrib
uted
ra
ndom
var
ia
ble
with
m
ean
1/
μ
h
an
d
1/
μ
l
resp
ect
ively
.
T
he
data
traf
fic
of
both
t
he
net
works
offer
e
d
t
o
the
que
ue
if
t
he
se
rv
ic
e
poin
t
is
consi
de
red
to
be
bu
sy.
The
t
raffic
load
is
de
note
d
by
ρ
i
=
λ
i
/
μ
i
,
w
her
e
i
=h
or
l,
dep
e
nds
on
MC
N
or
WSN
data
s
erv
ic
e
resp
ect
ively
.
A
m
ob
il
e
dev
i
ce
is
as
su
m
ed
to
be
e
qu
i
pp
e
d
with
a
m
e
m
or
y
siz
e
of
N
(buffer
s
pace
hav
i
ng
N=
1,
2…..
N
)
to
ho
l
d
inc
omi
ng
data
packet
s
in
a
qu
e
ue
(
both
f
r
om
MC
N
a
nd
WSN)
ti
ll
they
get
proce
ssed.
Wh
e
n
t
he qu
e
ue
size i
s equal t
o N,
t
he
inc
omi
ng
data
pac
kets are d
rop
ped.
Figure
2. Stat
e
transiti
on d
ia
gra
m
f
or
non
-
pr
e
e
m
ptive prio
rity
syst
e
m
In
t
his
pa
pe
r
Q
oS
m
easur
em
e
nt
is
done
by
a
naly
zi
ng
t
he
ar
rival
m
echan
ism
,
serv
ic
e
m
ec
han
ism
a
nd
buff
e
rin
g
m
echan
ism
to
reduc
e
data
dro
pp
i
ng
an
d
wait
in
g
tim
e
in
the
data
qu
e
ue
[
21]
.
He
re
we
hav
e
pro
po
s
ed
a
no
n
-
pr
e
em
pt
ive
pr
i
or
it
y
a
nd
preem
ptive
pr
i
or
it
y
m
od
el
with
2
-
Pr
i
or
it
y
finite
buf
fer
qu
e
uing
syst
em
.
Th
e
resu
lt
s
of both
the m
od
el
s ar
e
com
par
ed
a
nd
discusse
d i
n n
um
erical
secti
on.
2
.
1.
Analysis
of Non
-
Preem
p
tive
Pri
orit
y Mode
l
I
n
this
m
od
el
we
assum
e
a
non
-
pr
eem
ptive
pr
i
or
it
y
syst
em
,
where
the
ongo
i
ng
se
r
vice
of
WSN
dat
a
is
al
lowe
d
t
o
c
om
plete
it
s
serv
ic
e
e
ven
if
a
high
pr
i
or
it
y
MC
N
da
ta
ar
rives
t
o
t
he
m
obil
e
dev
ic
e
f
or
s
erv
ic
e
.
The
MC
N
data
adm
it
s
into
th
e
queu
e
a
nd
w
ai
t
un
ti
l
the
ser
vice
po
i
nt
is
f
r
ee.
He
re
we
co
ns
ide
r
a
tw
o
pri
or
it
y
syst
e
m
w
it
h buff
e
r
ca
pacit
y N
=
2. Stat
e tra
nsi
ti
on
diag
ram
f
or t
he
a
bove
m
od
el
u
sin
g M
arko
v
c
hain
is
sho
w
n
in Figu
re
2.
T
he
steady st
at
e
pro
ba
bili
ti
es are d
e
rive
d usin
g balance
d
e
quat
io
ns.
Let
P
i,
j
,
k
be
the
sta
te
pr
ob
a
bili
ti
es
wh
e
re
i
(i=0
,1,2,..
.)
de
no
te
s
the
sta
te
of
lo
w
pr
io
rity
WSN
data,
j
(
j
=0
,1,2,..
.)
de
no
te
s
the
sta
t
e
of
hi
gh
pr
i
ori
ty
MC
N
data
and
k
de
note
s
the
c
urre
nt
se
r
vice
sta
tu
s
of
data
i.e
.
MC
N
or
WSN
data.
The
value
of
k
=
1,
de
note
s
WSN
data
a
nd
k=
2,
de
note
s
MC
N
data.
Th
e
al
gorithm
for
data
flo
w
in
Non
P
r
ee
m
ptive
pr
i
ori
ty
m
od
el
is
presented
i
n
al
go
rithm
1.
In
t
his
al
gorithm
we
hav
e
i
niti
al
iz
e
d
the
syst
e
m
at
init
i
al
sta
te
wh
en
tim
e
t=
0
and
t
he
m
ob
il
e
sta
t
ion
is
f
ree.
T
he
bu
ff
e
r
ha
s
e
m
pt
y
sp
ace
to
adm
i
t
requesti
ng
MC
N
or
WSN
se
r
vice.
When
a
new
re
quest
is
adm
it
te
d
to
queue
the
c
urre
nt
sta
tus
of
qu
eue
is
increase
d by one.
I
n
case t
he buffe
r
is
f
ull t
he
n
et
w
ork
r
e
quest
is drop
ped.
Using
pro
bab
il
ist
ic
arg
um
ent
we ob
ta
in
the
fol
lowi
ng b
al
a
nc
ed
e
qu
at
i
on
s:
0
0
,
1
,
2
1
,
0
,
1
h
l
h
l
P
P
P
(1)
0
,
1
,
2
0
,
2
,
2
1
,
1
,
1
0
l
h
h
h
l
h
P
P
P
P
(
2)
1
,
0
,
1
0
1
,
1
,
2
2
,
0
,
1
l
h
l
l
h
l
P
P
P
P
(
3)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1
159
–
1
170
1162
0
,
2
,
2
0
,
1
,
2
hh
PP
(
4)
1
,
1
,
2
0
,
1
,
2
hl
PP
(5)
1
,
1
,
1
1
,
0
,
1
lh
PP
(
6)
2
,
0
,
1
1
,
0
,
1
ll
PP
(
7)
So
lvi
ng for i
nd
ividu
al
sta
te
pr
ob
a
bili
ti
es, let
u
s ass
um
e
:
2
,
0
,
1
P
=
k
1
,
0
,
1
P
=
l
l
k
1
,
1
,
1
P
=
h
l
k
0
P
=
2
()
()
l
h
h
l
l
l
h
l
l
h
h
k
0
,
1
,
2
P
=
2
()
()
l
h
l
h
l
l
l
h
h
k
0
,
2
,
2
P
=
2
2
()
()
l
h
l
h
l
l
h
l
h
h
k
1
,
1
,
2
P
=
()
()
l
h
l
h
l
l
h
l
h
h
k
Using
norm
al
i
zat
ion
c
onditi
on
for
the
equati
on as
:
0
0
,
1
,
2
1
,
0
,
1
0
,
2
,
2
1
,
1
,
2
1
,
1
,
1
2
,
0
,
1
1
P
P
P
P
P
P
P
Wh
e
re
k
=
1 +
h
l
+
l
l
+
lh
lh
r
+
2
2
lh
lh
r
+
2
lh
l
r
+
2
()
()
l
h
h
l
l
l
h
l
l
h
h
And r =
l
h
l
l
h
h
Algori
th
m
1
Algorithm
f
or
Mob
il
e Stat
io
n (MS)
A
ll
ocati
on U
si
ng No
n Preem
ptive prior
it
y M
odel
MS (t)
is
the i
ni
ti
al
Stat
e o
f
th
e MS at
Tim
e t
=0
a
nd Que
ue Si
ze <N
If
MC
N Data
S
erv
ic
e
request t
hen
Allocat
e to
MS
for
Ser
vice;
MS is B
us
y;
Else
If
MS
is Bu
sy
AND
MC
N
Re
qu
e
st OR
WSN Re
qu
est
t
he
n
Qu
e
ue
t
he
Re
quest
;
N=N
+
1;
Else
Drop the
Req
ue
st;
En
d
if
En
d
if
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
QoS
M
anage
m
ent in W
SN
-
M
CN
Con
ver
ge
nc
e Net
work
Us
ing
Priorit
y B
ase
d
T
ra
ff
ic
... (
Anita
Swa
i
n
)
1163
If
WSN
Data
S
erv
ic
e
request t
hen
Allocat
e to
MS
for
Ser
vice;
MS is B
us
y;
Else
If
MS
is Bu
s
y AND
WCN
Re
qu
est
OR M
SN
Re
quest
th
en
Qu
e
ue
t
he
Re
quest
;
N=N
+
1;
Else
Drop the
Req
ue
st;
En
d
if
En
d
if
2
.
2
.
Perf
orm
ance
Ind
ic
es
(N
on
-
Preem
p
tiv
e Pri
orit
y M
odel
)
In
t
his
sect
io
n
var
i
ou
s
pe
rform
ance
m
a
tric
es
cal
culat
ed
t
o
m
ai
ntain
QoS
of
t
he
syst
em
.Th
e
prob
a
bili
ty
that
arr
i
ving
da
ta
pack
et
s
f
ro
m
MC
N
a
nd
WSN
is
dro
pp
e
d
due
to
busy
m
obil
e
sta
ti
on
an
d
buff
e
r
s
pace
i
n
qu
e
ue
is fu
ll
is
de
no
te
d
as
P
D
(
non
-
pree
m
ptive):
P
D
(
non
−
preem
ptive
)
=
P
(0,2,2)
+
P
(1,1,1)
+
P
(2,0,1)
+
P
(1,1,2)
.
The pr
obabili
ty
that m
ob
il
e stat
ion
is
bu
sy i
n
se
rv
i
ng lo
w pr
i
or
it
y
W
S
N data
p
ac
kets
w
it
ho
ut
pr
eem
ption
is:
P
ws
n
=
P
(1,
0, 1)
+
P
(
2, 0,
1)
+
P
(1,
1, 1)
.
The pr
obabili
ty
that m
ob
il
e stat
ion
is
b
us
y i
n
se
rv
i
ng at le
a
st o
ne
MC
N
d
a
ta
p
acket
is:
P
m
cn
=
P
(
0, 1, 2)
+
P
(0,
2,
2)
+
P
(
1, 1,
2)
.
Wh
e
n
t
he
m
obil
e
dev
ic
e
is
busy,
a
rr
i
ving
MC
N
an
d
WSN
data
pa
c
kets
has
t
o
wait
in
the
qu
e
ue.
The wai
ti
ng ti
m
e in queu
e
is
denoted
as
W
q
(non
-
Pr
eem
ptive)
an
d
ca
n be
expresse
d
as
W
q
(
non
−
Preem
pt
iv
e
)
=
P
(0,2,2)
+
P
(1
,1,1)
+
P
(2,0,1)
+
P
(1,1,2)
.
2
.
3
. An
aly
sis
of Preem
pt
ive
Pri
ority
Mode
l
In
WSN
-
MC
N
co
nv
e
r
gen
ce
netw
ork,
MC
N
data
pac
kets
ha
ve
it
s
pri
ori
ty
ov
e
r
WSN
data
pac
kets
because
of
it
s
r
eal
tim
e
app
li
c
at
ion
.
He
nce,
WSN
data
pa
c
kets
ca
n
be
del
ay
ed
by
passing
t
he
MC
N
dat
a
on
a
pr
i
or
it
y
basis.
Hen
ce
a
pree
m
pt
ive
pr
i
or
it
y
m
od
el
has
be
en
intr
oduc
ed
to
ha
ndle
the
pr
eem
ption
of
WSN
pack
et
s
w
hich
are
cu
rr
e
ntly
in
serv
ic
e,
upon
a
rr
ival
of
MC
N
data.
I
n
this
m
od
el
the
ser
vice
t
i
m
e
is
exp
o
nent
ia
lly
distrib
uted
a
nd
it
will
sat
isfy
m
e
m
or
y
le
ss property,
he
nce
the r
esults
will
be
sam
e
f
or
pr
ee
m
ptive
re
sum
e
an
d
pr
eem
ptive
non
res
um
e
case.
I
n
t
his
m
od
el
bo
t
h
MC
N
a
nd
WSN
data
are
adm
it
te
d.
If
M
CN
data
is
in
s
erv
ic
e
and
WSN
data
requests
for
se
rv
i
ce
the
n
it
wi
ll
wait
in
queu
e
ti
ll
m
ob
il
e
st
at
ion
gets
f
ree.
And
if
the
queue
is
fu
ll
,
the
n
t
he
re
qu
e
sti
ng
WSN
data
a
re
dro
ppe
d.
O
n
the
ot
her
ha
nd
i
f
MC
N
data
a
rr
iv
es
for
ser
vice,
it
pree
m
pts
the on
go
i
ng se
rv
ic
e
of
WSN
data w
hich
w
il
l l
os
e f
r
om
the p
r
ocess
.
Figure
3. Stat
e
transiti
on d
ia
gra
m
fo
r
preem
ptive prior
it
y sy
stem
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1
159
–
1
170
1164
We
c
on
si
der
a
two
pr
i
or
it
y
syst
e
m
with
buf
f
er
ca
pacit
y
N
is
eq
ual
to
2.
T
he
sta
te
tra
ns
it
ion
dia
gr
am
for
the
a
bove
m
od
el
us
ing
M
arko
v
chai
n
is
sh
ow
n
in
Fig
ure
3.
Indivi
du
al
sta
te
pr
ob
a
bili
ti
es
can
be
re
presente
d
as
P
i,j
w
he
re
i
(i=0,1,
2,
.
..)
de
no
te
s
the
sta
te
of
lo
w
pr
io
rity
W
S
N
data
an
d
j
(j
=
0,1,2,
...)
denotes
the
st
at
e
of
high
pri
or
it
y
MC
N
data.
Th
e
al
gorithm
for
data
fl
ow
in
P
r
ee
m
ptive
pr
io
ri
ty
m
od
el
is
pre
sented
in
al
gor
it
h
m
2.
This
al
gorithm
discuss
e
d t
he
syst
e
m
is
a
t
ini
ti
al
sta
te wh
en
tim
e t=
0 and
th
e m
ob
il
e
sta
ti
on
is
fr
ee
. T
he s
yst
e
m
consi
ders
pr
io
r
it
y
m
e
tho
d wit
h pr
eem
ption
f
or MC
N
a
nd
WSN se
rv
ic
e
r
equ
e
st.
Fo
r
WSN
no
preem
ption
is
pe
rm
i
tt
ed
but
for
MC
N
th
e
on
go
i
ng
se
rv
ic
e
o
f
WSN
is
pre
e
m
pted
from
the
syst
e
m
.
The
buff
e
r
ha
s
em
pt
y
sp
ace
to
adm
i
t
req
uesti
ng
MC
N
or
WSN
se
rv
ic
e.
W
hen
a
ne
w
re
quest
is
adm
i
tt
ed
to
th
e
qu
e
ue
t
he
cu
rr
e
nt
sta
tus
of
qu
e
ue
is
inc
re
ased
by
one.
I
n
case
the
buf
f
er
is
f
ull
the
ne
twork
request is
dro
pped
.
U
sin
g pro
bab
il
ist
ic
arg
um
ent
the stea
dy stat
e
b
al
ance
d eq
ua
ti
on
s a
re as
foll
ow
s:
0
0
,
1
1
,
0
l
h
h
l
P
P
P
(8)
0
,
1
0
,
2
0
l
h
h
h
h
P
P
P
(
9)
1
,
0
2
,
0
1
,
1
0
l
h
l
l
h
l
P
P
P
P
(10)
1
,
1
1
,
0
1
,
2
0
,
1
l
h
h
h
h
l
P
P
P
P
(11)
1
,
2
1
,
1
0
,
2
h
h
l
P
P
P
(1
2)
2
,
1
1
,
1
2
,
0
h
l
h
P
P
P
(1
3)
0
,
2
0
,
1
l
h
h
PP
(14)
2
,
0
1
,
0
2
,
1
h
l
l
h
P
P
P
(
15)
So
lvi
ng for i
nd
ivid
ual
sta
te
pr
ob
a
bili
ti
es, let
u
s ass
um
e,
0
,
2
P
=
k
0
P
=
2
2
2
2
()
l
h
l
h
h
k
0
,
1
P
=
)
(
lh
h
k
1
,
0
P
=
l
h
h
h
l
h
lh
kM
1
,
1
P
=
kL
2
,
0
P
=
2
(
(
)
(
)
L
l
l
h
h
h
l
h
l
h
lh
kM
1
,
2
P
=
(
L
)
lh
h
k
2
,
1
P
=
l
h
h
h
l
h
lh
h
l
l
h
M
k
LL
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
QoS
M
anage
m
ent in W
SN
-
M
CN
Con
ver
ge
nc
e Net
work
Us
ing
Priorit
y B
ase
d
T
ra
ff
ic
... (
Anita
Swa
i
n
)
1165
0
,
3
P
=
h
h
k
3
,
0
P
=
2
2
l
h
h
h
l
h
l
l
l
h
M
k
L
Using
norm
al
i
zat
ion
c
onditi
on
for
the
abo
ve
equati
ons as:
0
0
,
1
1
,
0
1
,
1
0
,
2
2
,
0
1
,
2
2
,
1
0
,
3
3
,
0
1
P
P
P
P
P
P
P
P
P
P
Wh
e
re,
M =
2
22
2
l
h
l
h
h
a
nd L =
l
h
h
l
l
h
h
l
h
l
l
h
l
M
K=
1
1
1
1
1
l
h
l
h
h
h
l
h
l
h
l
h
h
l
h
h
l
h
l
l
h
M
MM
Algori
th
m
2
Algorithm
f
or
Mob
il
e Stat
io
n (MS)
A
ll
ocati
on U
si
ng P
ree
m
pt
ive prio
rity
Mod
el
MS (
t) is
the i
ni
ti
al
Stat
e o
f
th
e MS at
Tim
e t
=0
a
nd Que
ue Si
ze <N
If
MC
N Data
S
erv
ic
e
request t
hen
Allocat
e to
MS
for
Ser
vice;
MS is B
us
y;
Else
If
MS
is Bu
sy i
n
Se
r
ving MC
N
Re
qu
est
A
N
D Ar
riving M
CN Req
uest
O
R
W
S
N
Re
qu
e
st t
hen
Qu
e
ue
t
he
Re
quest
;
N=N
+
1;
Else
D
r
op the
Req
ue
st;
En
d
if
En
d
if
If
WSN
Data S
erv
ic
e
request t
hen
Allocat
e to
MS
for
Ser
vice;
MS is B
us
y;
Else
If
MS
is Bu
sy i
n
Se
r
ving
WSN Re
qu
est
A
N
D Ar
riving M
CN Req
uest t
he
n
Pr
eem
pt the WSN Ser
vice a
nd
Allow M
CN
in to Ser
vice;
Else
If
MS
is Bu
sy i
n
Se
r
vi
ng
WSN Re
qu
est
A
N
D Ar
riving
W
SN
Re
quest
th
en
Qu
e
ue
t
he
Re
quest
;
N=N
+
1;
Else
Drop the
Req
ue
st;
En
d
if
En
d
if
En
d
if
2
.
4
.
Perf
orm
ance
Ind
ic
es
(
P
reem
pt
ive
Pri
orit
y Model
)
The
perf
or
m
ance
m
a
tric
es
re
qu
i
red
f
or
Q
oS
m
easur
em
en
t
has
bee
n
cal
culat
ed.
The
pro
ba
b
il
it
y
of
arr
ivi
ng
data
pa
ckets f
ro
m
MC
N
an
d WS
N
is dro
pp
e
d du
e
to busy m
ob
il
e stat
ion
and
buff
e
r
sp
ace
in q
ueu
e is
fu
ll
is
denote
d as P
D(preem
pt
ive) :
P
D
(
Pree
m
ptive
)
=
P
(1, 2)
+
P
(2, 1)
+
P
(0, 3)
+
P
(3, 0).
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1
159
–
1
170
1166
Pr
oba
bili
ty
that the m
ob
il
e stat
ion
is
bu
sy i
n serv
i
ng lo
w pr
i
or
it
y
WSN d
at
a p
ac
kets is:
P
wsn
=
P
(1, 0)
+
P
(2, 0)
+
P
(3, 0).
Pr
oba
bili
ty
that the m
ob
il
e stat
ion
is
busy i
n serv
i
ng at le
ast
one MCN
d
at
a
p
ac
ket is:
P
m
cn
=
P
(0, 1)
+
P
(1, 1)
+
P
(2, 1).
Wh
e
n
t
he
m
ob
il
e d
evice i
s
busy i
n
ser
v
in
g d
at
a p
ackets
, arr
iving M
CN a
nd
WSN d
at
a
pa
ckets h
a
s to
wait
in
the
qu
e
ue
i
f
t
he
queue
is
not
f
ull.
Th
e
wait
in
g
ti
m
e
in
que
ue
is
de
no
te
d
by
W
q
(Pree
m
ptive)
and
ca
n
be
cal
culat
ed
as:
W
q
(
Preemp
tive
)
=
P
(1,2)
+
P
(2,1)
+
P
(1,1)
+
P
(0,2)
+
P
(2
,0)
+
P
(0,
3)
+
P
(3,0).
3.
NU
MER
IC
A
L
RES
ULTS
In
this secti
on
nu
m
erical
an
al
ysi
s o
f
both t
he
m
od
el
s ar
e presente
d.
T
he QoS o
f
the pr
opose
d
m
od
el
s
hav
e
bee
n
ach
ie
ved
by
e
valu
at
ing
th
e
perform
ance
par
am
et
ers.
T
he
pe
rfor
m
ance
is
e
va
luate
d
i
n
te
r
m
s
of
dro
pp
i
ng
probabil
it
y
(
P
D
),
pr
ob
a
bili
ty
of
hi
gh
(MCN)
pr
i
ori
ty
an
d
l
ow
(
WSN)
pr
io
rity
data
i
n
t
he
syst
e
m
an
d
wait
ing
ti
m
e
(
Wq)
of
the
dat
a
pack
et
s
in
t
he
qu
e
ue.
T
he
para
m
et
ers
ta
ken
f
or
Fi
gure
4
to
F
igure
11
a
re
μ
l
=
1.
0;
μ
h
= 1.0;
λ
l
=
0.5; λ
h
=
0.1 t
o 1
.0
.
Figure
4
an
d
F
igure
5
represe
nts
the
beh
a
vi
or
of
dro
ppin
g
pro
ba
bili
ty
P
D
with
res
pect
t
o
i
ncr
easi
ng
MC
N
data
tra
ffi
c
load
f
or
bo
t
h
no
n
-
pr
eem
ptive
an
d
preem
pti
ve
case
s
res
pec
ti
vely
.
It
ca
n
be
obse
r
ve
d
that
with
the
inc
reasin
g
traff
ic
l
oad
of
MC
N
data,
droppin
g
pro
ba
bi
li
t
y
increases
for
bo
t
h
t
he
m
od
el
s
.
But
by
var
yi
ng
the tra
ff
ic
loa
d
of
WSN t
he
dr
opping
prob
a
bi
li
ty
d
ecreases
with
dec
rease
of
WSN
data
tr
aff
ic
.
Fi
gure
4
sh
ows
a
lower
value
of
probabil
it
y
of
dro
ppin
g
for
non
-
pr
ee
m
pt
ive
m
od
el
com
par
ed
to
pr
eem
ptive
m
od
el
in
Figure
5.
This
is
due
to
the
fa
ct
that
data
pa
ckets
a
re
ser
ve
d
with
out
pr
ee
m
pt
ion
.
Hen
ce
,
by
c
hoos
i
ng
pro
pe
r
data traf
fic loa
d o
f
WSN
,
dro
pp
i
ng pr
ob
a
bili
ty
can
be
m
ini
m
iz
ed
to m
ai
nt
ai
n
Q
oS.
Figure
4. Tra
ffi
c intensit
y o
f M
CN V
s
pr
obabili
ty
o
f dro
ppin
g for
non p
r
ee
m
ptive prio
rity
m
od
el
Figure
6
an
d
F
igure
7
de
picts
the
pro
ba
bili
t
y
of
ser
vi
ng
WSN
data
in
t
he
syst
em
fo
r
var
i
ou
s
WSN
traff
ic
l
oad
with
re
sp
ect
t
o
M
CN
traf
fic
inte
ns
it
y
in
case
of
no
n
pr
eem
ptive
a
nd
preem
ptive
pr
i
or
it
y
m
od
el
resp
ect
ively
.
It
can
be
see
n
th
at
with
inc
reas
e
of
MC
N
tra
ffi
c
load
probabi
li
ty
of
se
rv
i
ng
WSN
data
decre
ases
for
both
the
m
od
el
s
.
T
his
is d
ue
t
o
co
ntin
uous
ser
vice of MC
N
data
i
n
non
-
pr
eem
ptive p
ri
or
it
y
m
od
el
w
it
hout
pr
eem
ption
of
WSN
data
w
he
re
as
WSN
da
ta
preem
pted
in
pr
eem
ptive
pr
i
o
rity
m
od
el
due
to
hi
gh
pri
or
it
y
MC
N
data.
I
n
Figure
7,
pr
ee
m
pt
ive
pri
ori
ty
m
od
el
s
hows
higher
nu
m
ber
of
WSN
data
serv
ic
e
in
com
par
is
on
with
non
pree
m
pt
ive
pri
ori
ty
m
od
el
in
Fig
ur
e
6
wh
ic
h
in
creases
t
he
W
SN
data
ser
vice.
Hen
ce
,
pr
ee
m
pt
ive
pr
i
or
it
y m
od
el
can
ac
hieve
th
e b
et
te
r Q
oS
for
MC
N
a
nd
W
SN
by m
ai
ntain
in
g
a
balan
ce
d WS
N
tra
ff
ic
.
The
pro
ba
bili
ty
of
MC
N
data
is
in
se
rv
ic
e
in
crease
with
inc
rease
i
n
MC
N
t
raffic
inte
ns
it
y
is
sho
wn
i
n
Figure
8
a
nd
Figure
9
f
or
non
-
pr
eem
ptive
pr
i
or
it
y
m
od
el
and
preem
pti
ve
pri
o
rity
m
od
el
res
pecti
vel
y.
It
is
ob
s
er
ved
t
hat
the
pr
ob
a
bili
ty
of
se
r
ving
MC
N
data
i
ncr
eas
es
with
inc
reas
e
in
MC
N
tra
f
fic
load
f
or
bo
th
the
m
od
el
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
QoS
M
anage
m
ent in W
SN
-
M
CN
Con
ver
ge
nc
e Net
work
Us
ing
Priorit
y B
ase
d
T
ra
ff
ic
... (
Anita
Swa
i
n
)
1167
Figure
5. Tra
ffi
c intensit
y o
f M
CN V
s
pr
obabili
ty
o
f dro
ppin
g fo
r pr
eem
ptive
pr
i
or
it
y
m
od
el
Figure
6. Tra
ffi
c intensit
y o
f M
CN V
s
pr
obabili
ty
o
f WS
N
d
at
a in
servic
e
for
non p
reem
ptive
pr
i
or
it
y
m
od
el
Figure
7.
Tra
ffi
c intensit
y o
f M
CN V
s
pr
obabili
ty
o
f WS
N
d
at
a in
servic
e
for
preem
ptive
pr
i
or
it
y m
od
el
As
MC
N
data
hav
e
hi
gh
e
r
pr
i
or
it
y
in
preem
ptive
pr
i
or
it
y
m
od
el
he
nce
wit
h
adm
issi
on
of
m
or
e
nu
m
ber
of
MC
N
traf
f
ic
,
W
S
N
data
is
pr
eem
pted
w
hi
ch
inc
reases
th
e
pro
bab
il
it
y
of
ser
ving
MC
N
data
as
pr
ese
nt
ed
in
Figure
8.
I
n
no
n
-
pr
eem
ptive
pri
or
it
y
m
od
el
pro
ba
bili
ty
of
se
rv
i
ng
MC
N
data
inc
reases
if
MC
N
data
a
dmi
ssion
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1
159
–
1
170
1168
increases
in
num
ber
s
as
i
n
Figure
9.
B
ut
with
i
ncr
eas
e
of
WSN
data
inte
ns
it
y,
se
rv
i
ng
MC
N
data
dec
reases
i
n
bo
t
h
m
od
el
s. No
n pr
eem
ptive p
ri
or
it
y
m
odel
in
Figure
8
s
hows
li
tt
le
d
ecrease in se
rv
i
ng MC
N data
co
m
par
ed
to
ser
ving
MC
N
tra
ff
ic
in
Fi
gure
9
as
WSN
da
ta
are
pr
e
e
m
pted
f
ro
m
serv
ic
e
i
n
pree
m
pt
ive
pr
i
or
it
y
m
od
el
wh
ic
h
s
hows
a
bette
r
re
su
lt
.
Figure
8. Tra
ffi
c intensit
y o
f M
CN V
s
P
rob
abili
ty
o
f
MC
N data
in se
r
vice f
or
non p
ree
m
p
t
ive prio
rity
m
od
el
Fig
ur
e
9. T
raffic
intensit
y
of
MC
N
Vs
pr
obabili
ty
o
f
MC
N data
in se
r
vice f
or
preem
pti
ve
pr
i
or
it
y m
od
el
Figure
10. T
raffic
intensit
y
of
MC
N,
WSN
V
s
wait
ing
ti
m
e in queu
e
for n
on
pr
eem
ptive pri
or
it
y
m
od
el
Figure
11.
T
raffic
intensit
y
of
MC
N,
WSN
V
s
wait
ing
ti
m
e in queu
e
for p
ree
m
pt
ive prio
rity
m
od
el
Evaluation Warning : The document was created with Spire.PDF for Python.