TELKOM
NIKA
, Vol.12, No
.2, June 20
14
, pp. 455~4
6
4
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v12i2.1933
455
Re
cei
v
ed
De
cem
ber 6, 20
13; Re
vised
Ma
rch 20, 20
14; Accepted
April 5, 2014
Additional Resource Allocation for improving Fairness
in WiMAX
Khalid Mah
m
ood A
w
a
n
,
Abdul Han
a
n Abdullah*,
Khalid Hu
ss
ain
F
a
cult
y
of Com
putin
g, Univ
ersiti T
e
knologi M
a
la
ysi
a
, Mala
ysia
T
e
l: +
60-19-7
0
406
23,
+
60-1
3
-
777
09
95, +
60-
13-6
032
97
8
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: makhal
id2
@
li
ve.utm.m
y
,
ha
nan
@utm.m
y
,
hkha
lid
2@l
i
ve.
u
tm.m
y
A
b
st
r
a
ct
IEEE 802.1
6
s
t
andar
d pr
ovid
es reso
urces t
o
both
servic
e
classes
Const
ant Bit R
a
te (
C
BR) a
n
d
Varia
b
le Bit R
a
te (VBR). T
hese services r
equ
ire en
ou
gh
resources to transmit data efficiently. In this
pap
er, a S
e
rvi
c
e Bas
ed F
a
ir Reso
urce A
l
l
o
catio
n
(SBF
R
A
) Mecha
n
is
m is pr
opos
ed,
for assessi
ng
th
e
requ
ired serv
ic
e from Subscr
i
ber (SS) alon
g w
i
th t
he chann
el con
d
itio
n. T
he prop
osed
mo
del ev
alu
a
tes the
requ
ired
add
iti
ona
l reso
urces
in ord
e
r to pr
ovid
e t
he re
qu
ired serv
ice to
SS. In this propos
ed
mo
del,
a
Priority Queu
e
Schedu
li
ng
Method
olo
g
y i
s
presente
d
in
order to prov
ide a
d
d
i
tion
al
resourc
e
s as
per
chan
nel c
o
n
d
iti
on. A co
mp
ari
s
on is
ma
de w
i
th an
d w
i
thout
the pro
pose
d
mo
de
l
on
both
traffic, CBR and
VBR. T
he
exp
e
ri
ments
show
that th
e
prop
osed
d
e
si
g
n
ma
na
ges
the
user r
equ
est
by pr
ovid
in
g t
h
e
m
add
ition
a
l res
o
urces as
per r
equ
ire
m
e
n
t. T
he re
quests
are ha
ndl
ed
in
Priority Que
u
e
Sched
uli
ng
b
a
sed
mec
h
a
n
is
m. Results show
that an im
prov
e
m
e
n
t is achi
ev
ed by prov
idi
n
g the add
itio
n
a
l reso
urce on
fair
sched
uli
ng b
a
s
i
s. T
o
achieve t
he req
u
ire
d
per
forma
n
ce for fa
irness ther
e is a compro
mise
on de
lay.
Ke
y
w
ords
: Resourc
e
Alloc
a
tion, SNR, CB
R, VBR, Priority Queue
.
1. Introduc
tion
WiMAX is a
4G wirel
e
ss network tech
nolo
g
y. This sy
stem
is desi
gne
d for the
metrop
olitan
area
network
(MA
N
).
Wirel
e
ss n
e
twork a
r
e
gene
rally le
ss
efficient
and
unpredi
ctable
as
comp
are
d
to a
wire
d
network
du
e
to
limite
d
resou
r
ces. M
anag
ement of
available re
source
s
is a critic
al issu
e
in WiMAX [1
]. This re
se
a
r
ch fo
cu
se
s
on providing
th
e
quality of service (Q
oS) in
a wirel
e
ss en
vironm
e
n
t. To handle thi
s
challen
ge WiM
AX provides
stron
g
co
nten
der for b
r
oa
d
band
wirel
e
ss technol
ogy.
In WiMAX when a ne
w subscri
b
e
r
req
uest
s
for the
service a call admissio
n
control
pro
c
e
ss i
s
initiated. Acco
rding
to the proce
dure, Base Station
(BS) monitors existing re
so
urces
for the provisi
on of servi
c
e to new users.
This
en
sures that the system has e
nou
gh re
sou
r
ces
to
accomm
odat
e the se
rvice
s
req
u
ired to newly re
que
sting use
r
. sch
edulin
g algo
ri
thms should
be
c
a
pable of providing minimum level for QoS [2],[3].
Sc
heduling is
a
c
r
itic
al process
for allo
c
a
ting shared res
o
urc
e
s [4],[5]. The proc
ess
involves allo
cating
band
width am
ong
the co
nten
ded
u
s
e
r
s. Sched
uling algorith
m
s
f
o
r a
particula
r sy
stem is sele
cted ba
sed
on
the type of
service
req
u
ire
d
by use
r
s in
the network a
n
d
its QoS re
qui
reme
nts. In real-time a
ppli
c
ation
su
ch a
s
video
confe
r
en
cing, voi
c
e and
strea
m
ing
delay o
r
jitter
plays
a vital
role in
QoS
re
quire
ment. A
task of the
scheduli
ng
algo
rithm in
a
mul
t
i-
cla
s
s traffic i
s
to catego
rize
the u
s
e
r
s in
one
of
the p
r
e-defin
ed
cla
s
ses. Ea
ch
u
s
er is a
ssi
gn
ed a
prio
rity of class, a
s
well a
s
ensu
r
ing that
fairness bet
wee
n
the use
r
s i
s
maintain
ed [6].
Sched
uling
a
l
gorithm i
s
i
m
pleme
n
ted
for both
sid
e
s
which a
r
e
uplin
k an
d d
o
wnli
nk i
n
WiMAX network. Th
e focus of the scheduli
ng focuse
s on allo
cation of ba
ndwi
d
th to the
sub
s
crib
er. A
su
bscri
b
e
r
may get
sep
a
rate
re
so
urces fo
r e
a
ch
appli
c
ation if
the
sched
uli
ng
algorith
m
i
s
n
o
t implem
ent
ed at th
e SS
side. If g
r
a
n
t
per Sub
s
crib
er Station
(S
S) is ap
plied
then
scheduling al
gorithm at SS need
s to decide the allocation
for each connection [7],[8].
To co
ntrol th
e sched
uling
mech
ani
sm
in WiMAX ne
twork it is ne
ce
ssary to m
a
intain
efficient me
chani
sm on th
e BS. Numbe
r
of issue
s
arise
s
at the time of uplin
k. To mana
ge
the
resou
r
ce opti
m
ally and effi
ciently the
scheduli
ng al
g
o
r
ithm at BS coordi
nate
s
wi
th the sche
du
ler
on the Subscribe
r
sid
e
[9].
In different
sched
uling
mech
ani
sm
s, variou
s ap
proa
ch
es
are use
d
for
resou
r
ce
manag
eme
n
t. Most com
m
o
n
sched
ulers
are FIFO
, WFQ, EDF and
EDF whi
c
h a
r
e u
s
ed at SSs
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 2, June 20
14: 455 – 46
4
456
for rtPS cl
ass. To han
dle t
he sch
eduli
n
g of no
n-real time
Polling Service Wei
g
hted
Fai
r
Q
u
eue
is a
dopte
d
.
Whe
r
ea
s for
providin
g the
resou
r
ce to
Best Effort
se
rvices FIFO
m
e
thod i
s
ado
p
t
ed
[10]. Taki
ng i
n
to a
c
count t
he Q
o
S fram
ewo
r
k,
a l
o
w cla
s
s p
r
io
rity se
rvice p
r
ov
ides ad
dition
al
requi
re
d re
so
urces in p
r
io
ri
ty sched
uling
para
d
igm.
WiMAX is no
t only ba
sed
on a
stati
c
m
odulatio
n sch
e
me, vari
ou
s ada
ptive mo
dulation
and
co
dding
method
olog
ies
are
u
s
e
d
. In a
wireless
enviro
n
ment m
odu
lation, sch
e
m
e
s
synchro
n
ization play
s an
i
m
porta
nt rol
e
for ad
di
tional
re
sou
r
ces wi
th re
spe
c
t to
the se
rvice a
nd
modulatio
n schem
e [11]. J. Lin et al p
r
opo
se
d
the
uplin
k fair
scheduli
ng st
ru
cture i
n
Wi
M
AX
netwo
rk [12].
This stru
ctu
r
e descr
i
b
e
s
the factor of throu
ghp
ut
an
d the delay handlin
g in multi
cla
ss’
s traffic.
This stru
ctu
r
e pre
s
ente
d
t
he Modified
Weig
hted Ro
und Ro
bin m
e
ch
ani
sm in BS.
The m
e
thod
i
s
to
man
age
the
sched
ulin
g p
r
oble
m
at
the sub
s
cribe
r
side,
and
all
o
catio
n
i
s
ma
de
based
on th
e
num
ber of
u
s
ers. SS
han
dles UGS a
n
d
real-time polling
service with MWFQ.
To
handl
e non re
al time Polling Service M
W
FQ i
s
used
and for BE traffic FIFO is
use
d
.
To allo
cate
b
and
width o
n
prio
rity basi
s
to
all the
su
bscrib
ers in t
he sy
stem a
hybrid
sched
uling
mech
ani
sm i
s
p
r
e
s
ente
d
by M. Settembe
r
et
al
[13]. WRR ba
sed
sch
e
duling
techni
que is
use
d
for ban
dwidth allo
cat
i
on to SSs
, fo
r rtPS and nrt
PS classe
s on priority ba
si
s.
Re
st of
re
sou
r
ce
s allo
cate
d to BE
cla
sses
by
u
s
ing
round
robin
(RR) m
e
chani
sm.
Dra
w
b
a
ck o
f
this meth
od i
s
if a lo
w p
r
i
o
rity cla
s
s re
quire
s a
ddi
tio
nal re
so
urce
s then it will h
a
ve to wait fo
r a
longe
r pe
riod
of time [14].
In [15] WF
Q
approa
ch i
s
u
s
ed
for SS
of bot
h n
r
tPS a
nd BE
classe
s. The
sta
nda
rd d
o
e
s
not req
u
ire
a
n
y spe
c
ific Q
o
S requi
rem
ent for the B
e
st Effort traffic [16]. To apply a com
p
l
e
x
architectu
re f
o
r the man
a
gement of re
sou
r
ces fo
r the BE traffic is not feasi
b
le for practi
cal
impleme
n
tation du
e to resource
hung
ry
archite
c
tu
re.
The ove
r
h
e
a
d
of the
syst
em will
be m
o
re
than the actu
al efficien
cy of the system
. The
main compa
r
ison in
this pape
r is made ba
sed
on
the
g
r
ant pe
r sub
s
crib
er
an
d
g
r
ant pe
r conne
cti
on. It i
s
a
nalysed
a
nd
re
comm
e
nded
that
GPCC
is better than
GPC in term
s of
delay in the netwo
rk traffic.
It is observed
that there i
s
a defici
e
n
c
y of an algo
rith
m whi
c
h
can
handl
e wi
rele
ss t
r
affic
based on th
e
fairne
ss fa
ct
or an
d allo
ca
tes the re
sou
r
ce
s b
a
sed o
n
use
r
requi
rements.
WiM
AX
hav
e mult
ipl
e
cla
s
se
s,
w
i
t
h
re
spe
c
t
t
o
t
hos
e cl
asse
s multiple
sched
uling
mech
ani
sm
are
requi
re
d. To handl
e this si
tuation an efficient algo
rith
m is requi
red
to compute and allo
cate the
existing re
so
urces
ba
sed
on re
quireme
nts. The
alg
o
r
ithm sh
ould
be able to ha
ndle the mo
b
ile
as
well as
the us
er in hard loc
a
tion.
In WiMAX m
i
nimum
re
so
urce all
o
catio
n
can b
e
m
a
de by
slot p
e
rmutatio
n al
locatio
n
methodol
ogy
[18]. Four Types of
perm
u
tati
on
are
used in
WiMAX, p
a
rtial u
s
a
g
e
sub
cha
nneli
z
atio
n (P
USC), full usa
ge
su
b
chann
eliz
atio
n
(F
USC), ada
ptive modul
ation an
d
codi
n
g
AMC an
d tile
usa
ge
sub
chann
elizatio
n
TUSC. In [1
5] a ne
w al
g
o
rithm i
s
p
r
o
posed
whi
c
h
is
based o
n
the
ch
annel
con
d
ition an
d th
e buffer capa
city. This
cro
s
s layer ap
proach in
WiM
AX
sho
w
s that
wirel
e
ss li
nk effects the
perfo
rm
an
ce
of the
syst
em. To im
pl
ement
sched
uling
mech
ani
sm a
t
the BS can
prod
uce bett
e
r results. M
a
in focus
of this
work i
s
to get maxim
u
m
throug
hput a
nd
fair sched
uling amon
g the
servi
c
e fl
ows, but
doe
s n
o
t con
s
id
er th
e
cha
n
n
e
l
status fo
r sch
edulin
g.
2. Proposed
Fairness M
o
del
Figure 1
sho
w
s the
pro
p
o
s
ed
Servi
c
e
Based
Fai
r
ne
ss
R
e
s
our
ce
A
llocat
i
o
n
(
S
B
F
RA
)
Mech
ani
sm; an ada
ptive resou
r
ce allo
cation mo
del
is develope
d by eval
uating the se
rvice.
Whe
n
use
r
submits a re
q
uest for a particular
servi
c
e, the propo
sed SBF
RA
Mech
ani
sm ch
ec
ks
the requi
red
se
rvice al
o
ng
with the
cha
nnel
con
d
ition. In thi
s
m
odel, t
w
o sub
mod
e
l
s
a
r
e
pre
s
ente
d
; Service Evalu
a
t
or and Reso
urce Man
age
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Additional Re
sou
r
ce Alloca
tion for Im
proving Fai
r
n
e
ss in WiMAX (Khalid Mahm
o
od Awan
)
457
Figure 1. Fairne
ss b
a
sed reso
urce allo
cation
In ou
r p
r
op
osed mo
del
we
con
s
id
er
both
se
rv
ice
s
Co
n
s
tant Bit
Rate
(CBR) a
nd v
a
riabl
e
Bit Rate (VBR).
2.1 Cons
tan
t
Bit Ra
te Ser
v
ice
CBR i
s
traffic which keep
s the bit rate same
th
rou
g
hout the process, in real ti
me data
strea
m
con
s
i
s
ting of fixed
size data
pa
ckets.
Th
e Q
o
S that re
qui
res the u
s
e
r
to determi
ne
a
fixed band
wi
dth requi
rem
ent at the time the co
nn
ection i
s
set
up. The data
can be
sent
in a
steady
stre
a
m
. CBR se
rvi
c
e i
s
often
u
s
ed
when
tra
n
s
mitting fixed
rate
un
comp
resse
d
vide
o. A
cla
ss of traffic that need
s acce
ss
to time slots at reg
u
lar an
d pre
c
ise interval. QoS para
m
et
er
inclu
d
e
s
cell
delay vari
ation and
cell tra
n
sfer del
ay. We n
eed
to tran
sfer
un
co
mpre
ssed voi
c
e
and vid
eo u
s
ing
CB
R. CBR is de
sig
n
ed for ATM
vi
rtual ci
rcuit
wh
ere
stati
s
tically con
s
ta
nt
amount of b
and
width is required
for t
he duration
of active con
nectio
n
. CBR service cla
s
s is
desi
gne
d for real time appli
c
ation [20].
2.2 Variable Bit Ra
te Ser
v
ice
The te
rm VB
R i
s
u
s
ed
in
telecom
m
uni
cation
and
computing
tha
t
relate
s to t
he bitrate
use
d
in the sound a
nd vid
eo en
codi
ng. VBR files
vary from the amount of out
put data per t
i
me
segm
ent. VBR allo
ws
a hi
gher
bitrate
and the
r
efo
r
e
requi
re
s mo
re sto
r
a
ge to
be allo
cated
to
compl
e
x seg
m
ent of medi
a files. The disa
dvantag
e
o
f
V
B
R
i
s
t
h
a
t
i
t
m
a
y
t
a
k
e
m
o
r
e
t
i
m
e
t
o
e
n
c
o
d
e
as
the
pr
oc
es
s is
mo
r
e
co
mp
lex. VBR
ma
y
a
l
s
o
p
o
s
e
a pr
o
b
l
em du
r
i
ng
s
t
r
e
a
m
in
g
w
h
en
the instant
an
eou
s bit rate
excee
d
s the
data rate
of the communi
cation path. Bi
ts available
can
use mo
re flex
ibility to enco
de so
und o
r
video data [20]
.
2.3
Serv
ice Ev
aluator
WiMAX Serv
ice
s
a
r
e
usu
a
lly cate
gori
z
ing in
five m
a
in
cla
s
ses.
In pro
p
o
s
ed
model
Service Eval
uator i
s
resp
onsi
b
le for re
ceiving
user
reque
st for
an
y of the
five required
se
rvice
s
.
Service Eval
uator
con
s
ist
s
of two mai
n
feature
s
, which a
r
e
Req
u
ired Se
rvice
R
s
and Cha
nnel
Condition
C
c
. Wh
en
a
use
r
requ
est
s
fo
r a
se
rvice
R
s
identify the
required S
e
rvi
c
e
wh
erea
s
C
c
will asse
ss t
he ch
annel
condition from
where the
service i
s
req
u
ired. After a
s
sessin
g the
s
e
para
m
eters final value for
α
is obtaine
d.
2.4 Reso
urc
e
Manag
e
r
Re
sou
r
ce Manag
er re
cei
v
es the value of
α
and
estimate
s the additional
resou
r
ce
requi
re
d for
providin
g the
req
u
ire
d
se
rv
ice to the
u
s
er. After fin
a
lizin
g the
R
s
the Resource
Manag
er eval
uates ho
w
m
u
ch re
se
rve
reso
urce
it
h
a
s
an
d h
o
w m
u
ch it
ca
n all
o
cate th
e u
s
e
r
for
the
R
s
. T
hen
Re
sou
r
ce M
a
nage
r fo
rwards th
e
re
sult t
o
Qu
eue
Sch
edule
r
Q
s
pla
c
e
s
the
reque
st
in the three a
v
ailable que
u
e
s, First Priority
F
p
,
Sec
o
nd Priority
S
p
,
and Thi
r
d Pri
o
rity
T
p
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 2, June 20
14: 455 – 46
4
458
Figure 2. Service Evaluation
Let
C
c
an
d
R
s
be the rand
o
m
variable
s
repre
s
e
n
ting
t
he input a
nd
output of the
cha
nnel,
r
e
spec
tively. Let
PR
|C
R
|C
be the condition
al distribution fu
n
c
tion of Rs given Cc, which
is
an inhe
rent fixed prop
erty
of the communication
s cha
nnel. The
n
the choi
ce
of the marg
inal
distrib
u
tion
P C
C
com
p
letely
determi
ne
s the joint di
st
ribution
P C
,R
C
,R
due to the
identity.
P C
,R
C
,R
P
R
|C
R
|C
PR
R
(1)
Whi
c
h, in turn, induce
s
mutual informatio
n
I R
;
C
. The overall
chann
el con
d
ition along
with SNR i
s
d
e
s
c
rib
ed a
s
O
I
R
; C
(2)
As per Sha
n
non theo
rem
for every req
u
ired
se
rv
ice
there is a
cha
nnel capa
city. That is
maximum re
quire
ment fo
r providing required
se
rvi
c
e. On the
bases of thi
s
theo
rem o
u
r
prop
osed mo
del cal
c
ul
ates the chan
nel
cap
a
city.
O
I
R
;
C
(3)
Secon
d
ly every ch
ann
el has
som
e
no
ise ratio whi
c
h i
s
call
ed
Signal to Noi
s
e
Rati
o
(SNR). In this equatio
n we
denote it a
s
with
α
, and
α
>
0
a
nd
R
c
< C
c
for the
req
u
ired
se
rvice
R
s
,
as
per the e
x
isting
code
for Rs
sh
oul
d be
>
Rs a
s
p
e
r
co
ding
algo
rithm a
nd for the
R
s
the
maximal pro
b
ability of the
SNR shoul
d be <
α
.
2.5 Queue S
c
heduler
In our p
r
op
osed mod
e
l we
also int
r
odu
ced Prio
rity Queue S
c
he
du
ling metho
dol
ogy for
providin
g re
source a
s
per
the r
equi
rem
ent of the user se
rvice.
A
c
cordi
ng to the cha
nnel
status
Queu
e Sch
e
duler
pla
c
e
s
the de
sire
d reque
st in the
first, se
con
d
and third qu
eue a
c
cordin
g to
the requi
re
d service a
nd ch
annel
con
d
ition.
First P
r
iority
(F
p
): Q
ueu
e
Sched
ule
r
ev
aluate
s
an
d
places all th
ose
re
que
sts whi
c
h
requi
re
d 10%
additional
re
sou
r
ces fo
r fulfil the requi
red
requ
est.
All those req
uest pla
c
e i
n
the
F
p
.
Sec
o
nd Priority (S
p
): Our propo
se
d model ba
sed
queu
e sched
uler pla
c
e al
l those
requ
est in Sp
which are
re
quire
d 20% a
dditional
reso
urces fo
r fulfilment to requi
red servi
c
e.
Third P
r
iority
(T
p
): In the l
a
st qu
eue
all
those
re
que
st ha
s be
en
placed
whi
c
h
req
u
ire
d
more tha
n
30
% additional reso
urce
s.
Propo
se
d qu
eue sch
edul
er is b
a
sed
on prio
rity queue
sched
u
ling, so after certai
n
perio
d of time every requi
re
d use
r
got pla
c
e in the next
S
p
and T
p
.
3. Simulation and Re
sults
To eval
uate t
he p
r
o
p
o
s
ed
model
a five
node
ba
se
d
WiMAX
scen
ario
is devel
oped
in
NCT
Un
s version 6.0 sim
u
l
a
tor. The si
m
u
lation is exe
c
uted fo
r mul
t
iple times wit
h
re
spe
c
t to the
servi
c
e
s
. Foll
owin
g param
eters a
r
e u
s
e
d
for both services.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Additional Re
sou
r
ce Alloca
tion for Im
proving Fai
r
n
e
ss in WiMAX (Khalid Mahm
o
od Awan
)
459
Table 1. Simulation Para
meters
Figure 3. Five SS based scenari
o
of Wi
MAX
The Fi
gu
re
3
sh
ows th
e
real
simulatio
n
scr
ee
n
sho
t
of five no
de
s b
a
sed SS
scena
rio
with one BS and one
Ho
st. In the first a
s
sumption, e
v
ery node de
mand
s the CBR se
rvice from
BS and BS
whi
c
h i
s
e
qui
pped
with S
B
FRA evalu
a
t
es the
dem
a
nded
service
and it
s
cha
n
nel.
Then th
e p
r
o
posed SBF
RA initiates
an
d re
ceive
s
th
e re
que
st fro
m
the u
s
e
r
a
nd p
r
ovide
s
t
h
e
requi
re
d se
rvice a
c
cordi
ngl
y
.
After exe
c
uting
simul
a
tion
for
100
su
sin
g
p
r
op
osed
model. Ba
se
Station
re
ce
ived th
e
real time re
q
uest from u
s
er form different
servi
c
e
s
. Figure 4 sh
ows the use
r
demandi
ng the
CBR servi
c
e from the ba
se
station.
Figure 4. SS
comm
uni
cati
on with Base Station for Re
quire
d Service
Descrip
tion
Range
/ Mo
del /
Qua
n
ti
t
y
Mobile Station
20 to 30
Base Station
01
Host
01
Number
Channel
10 to 20
Band
w
i
dth bet
w
e
en Host and BS
10 MB
Base Station Antenna Height
30 m
Mobile Station Height
10 m
Protocol 802.16e
Fading Variance
10.0
Path loss Expon
ent
2.0
Ricean Factor
10.0 K(db
)
Empirical Channel
Cost_231_Hat
a
Path loss Model
Ereg path loss
Freque
nc
y 2300
MHz
Transmission Pow
e
r
43 dbm
Receive Sensitivit
y
-99 dbm
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 2, June 20
14: 455 – 46
4
460
3.1 Simulation Scenario
1.
In simul
a
tion
scena
rio, the
prop
osed m
o
del i
s
ap
plied
on o
n
ly five
node
s b
a
sed
WiMAX
netwo
rk.
In
whi
c
h th
e a
s
sumptio
n
wa
s that
every
use
r
dema
n
d
s
CBR servi
c
e
with
varia
b
le
cha
nnel
con
d
ition. It is obse
r
ved that
how p
r
opo
sed model pl
a
c
e
s
the user reque
st
s in the
spe
c
ified q
u
e
ue and h
o
w it
manage
s the
requi
red resource.
In this simul
a
tion, every SS uses the
802.16e pro
t
ocol alon
g with som
e
additional
pred
efined p
a
ram
e
ters. In the first run has the capa
bility to submit require
d for all of traffic i
.
e.
UGS, rtPS nrtPS and BE. In this stu
d
y, we eval
uat
ed
the re
sou
r
ce
allocation to
use
r
s
with a
n
d
without SBFRA and monito
red the pe
rformance of SBFRA Me
chani
sm.
Figure 5.Ch
a
nnel utilization without fairness
Figure 5
sh
o
w
s the
real
simulatio
n
of
the 10
SS along
with th
eir
cha
nnel
condition,
belo
w
the
re
d line
explai
ns
on lin
e u
s
er
req
u
ire
d
e
x
tra re
so
urce
to fu
lfil
the required se
rvice
whe
r
ea
s a
b
o
v
e the red li
ne explai
ns
on line
user
has
req
u
ired
the minimu
m re
sou
r
ces.
By
evaluating th
e req
u
ire
d
reso
urce the
prop
os
ed mo
del SBFRA
Mech
ani
sm
places the
u
s
er
requ
est
s
accordin
g to the pre
s
crib
ed qu
eue sch
edul
e
r
Figure 6.Ch
a
nnel utilization with fairne
ss
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
ubs
c
r
i
b
er
C
h
ann
el
Ut
i
l
i
z
at
i
o
n
C
h
a
n
n
e
l
U
t
iliz
a
t
io
n
w
i
t
h
o
u
t
Fa
ir
n
e
s
s
T
h
r
e
s
h
ol
d
Lev
el
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
u
b
scr
i
b
e
r
C
han
nel
U
t
i
l
i
z
a
t
i
on
C
h
anne
l
U
t
iliz
a
t
io
n
w
i
t
h
Fa
ir
n
e
s
s
T
h
r
e
s
h
ol
d L
e
v
e
l
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Additional Re
sou
r
ce Alloca
tion for Im
proving Fai
r
n
e
ss in WiMAX (Khalid Mahm
o
od Awan
)
461
Figure 6
presents th
e p
r
op
ose
d
mo
del f
a
irne
ss
re
sou
r
ce
meth
odol
ogy an
d a
c
co
rding
to
the requi
rem
ent SBFRA
Mech
ani
sm
p
r
ovide
s
th
e
required
re
so
urce
and
all
use
r
s receive
d
the
servi
c
e
with a
dditional
re
so
urce b
u
t with
some
del
ay. In this
study, t
he p
r
ime fo
cu
s i
s
to p
r
ovid
e
the use
r
re
qui
red servi
c
e
s
but with a co
mpromi
se o
n
delay.
Figure 7. Co
mpari
s
o
n
of chann
el
utilizat
ion with an
d without SBFRAM
Figure 7 p
r
e
s
ents the
overall perfo
rma
n
c
e of
the
SBFRA Me
ch
ani
sm, in
whi
c
h
use
r
s 2,
3,4,7,8 an
d
10
req
u
ire
a
dditional
re
source,
wh
ere
a
s
only
1, 5
and
9
chan
nel
con
d
ition
i
s
according
to
the requi
re
d se
rvice, so witho
u
t
ad
ditional
re
so
urce thei
r re
quire
d
servi
c
e is
provide
d
imm
ediately. Whe
r
ea
s the re
st of the users i
s
pla
c
ed in th
e F
p
, S
p
and
T
p
accordi
ng
to
their ch
annel
co
ndition, b
u
t
after a ce
rtain peri
od
of time eve
r
y use
r
re
cei
v
ed its
re
qui
red
serv
i
c
e.
3.2 Simulation Scenario
2.
Figure 7 p
r
e
s
ents the
overall perfo
rma
n
c
e of
the
SBFRA Me
ch
ani
sm, in
whi
c
h
use
r
s 2,
3,4,7,8 an
d
10
req
u
ire
a
dditional
re
source,
wh
ere
a
s
only
1, 5
and
9
chan
nel
con
d
ition
i
s
according
to
the requi
re
d se
rvice, so witho
u
t
ad
ditional
re
so
urce thei
r re
quire
d
servi
c
e is
provide
d
imm
ediately. The
rest
of the u
s
ers a
r
e pl
aced in the F
p
, S
p
and T
p
according to th
eir
cha
nnel
con
d
i
tion, but after a certain p
e
ri
od of
time every use
r
re
cei
v
ed its requi
red se
rvice.
Figure 8.Com
pari
s
on of ch
annel utili
zation with an
d without fairne
ss for ertPS cl
ass
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
ubs
c
r
i
ber
C
hannel
U
t
iliz
a
t
io
n
C
hannel
U
t
iliz
a
t
io
n
w
i
t
h
o
u
t
Fa
ir
n
e
s
s
C
hannel
U
t
iliz
a
t
io
n
w
i
t
h
Fa
ir
n
e
s
s
T
h
r
e
s
hol
d Lev
el
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
u
b
scr
i
b
e
r
C
hann
el
U
t
i
l
i
z
at
i
o
n
er
t
P
S
w
i
t
hout
F
a
i
r
n
e
s
s
e
r
tPS w
i
th
F
a
i
r
n
e
s
s
T
h
r
e
s
hol
d Lev
el
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 2, June 20
14: 455 – 46
4
462
Figure 8 sh
o
w
s that all 10
sub
s
crib
ers require
the ert
PS service, b
u
t 1,5 and 10
cha
nnel
con
d
ition is n
o
t as pe
r the requi
re
d para
m
eter, so u
s
e
r
s a
r
e pla
c
e
d
in their re
sp
e
c
tive queu
es
for an additio
nal re
sou
r
ce, and when the
requi
red i
s
provided by the
Resource M
ange
r then B
S
provide
s
the
m
the requi
re
d servi
c
e.
Figure 9.Com
pari
s
on of ch
annel utili
zation with an
d without fairne
ss for nrtPS cl
ass
Usi
ng sam
e
cha
nnel co
nd
ition,
the
exp
e
ri
me
nts
are
perfo
rmed
by
ch
angin
g
th
e othe
r
servi
c
e
nrtPS. Figure
9 p
r
ese
n
ts that i
n
pro
p
o
s
ed
m
odel u
s
e
r
s 1,
2, 3 an
d 4
require a
dditio
nal
resou
r
ces for providing th
e requi
re
d se
rvice.
The p
r
opo
sed mo
d
e
l is appli
ed
on every service,
and it is fo
un
d that for diff
erent u
s
e
r
s require
ad
ditio
nal re
so
urce
s. In Figure
1
0
, the propo
sed
model i
s
eval
uated o
n
rtP
S
servi
c
e a
n
d
marked
2,
5,
7 and
8 for requiri
ng a
ddit
i
onal resou
r
ces.
Whe
r
ea
s in F
i
gure 1
1
only 1, 8 and 9 re
quire m
o
re
re
sou
r
ce for BE service.
Figure 10.Re
sou
r
ce utilizat
ion ba
sed o
n
cha
nnel
con
d
i
tion for rtPS Service
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
ubs
c
r
i
b
er
C
h
a
n
n
e
l
U
t
iliz
a
t
io
n
nr
t
P
S
w
i
t
hout
F
a
i
r
n
e
s
s
nr
t
P
S
w
i
t
h
F
a
i
r
nes
s
T
h
r
e
s
h
o
l
d Lev
el
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
u
bs
c
r
i
ber
C
hannel
U
t
iliz
a
t
io
n
r
t
PS w
i
t
h
o
u
t F
a
i
r
n
e
s
s
r
t
P
S
w
i
th
F
a
i
r
n
e
ss
T
h
r
e
s
h
o
l
d Lev
el
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Additional Re
sou
r
ce Alloca
tion for Im
proving Fai
r
n
e
ss in WiMAX (Khalid Mahm
o
od Awan
)
463
Figure 11.Re
sou
r
ce utilizat
ion ba
sed o
n
cha
nnel
con
d
i
tion for BE service
Figure 12 p
r
e
s
ent
s the
su
mmary at BS level
whe
r
e
multiple servi
c
e
s
are dem
ande
d by
the SS, and
SBFRA ha
s
queu
ed
all th
ose
services
whi
c
h
re
quire ad
ditional
reso
urce
s. After
providin
g
the
req
u
ired re
source every SS
entertain
s
with that
servic
e. In this
proc
es
s thos
e
use
r
s
which
required
more
than
20%
ad
ditional
re
sou
r
ce
h
a
s be
en
comp
romi
se
d
on
bit d
e
lay f
o
r
providin
g the
said
se
rvice. The
solid t
r
e
nd line
pr
ese
n
ts that eve
r
y se
rvice
ho
w
many ad
ditio
nal
resou
r
ces a
r
e
requi
red by e
v
ery servi
c
e
and the pr
op
ose
d
model p
r
ovide
s
the re
quire
d re
so
urce
as pe
r the pre-set prio
rity on the ba
sis
of t
he assum
ed se
rvice p
r
i
o
rity scheduli
ng mechani
sm.
Figure 12.Co
mpari
s
o
n
of chann
el utilizat
ion with
fairn
e
ss for differe
nt classe
s of VBR traffic
4. Conclusio
n
In this paper,
a SBFRA Mech
ani
sm is prop
os
ed by usin
g cha
nne
l status. After studyin
g
the problem
of fair resou
r
ce m
ana
gem
ent the p
r
o
p
o
se
d me
ch
a
n
ism
allo
cate
s b
a
se
d o
n
t
h
e
status
of ch
annel fai
r
ly and u
s
e
r
satisfacti
o
n
le
vel is achie
v
ed. Simulation shows t
h
e
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
S
ubs
c
r
i
b
er
C
h
anne
l
U
t
iliz
a
t
io
n
B
E
w
i
t
hou
t
F
a
i
r
nes
s
BE w
i
th
F
a
i
r
n
e
s
s
T
h
r
e
s
hol
d
Le
v
e
l
1
2
3
4
5
6
7
8
9
10
0
5
10
15
20
25
30
35
S
ubs
c
r
i
ber
C
h
a
nnel
U
t
i
l
i
z
a
t
i
o
n
er
t
P
S
nr
t
P
S
rt
P
S
BE
T
h
r
e
s
hol
d Lev
el
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 2, June 20
14: 455 – 46
4
464
improvem
ent
in the allocation of re
sou
r
ces in te
rms o
f
fairness whi
c
h sho
w
s th
a
t
the addition
al
resou
r
ce i
s
allocated to
the u
s
e
r
with lo
w S
N
R
value in
o
r
d
e
r to
a
c
hiev
e the
minim
u
m
satisfa
c
tion.
The expe
rim
ents
sho
w
th
at both co
nst
ant bit rate and variable bit rate traffic
gets
improvem
ent
in fairn
e
ss
an
d level
of u
s
e
r
a
c
cept
an
ce.
Thi
s
me
ch
an
ism al
so
imp
r
oves
wait tim
e
for lo
w
prio
rity cla
s
ses to
provide
fair a
nd d
e
si
red
service
to
all t
he u
s
e
r
s
without affectin
g
the
QoS u
p
to th
e maximum
a
v
ailability of reso
urce
s.
Thi
s
m
e
chani
sm
doe
s
not p
e
rform effici
entl
y
if
the use
r
s
with low SNR wi
ll be greate
r
i
n
numbe
r.
Ackn
o
w
l
e
dg
ement
This
re
sea
r
ch is
su
ppo
rte
d
by the Mi
ni
st
ry of
Highe
r Edu
c
ation
Malaysia
(M
OHE) and
wa
s con
d
u
c
ted in colla
boration with the
Rese
arch
M
anag
ement Centre (RM
C
)
at the Univeri
s
ti
Tekn
ologi M
a
laysia (UTM
) unde
r Vot Nu
mber Q. J 1
3
0000.2
528.0
3
H
19.
Referen
ces
[1]
Rua
ngch
a
ij
atu
pon N, Y Ji.
Proporti
ona
l fa
irness w
i
th
mi
ni
mu
m rate
g
uara
n
tees sch
edu
lin
g in
a
mu
ltius
e
r OFDMA w
i
reless ne
tw
ork
. Leipzig. 200
9.
[2]
Haita
ng W
,
W
e
i L, D
harm
a
P.
Agra
w
a
l.
Dynam
ic
admission
control an
d QoS for IEEE 802.16Wireless
MAN
. Proc. of
W
i
reless T
e
lecommunic
a
tio
n
s
S
y
mp
osium.
WT
S. 2005: 60-66.
[3]
F
a
rhad
i R. A N
o
vel Cr
oss-L
a
y
er Sche
dul
in
g
Algorit
hm for OF
DMA-Based
W
i
MAX N
e
t
w
or
ks.
Int. J. of
Co
mmun
icati
o
ns, Netw
ork and System Sci
e
nces.
201
1; 4(
2): 98-10
3.
[4]
Cho
w
d
h
u
r
y
,
P
r
asun, Iti SM.
A Fair a
nd
Efficient Pack
et Sche
dul
ing
Scheme
for
IEEE 802.1
6
Broadband Wireless Access S
y
stems.
ArXiv prepri
n
t arXiv.
201
0; 100
9: 60
91.
[5]
Mua
y
a
d
S, Al
S. Sched
uli
n
g an
d R
e
sour
ce All
o
catio
n
Strateg
y
for OF
DMA S
y
stem
over T
i
me
Var
y
in
g Cha
n
n
e
ls.
Int J Wireless Inf Networks.
2011.
[6]
J Lin, H Sirisena
. Quality of
Service Sche
d
u
lin
g in
IEEE
802.16 Broa
d
ban
d Wireless
Netw
orks
.
Procee
din
g
s o
f
F
i
rst
Internation
a
l Co
nfere
n
c
e
on Ind
u
stri
al an
d Inform
ation S
y
stems.
200
6: 396-
401.
[7]
J Sun, Y Y
ao,
H Z
hu.
Qua
lity
of Service
Sch
edu
lin
g F
o
r
80
2.16 Br
oad
ba
n
d
W
i
rel
e
ss Acc
e
ss Syste
m
.
Advanc
ed s
y
stem techno
lo
g
y
teleco
m l
ab IEEE. Beijin
g, Ch
ina. 20
06.
[8]
K Wongthav
ar
a
w
at, A Gan
z
. Packet sched
uli
ng for
QoS supp
ort i
n
IEEE 80
2.1
6
bro
adb
an
d
w
i
reless acces
s
s
y
stems.
Internatio
nal Jo
urn
a
l of Co
mmu
n
i
c
ation Syste
m
s
.
2003; 16(
1): 81-96.
[9]
JH Jeo
n
, JT
Lim. D
y
n
a
mic
b
and
w
i
dth
al
lo
cation
for Qo
S in IEEE
80
2.16
broa
db
a
nd
w
i
rel
e
ss
access net
w
o
r
k
.
IEICE
Trans Communication
. 2008; E9
1-B
(
8): 2707-2
7
1
0
.
[10]
J Lin, H Siris
ena.
Qual
ity of Service Sch
edu
lin
g in IE
E
E
802.16 Bro
adb
an
d W
i
rele
ssNetw
o
rks
.
Procee
din
g
s of
F
i
rst International C
onfere
n
ce
on I
ndustria
l a
nd Informatio
n
S
y
stems. 20
06
: 396-40
1.
[11]
H Safa,
H Art
a
il, M
Karam,
R S
oud
ah, S
Kh
yat.
New
Scheduling Ar
chitecture for I
EEE 802.16
Wireless Metro
polit
an Are
a
Netw
ork
. Ameri
c
an un
iversit
y
of Beirut. IEEE. Leban
on. 20
0
7
.
[12] M Settembre, M Puler
i
, S Garritano,
P
T
e
sta, R Albanese, M Manci
n
i
,
V Lo C.
Perfo
r
ma
nce a
naly
s
is
of an efficie
n
t
packet-bas
e
d
IEEE 802.
16 MAC sup
portin
g
ad
apti
v
e modu
lati
on
and co
di
ng.
Procee
din
g
s of
Internation
a
l S
y
mp
osi
u
m on
Comp
uter Net
w
o
r
ks. 200
6: pp.11-1
6
.
[13]
So-In, Chakc
h
ai, Ra
j J, Abd
e
l-Karim
AT
.
G
enera
lize
d
w
e
ig
hted fa
irne
ss and
its ap
plicati
on f
o
r
resourc
e
a
lloc
a
tion
in IEEE
80
2.16
e m
obi
le
w
i
ma
x.
C
o
mp
u
t
er an
d A
u
to
mation
En
gin
eeri
ng (IC
C
AE),
IEEE.
2010; 1: 784-
788.
[14]
IEEE Std. 802.16e, IEEE St
andar
d for local
and me
tropol
itan ar
e
a
net
w
o
rks,
part 16: Air
Interface for F
i
xe
d a
nd M
o
b
i
l
e
Broa
db
an
d
W
i
rele
ss Acc
e
ss S
y
stems, A
m
endm
ent 2:
Ph
y
s
ica
l
a
n
d
Medi
um Access Control La
yers for Com
b
in
ed F
i
xed a
nd Mobi
le Op
eratio
n in Lic
ense
d
Ban
d
and C
o
rrig
e
n
d
u
m 1”, Ma
y
2
0
0
5
[15] DS
Shu. A Cross
Lay
er approac
h for
Packet
Sc
heduling
at Do
w
n
link of
WiMA
X
of IEEE802.16e.
Europ
e
a
n
jour
nal of scie
n
tific
research
. 20
1
0
; 45: 529-
539.
[16]
W
i
MAX F
o
rum,
Krishna R, Ra
j
J.
W
i
MAX System Eval
uati
o
n
Methodo
lo
gy.Versio
n
2
. 200
8.
[17]
Khal
id MA, Abdu
l HA, Khal
id H. Ch
ann
el
Based R
e
so
urce All
o
cati
o
n
Mecha
n
ism
(CBRAM) i
n
WiMAX
.
Life S
c
ienc
e Journ
a
l
.
2013; 1
0
(3).
[18]
Mammeri Z, D
Bouzid, P Lorenz.
Auto
matic
Mapp
ing
of
Re
al-ti
m
e T
r
affic
Constra
i
nts o
n
to CBR
a
n
d
rt-VBR Services of ATM.
ATM (ICAT
M
2001) and H
i
g
h
Spee
d Intell
ige
n
t
Internet S
y
m
posi
u
m, 200
1.
Joint 4th IEEE Internatio
na
l C
onfere
n
ce o
n
. IEEE, 2001.
Evaluation Warning : The document was created with Spire.PDF for Python.