TELKOM
NIKA
, Vol.11, No
.5, May 2013, pp. 2739 ~
2
746
ISSN: 2302-4
046
2739
Re
cei
v
ed
De
cem
ber 2
0
, 2012; Re
vi
sed
March 19, 20
13; Accepted
March 26, 20
13
Intelligent Caching Wireless D
a
ta Access in t
h
e
Wireless Spectrum
Sy
az
w
a
Mad
Jais*, Sarina Sulaiman,
Siti Mariy
a
m Shamsuddin
Soft Computin
g Rese
arch Group
F
a
cult
y
of Com
puter Scie
nce
and Informati
o
n
S
y
stems
Univers
i
ti T
e
knolo
g
i Mal
a
ysia,
Johor, Mala
ys
i
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: s
y
az
w
a
_
w
a
w
a@
ya
hoo.com.
m
y
, sarin
a
@
u
tm.m
y
,
mari
ya
m@utm.m
y
Ab
stra
ct
T
he
evo
l
uti
on w
i
reless
tech
n
o
lo
gies are gr
ow
ing
ra
pi
dly i
n
upc
o
m
i
ng y
e
ars. It is expe
cted tha
t
ma
ny users w
i
ll shift to more
advanc
ed
de
vices that w
ill
contrib
u
te to g
a
in h
i
g
her d
e
m
a
nd i
n
w
i
rele
ss
spectru
m
. How
e
ver, the cap
a
c
ity for the allo
cation freq
u
e
n
cy in the w
i
rele
ss spectrum is
typically l
i
m
ite
d
in
w
i
reless d
a
ta t
r
ans
missi
on. T
herefor
e, w
hen
the lo
ads
of w
i
reless us
ers
in the w
i
re
les
s
commun
i
cati
ons
are i
n
cre
a
sin
g
, the
nee
d
of c
a
che
mech
an
i
s
ms s
u
ch
as
W
eb cac
hes
o
r
W
eb serv
ers
are cr
uci
a
l. T
h
e
scala
bil
i
ty de
ma
nds o
n
int
e
rnet infrastru
c
ture keep
i
n
creasi
ng as the inter
net co
ntinu
e
s to gro
w
in
pop
ular
ity an
d si
z
e
. Th
erefor
e
,
the
existenc
e
and
deve
l
o
p
m
ent in W
eb c
a
c
h
in
g techn
o
l
ogi
es w
ill contri
bu
te
to ban
dw
idth s
a
vin
g
s, netw
o
r
k
latency r
e
d
u
c
t
ion, i
m
pr
ove c
ontent
avai
lab
i
l
i
ty and
s
ubs
eq
uently s
e
rver l
o
ad
bal
anci
ng. T
h
i
s
paper w
i
l
l
studi
es an
d inv
e
stigates the ca
che perf
o
rman
c
e in w
i
rel
e
ss spectru
m
w
i
th th
e
purp
o
se
of d
e
a
lin
g w
i
th
the
data
grow
th si
nce th
e s
pectr
um crisis
b
e
co
mes
a
seri
ous
matter l
a
tely.
T
h
e
perfor
m
a
n
ce
i
m
pr
ove
m
ent w
ill
be
obs
erved
usin
g cac
h
i
n
g
sche
m
e
w
h
ich
all
o
w
s
for ti
me shifti
ng
and
l
o
a
d
shifting in accessing the wirele
ss data with the better cac
h
e dep
loy
m
ent in the network system
.
Ke
y
w
ords
: caching, wireless spectrum
, fr
eq
uency a
lloc
a
tio
n
, cellu
lar netw
o
rk
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
Wirel
e
ss spe
c
trum i
s
the publi
c
radi
o wave
s whi
c
h
refer to ele
c
tromag
netic
spectrum
that have th
e wi
rele
ss
capa
city whi
c
h comp
risi
ng
of different
freque
ncy
ra
nge
s for diff
erent
types
of app
lication. For instan
ce
s,
all
wi
rele
ss
sig
nals such a
s
televisio
n
,
Global
Positi
on
System (GPS), bro
adb
an
d an
d radio
that travel
th
rough
the
air
on p
ubli
c
ov
er g
eog
ra
phi
cal
distan
ce
whi
c
h
are
the i
n
visible
sig
n
a
ls al
so
call
ed
spe
c
trum.
The
s
e
days, wirel
e
ss
ra
dio
spe
c
tru
m
be
comes
a cru
c
i
a
l issue in th
e up
comin
g
year b
e
cau
s
e i
t
is a finite precio
us
re
sou
r
ce.
Wirel
e
ss net
work ope
rato
rs have
thei
r
own
spec
t
r
u
m
allo
cation
that they ha
ve boug
ht from
comm
uni
cati
on com
m
issio
n
for every si
ngl
e ap
plicati
on that being
impleme
n
ted.
Ho
wever, m
o
st of the
net
work
ope
rato
rs
clai
m
that
their bl
ocks o
f
spe
c
tru
m
cannot b
e
su
staine
d wit
h
the expl
osi
v
e dema
nd t
hat can
slug
gish
wi
rele
ss intern
et serv
ice
su
ch
wh
a
t
is
being fa
ced
by Smartpho
ne and i
-
Pho
ne users an
d
also forcin
g drop
ped
cellu
lar pho
ne
cal
l
s.
They cann
ot
use
mo
re
sp
ectru
m
tha
n
they
have
bo
ught for thei
r user
s’ sa
ke.
Co
nseque
ntly,
they are fight
ing to
get
mo
re
blo
c
ks
of
spe
c
tr
u
m
fro
m
unused television spe
c
trum that is also
kno
w
n
as
whi
t
e spa
c
e. Thi
s
white sp
ace
spe
c
tru
m
so
metimes
calle
d su
per
Wi-Fi
is allo
cated f
o
r
television n
e
twork, but it i
s
not in u
s
e
by t
he television
station.
One of the
mobile n
e
two
r
k
provide
r
s AT&T rece
ntly,
prop
osed spe
c
trum
comb
i
n
ation tech
niq
ue with othe
r mobile network
operator
whi
c
h is T-M
obile
for AT&T’s comp
any
to h
e
lp them in
mobile
servi
c
e for their u
s
ers
in Ameri
c
a. If
the d
eal
goe
s th
roug
h, it d
oes not
put a
n
y more
spe
c
trum into
wi
re
less b
r
oa
dba
nd
overall [1].
Neverth
e
le
ss,
the efficient
use
of the m
e
rge
r
can i
m
proved
its o
w
n se
rvice. On
e thing
that nee
ds to
be
su
re to
a
v
oid the inte
rferen
ce
between
sign
als i
s
two
signal
s that are b
e
i
n
g
broa
dcast ca
nnot be in the
same sli
c
e of
spe
c
trum o
r
otherwise the
signal
s will b
e
com
e
usele
ss.
Ho
wever,
the
Fed
e
ral
Co
mmuni
cation
s
Commi
ssio
n (F
CC) en
su
res that if
an
y two tran
smi
tters
(whi
ch
are
ce
ll towers) i
n
o
ne area an
d frequ
en
cy
, the signal
s
will n
o
t overlap a
n
d
interferes t
o
each other [
2
]. If the frequen
cy ran
g
e
s that bei
ng
allocated a
n
d
the implem
entation are not
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 5, May 2013 : 2739 – 274
6
2740
organized in efficient way, a
ccess to the wireless spectrum
will be a problem in wi
reless
serv
i
c
e
s
.
The
network that b
e
ing
built by th
e
wirel
e
ss com
pany i
s
ba
se
d on
the
ran
k
of
the
averag
e expe
cted u
s
a
ge.
The mo
st cases that h
ad b
een fa
ced
by the network
provide
r
a
r
e t
h
e
huge
su
rg
e i
n
volume th
a
t
flooded th
e
wirele
ss
sp
e
c
trum, l
eadin
g
to many
b
u
sy
signal
s [
13].
This
situation
happe
n whe
n
, there i
s
hu
ge num
ber
of use
r
s t
r
ies to
make
call
s a
t
the same ti
me.
For insta
n
ce, the m
o
me
nt wh
en
wi
rele
ss get
s
bla
c
ked o
u
t in
Ne
w Yo
rk o
n
2
003
sh
ows a
n
d
proved
to th
e
network
prov
ider and
cell
phon
e u
s
e
r
s
that the
network u
s
a
ge th
at bein
g
u
s
e
d
at
that time was overload [3].
A study fro
m
Cisco
sai
d
that band
width
de
man
d
s a
r
e expe
cted to g
r
o
w
30 times
betwe
en 201
0 and 201
5 [4]. Having talked to a num
ber
of co
mpa
n
ies, it seem
s that there a
r
e
four ba
si
c techniqu
es for
a
ddre
s
sing the
probl
em
ove
r
the next few years which are a
dditio
nal
spe
c
tru
m
, be
tter use of t
hat spe
c
trum
, offloading
data o
n
to
wired
networks, and i
n
crea
sed
ca
chin
g [4].
One of the solution
s to overcome
the p
r
oble
m
that have been
sta
t
ed is ca
chi
n
g.
Ca
chin
g is o
n
e
of the impo
rtant metho
d
s that c
an
be u
s
ed to
deal
with the data traffic espe
ciall
y
involving the
video data [4]
.
It is better to pre
d
ic
t
wha
t
content u
s
e
r
s will
want a
n
d
then
ca
che
it
to their devices, eithe
r
wh
en they are
con
n
e
c
t
ed to
a wire
d net
work, or
wh
e
n
the wirele
ss
netwo
rk i
s
un
derutili
zed [4]
.
The idea is
to le
verage l
o
cal
storage
to intelligently cache net
wo
rk
conte
n
t whi
c
h allows fo
r time shiftin
g
and m
o
re
importa
ntly load shifting whi
c
h
can
be
impleme
n
ted
by moving the band
width requireme
nts
from cong
este
d netwo
rk at
a busy time to
downloadi
ng
from a net
wo
rk in n
on-pea
k time (say
o
v
ernight
) or from a Wi
-Fi n
e
twork [4]. For
example, let
say a
use
r
watche
d the first few
epi
so
d
e
s of
a dram
a se
rie
s
, it
wil
l
be expe
cted
that
the
user will want
to watch
the
next episodes
until the end of the
dram
a. So, this i
s
where the
caching system will take the role
to cache the next epi
sodes of the dram
a to
the user’s device
within non-peak
time.
1.1. Summary
of Crucial Issues o
f
Finite Wirele
ss
Spectr
u
m an
d Caching S
c
heme
Intelligent ap
proa
ch
es i
n
Web
ca
chi
ng
are im
porta
nt to wirele
ss
n
e
twork
enviro
n
ment in
orde
r to
ada
pt with the
ra
pid g
r
o
w
th of
the net
wo
rk.
The
r
e a
r
e
a
numb
e
r
of o
ngoin
g
effort
s in
Web
ca
chin
g
technol
ogie
s
that have the solutio
n
s to
overcome th
e pro
b
lem
s
that occu
rs in
the
wirel
e
ss
network.
Even th
o
ugh th
ere a
r
e
many
studi
e
s
in
this kind
of
tech
nolo
g
i
e
s,
the
shifting
method i
n
te
rms of time
shifting
and
l
oad
shifting
are
still n
o
ve
l in
Web
caching. By goi
ng
throug
h the
other re
se
archer pap
ers, j
ourn
a
ls an
d
web
s
ite, it
wi
ll help
the
au
thors to
have
a
good vie
w
on
Web
ca
chin
g and lo
ok
u
p
for bette
r
n
e
w
solution t
o
overco
m
e
the problem
s
that
occurs in the
wirel
e
ss com
m
unication n
e
twork
.
Talley (199
7) stated that the Artour G
a
te
way and WebExpre
ss use sm
art ca
ching an
d
solid
man
age
ment to g
a
in
high
er
sp
ee
d in a
c
ce
ssi
n
g
wi
rele
ss n
e
twork. All I
P
function
s,
with
data en
crypti
on an
d comp
ressio
n a
r
e p
r
ovided
by th
e Gate
way. After the data i
s
bei
ng
re
cei
v
ed
by the server, the data will be translate
d to the appr
opriate n
e
two
r
k p
r
oto
c
ol b
e
fore it send
s to
the client thro
ugh wi
rele
ss netwo
rk. Th
e
conn
ectio
n
to the internet
and intran
et can b
e
don
e by
usin
g Gate
way alone a
n
d
then the no
rmal intern
et
activities
can
be eng
age
d. Neverth
e
le
ss, it
will be quite slow i
n
Web browsing
since the
pr
oduct offers
only
gener
al
IP transport
with
bro
w
si
ng
sp
e
ed at
9600
bp
s (at be
st) which
is ve
ry
painful. Th
ere is a
com
b
i
nation
of sev
e
ral
ways to o
p
timize the
We
b bro
w
sing b
y
the WebE
xpre
ss i
n
orde
r to make the
slow
co
nne
ction
seem m
u
ch
faster. Web
E
xpress intel
ligent ca
chin
g techniq
u
e
cache
s
different pa
ge
s for
different le
ngt
hs of time
[5]. A parti
cula
r
page
ca
n al
ways be
provid
ed from
cach
e by setting up
the system a
nd for differe
nt refre
s
hin
g
strategi
es
an
d ca
chin
g also can b
e
do
n
e
by configu
r
i
ng
the addition
al page
s.
WebExp
re
ss also offe
rs retrieving
document
s i
n
the ba
ckg
r
oun
d
automatically, so when the
use
r
ne
ed
s them, t
he
do
cuments are a
l
ready
the
r
e. The
do
cum
e
nts
can
be do
wnl
oade
d immed
i
ately or leav
e them in
the
backg
rou
nd
as the
syste
m
is directe
d
by
the u
s
er.
Howeve
r, the
lack of
ce
ntralized
ad
ministration t
ools is
an
obsta
cle
to
the
WebExp
re
ss.
This is be
ca
use
the
ca
chi
ng info
rmat
io
n ne
ed
s to b
e
man
ually a
dd to
each of
the
client
s units b
y
the administrator if an intranet serve
r
is add
ed.
The
su
rvey p
aper that
bee
n wr
itten by Barish
a
nd O
b
ra
czka
(200
0)
describe
the trend
s
and techniq
u
e
that being
use
d
in Web
ca
chin
g te
ch
nologi
es. Th
e
y
stated that Web
ca
chin
g
is
still young in
dustry a
nd a
numbe
r of
comm
ercial v
endo
rs
pu
shi
ng ne
w solut
i
ons eith
er
with
dire
ct to rese
arch system
s or architecte
d by i
ndividuals with nota
b
le research b
a
ckgroun
d [6]. In
the article, several cachi
ng archite
c
tu
res,
the sele
ction of ca
che depl
oyme
nt, and spe
c
ific
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Intelligent Ca
chin
g Wirele
ss Data A
c
cess in the Wi
rel
e
ss Spect
r
um
(Sya
zwa Ma
d Jai
s
)
2741
desi
gn tech
ni
que
s are b
e
in
g pre
s
ente
d
. A number
of
ongoi
ng issu
es are rem
a
in
ing even thou
gh
better pe
rformance and
scala
b
ility improveme
n
t is
being offered
by Web ca
ching. One of the
issue
s
i
s
the
obje
c
t inte
grity. Let say if
an
obj
e
c
t i
s
ca
ch
ed, i
s
t
he u
s
e
r
gua
ranteed
that t
he
c
a
c
h
ed
c
o
p
y
is
up
-
t
o-
d
a
t
e
w
i
th
th
e ve
rs
io
n
on
th
e
o
r
i
g
inating
serv
er [6]. Som
e
techni
cal
issu
es
are
co
ntent secu
rity, the p
r
acti
cally of h
andlin
g dyna
mic a
nd
real
time data, an
d deali
ng
wit
h
compl
e
x functional obje
c
ts [6]. Different sce
na
ri
os m
a
y apply different archite
c
t
u
re
s dep
endi
ng
on the
suitab
ility of the system
by the
deploye
r
. Ho
wever, to
ch
oose which o
ne is th
e be
st is
quite difficult
since there
are ma
ny cachi
ng soluti
on that exist
nowa
days
and the
ca
chin
g
scheme n
o
rm
ally depend
s
on the nee
d a
nd co
nstraint
s of the deplo
y
er.
Tay, et al. (2000
) introd
uce
a cachin
g schem
e which i
s
a
rev
o
lutiona
ry di
stribute
d
internet
caching
system. I
n
this pa
pe
r,
the
auth
o
rs
p
r
opo
se
d n
e
w structu
r
e,
ca
che
mo
del
a
nd
proto
c
ol th
at will
be im
pl
emented
in t
heir
re
se
arch. The
theo
retical m
odel
that ha
s
b
een
prop
osed ha
s the large
s
t combine
d
storage spac
e an
d has a gu
aranteed data
consi
s
ten
c
y. This
kind of
ca
chi
ng syst
em in
theoreti
c
ally i
s
mo
re
reliabl
e com
pare to
any singl
e or grou
p of ca
che
serve
r
s. However, the
met
hod to
utilize
the ad
di
tional
po
wer of the
clie
nt no
de
s
become
a
ke
y
issue
s
for th
e
authors in
order to in
crea
se the
pe
rformance of the
se
rver fun
c
ti
ons
so th
at the
overall netwo
rk
will not
aff
e
cted
by the
i
nherent u
n
rel
i
ability ch
ara
c
teristic.
A cli
e
nt co
mpute
r
wil
l
have as its fi
rst pri
o
rity pe
rform
s
locally init
iated tasks and the
se
con
d
prio
rity task
will be the
server functi
on and thi
s
will add to
several
effe
cts such as the client
s com
p
uter will
become
unreli
able fro
m
the serve
r
functio
n
s.
The in
cid
ent
that hap
pen
s in
Ne
w Yo
rk on
200
3 i
s
sho
c
king
eve
r
y re
sid
ent
si
nce
the
blacko
ut also
give an impact to wireless network
as
stated by La
Monica (2
003
). The situati
o
n
become fru
s
t
r
ating an
d turns o
u
t to be more ann
oying whe
n
ma
ny wirele
ss custome
r
s we
re
unabl
e to
use
their mo
bile
phon
e d
u
rin
g
that time. T
h
is
circum
stan
ce
hap
pen
b
e
ca
use of
hu
ge
numbe
r of
ca
llers t
r
ie
s to
make
a call
at the sa
me
time. This
wil
l
contri
bute t
o
hug
e surg
e
in
volume and l
ead to many
busy si
gnal
s
and drop call
s. Spectrum i
s
a finite re
source
s and t
h
e
netwo
rk p
r
ovi
der buil
d
up their net
wo
rk
within
the fre
quen
cy allo
ca
tion that have been gai
n from
the a
s
sociate
d
gove
r
nme
n
t
commi
ssion
e
r b
a
sed
o
n
averag
e exp
e
c
ted
usage l
e
vels. Since th
e
wirel
e
ss is to
o de
pen
dent
on the
extern
al po
we
r
s
u
p
p
ly, the bl
ackout shoul
d p
u
t
pre
s
sure
on
the
major
wirel
e
ss ca
rri
ers to bulk up
covera
ge in al
re
ady high-t
r
affic ne
twork and ta
kes mo
re ste
p
s
to ensu
r
e that
wirele
ss ante
nna
s ca
n ke
e
p
r
unni
ng eve
n
in the event of a power o
u
tage [6].
The deal of spe
c
tru
m
me
rge
r
by the AT&T
and T-Mobile give an issue to the Ovide
(201
1) rega
rd
ing on the lim
ited re
sou
r
ce
which is spe
c
trum. Th
e d
e
finition of sp
ectru
m
and t
h
e
importa
nce for u
s
e
r
s
of mobile d
e
vices al
so
have
been
provid
ed thro
ugh t
h
is p
ape
r. AT&T
proposed $39 billion to T-Mobile for the merger to
help the com
p
any to meet the demand of the
netwo
rk. If the merge
r
go
es thro
ugh, the deal do
es not put any
more
spe
c
tru
m
into wirele
ss
netwo
rk overall, but the
co
mbinat
ion
ca
n improve its
own
se
rvice.
Nev
e
rt
h
e
le
ss,
t
he in
suf
f
i
cie
n
t
spe
c
tru
m
limi
t
s a ca
rri
er’
s
coverage,
service
quality a
nd data
conn
ection
spe
e
d
s
[1]. Since the
spe
c
tru
m
ava
ilable fo
r th
e
netwo
rk p
r
ovi
der is al
ready
fixed for the
m
, they ne
ed
to organi
ze
th
eir
netwo
rk in eff
i
cient way to give better se
rvice to their
use
r
s.
Table 1. Summary of cru
c
ial issu
es of
finite Wirele
ss Spect
r
um a
nd Ca
chi
ng Schem
e
A
u
t
h
or
Brief
Des
c
ripti
o
n
Limit
a
ti
ons
T
a
lley
(1997) [
5
]
Gate
w
a
y client component p
r
ovides IP
connectivity
over
non-IP net
works and
provides all IP functions as w
e
ll as data
encr
y
ption an
d compression.
WebExpress add
s complementar
y
client and server pieces. It caches
different pages f
o
r different leng
t
h
s of
time.
Web bro
w
sing
would be slow
b
e
cause
the product off
e
r
s
only
:
1.
Gene
ral IP trans
port
2.
Web Bro
w
sing at
9600bps (at b
e
st)
WebExpress lack of centralized
administration tools.
Barish
and
O
b
r
a
czka
(2000) [
6
]
Surve
y
the state
of the art in caching
designs, presenting taxo
nom
y
of
architectures and
describing a variet
y
of
specific trends a
nd techniques.
To decide
w
h
ich one is the best since
the effectiveness of a Web Cachin
g
largel
y
depe
nd o
n
the need a
nd
constraints of the deplo
y
er
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 5, May 2013 : 2739 – 274
6
2742
The appe
al of Web Caching beca
u
se of
the unpar
ralled g
r
o
w
th of t
he Inter
net.
There
remain a n
u
mber of op
en
issues:
1. Content
securit
y
2.
The practicall
y
of
handling d
y
nami
c
and real time dat
a.
3.
Dealing w
i
th com
p
lex functional o
b
ject
(for e
x
ample Jav
a
progr
ams)
Ta
y
,
e
t
al.
(2000) [
7
]
In the prop
osed
structure, ever
y client
computer in the
net
w
o
rk takes on
the
additional role of a cache server,
serving the objects it had previously
requested an
d re
ceived to other client
computers that al
so request the o
b
ject.
A comprehensive set of protocol f
o
r the
access, storage and serving function is
developed.
When putting the
server functions as
second priorit
y
ta
sks:
1.
A definite quality
of service w
i
th
respect to the se
rver functions cannot
be guaran
teed.
2.
Makes the client
computer un
reliable
from the serve
r
f
unction perspective.
Ke
y
issues:
1.
The method
to ut
ilize the addition
al
po
w
e
r of t
he client nodes to incre
a
se
the perfo
rmance
of the server
functions.
2.
A direct implication of the p
r
opos
ed
OS initiated server functions in the
client node is sy
stem securit
y
.
La Monica
(2003) [
3
]
Explain the incident that happ
en in the
Ne
w
Y
o
rk met
r
o
politan area,
w
h
e
r
e
man
y
w
i
r
e
less cu
stomers, parts of
the
Midw
est and
Ca
nada
w
e
r
e
unabl
e to
use their cell phones during the
blackout, w
h
ich
made a frustr
ating
situation even more anno
ying.
In most cases, the biggest proble
m
w
a
s a h
uge surg
e in volume
(ever
y
bod
y all tr
ying to make calls at
the same time) t
hat flooded the
w
i
reless spectru
m
, leading to ma
n
y
bus
y
signals.
Wireless is too dependent on
the
exter
nal po
w
e
r suppl
y
and since
w
i
reless phones
are becoming th
e
main phone line for man
y
people,
the
industr
y
has to m
a
ke sure that
customers have the same kind of
reliability
as w
i
th landlines.
Ovide
( 2011)
[1]
Explain the definition of spectrum
and
the importance f
o
r users of mobil
e
devices.
Discussion on the spectrum mer
ger of
tw
o
net
w
o
rk prov
ider namel
y
A
T
&
T
and
T-Mobiles’s in order to look on th
e
improvement of t
he countr
y
’s
w
i
re
less
net
w
o
rk situation overall.
Because there ar
e onl
y
so man
y
usable frequencies, and things ar
e not
organized in the
most efficient w
a
y,
access to spectr
um is a limit
ing factor
for w
i
rel
e
s
s
s
e
rv
i
c
es
.
Insufficient spect
r
um limits a carrier’s
coverage, service qualit
y
,
and da
t
a
connection speeds.
1.2. Spectrum Allocations in Mala
y
s
i
a
The allo
catio
n
s of freq
ue
ncy sp
ect
r
u
m
in Malaysi
a
are all
o
cated by the M
a
laysia
n
Comm
uni
cati
ons an
d Mul
t
imedia
Com
m
issi
on
(M
CMC) in a
s
sig
n
ing th
e different f
r
eq
ue
ncy
allocation
s fo
r differe
nt ap
plicatio
ns. Ba
sically, the n
o
rmal f
r
eq
ue
ncy rang
es i
n
allo
cation
s of
spe
c
tru
m
tha
t
provided
b
y
governm
en
t commi
ssi
on
er a
r
e b
e
twe
en 3
kHz to
300G
Hz. F
o
r
microwave communi
catio
n
and Wi
rel
e
ss LA
N (L
ocal
Area Network), the freq
uen
cy rang
es are
betwe
en
300
MHz to
30G
Hz whe
r
eby t
he d
e
si
gnatio
n ba
nd
s
are
UHF
(Ultra
High F
r
eq
uen
cy)
and SHF
(Su
per Hi
gh Freq
uen
cy) as
sh
own in the Fi
gure 2.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Intelligent Ca
chin
g Wirele
ss Data A
c
cess in the Wi
rel
e
ss Spect
r
um
(Sya
zwa Ma
d Jai
s
)
2743
Figure 2. Spectrum Allo
cati
ons in Mal
a
ysia [2]
2. Rese
arch
Metho
d
This section
will present
a methodology
that
will
be i
m
plemented i
n
completion of this
resea
r
ch and
to enhance
the perform
ance of t
he data acce
ss in the wirel
e
ss
spe
c
tru
m
. A
flowchart of
the re
sea
r
ch
methodol
og
y is very
im
portant fo
r the proje
c
t to ensure th
at the
impleme
n
tation of the proj
ect ca
n be do
ne smo
o
thly with sp
ecifi
c
tasks.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 5, May 2013 : 2739 – 274
6
2744
The re
se
arch
is sta
r
ted wi
th
the study
and inve
stiga
t
ion
of the re
lationship bet
wee
n
Web
ca
chin
g
and wirele
ss sp
ect
r
um. This first sta
ge is e
s
sent
ial in orde
r
to develop an
unde
rsta
ndin
g
in both a
r
e
a
s
, namely
We
b ca
chi
ng an
d wirele
ss
sp
ectru
m
so tha
t
the kno
w
le
d
ge
that have b
e
e
n
gain
ed
ca
n
be a
pplied
through
this
re
search. T
he
cach
e po
sition
will b
e
the
n
e
xt
step to be d
o
ne for the sy
stem sin
c
e the
r
e are th
re
e main cache d
eployment
ch
oice
s in net
work
whi
c
h a
r
e
ne
ar the
conten
t con
s
um
er
(con
sum
e
r-
o
r
i
ented), nea
r
t
he content provider (p
rovid
e
r-
oriente
d
)
and
at strategi
c
points in the
network b
a
sed on
u
s
e
r
a
c
cess
pattern
s a
nd n
e
two
r
k
topology [6,
8]. Each
of t
he d
eployme
nt feat
ures h
a
s th
eir o
w
n
ben
efits
but
, to ap
ply th
e
deployme
nt, it will dep
end
on the
req
u
irement of t
he
netwo
rk sy
stem and
the
compatibility with
the system.
Input data or input messa
g
e
s are need
e
d
to have
the
simulatio
n
re
sult, so that the re
sult
can
be a
naly
z
ed
at the e
nd exe
c
ution
of the pr
oj
e
c
t. Whe
n
a
reque
st is m
a
de for a
n
obj
ect
curre
n
tly in th
e cache, it i
s
returned
to th
e cli
ent
but,
whe
n
a
requ
est i
s
m
ade
f
o
r
an
obje
c
t
not
curre
n
tly in the
ca
che, th
e cache
retri
e
ves th
e obj
ect a
nd
simu
ltaneou
sly fe
eds bytes to
the
client a
nd
sto
r
es the o
b
ject
into memo
ry.
If the network
at that time
is busy, it will
allows for time
shifting a
nd l
oad
shifting where it
will mo
ve the band
width req
u
ire
m
ents fro
m
con
geste
d net
wo
rk
at a b
u
sy tim
e
to d
o
wnloa
d
ing from
a n
e
twork i
n
n
o
n
-
pea
k tim
e
(say overni
ght) or from
a
Wi
-Fi
netwo
rk. La
st
ly, after the d
a
ta requi
red
have been
cach
e, the data will be sen
d
ing to the cli
ent
or user. Th
e rese
arch meth
odolo
g
y sho
u
l
d be com
p
let
ed as
sho
w
n i
n
flow ch
art in Figure 1.
Figure 1. Flowchart of the
methodol
ogy
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Intelligent Ca
chin
g Wirele
ss Data A
c
cess in the Wi
rel
e
ss Spect
r
um
(Sya
zwa Ma
d Jai
s
)
2745
3. Expecte
d
Resul
t
As dem
and f
o
r wi
rele
ss n
e
twork
comm
unicati
on
s co
ntinuou
s to e
x
plode, leveraging th
e
intelligent ca
chin
g wirele
ss data a
c
cess is vital to
the netwo
rk provider in o
r
d
e
r to give a h
i
gh-
quality consumer experience. In
this
research, the cachi
ng
scheme that will
be used i
s
the
ca
che
will
lea
r
n
and
predi
ct the d
a
ta e
s
peci
a
lly
data
that involved
with vide
o
co
ntent whe
r
e t
he
use
r
like
s
to surf and when
and where the use
r
wa
nts
to access it.
This
ca
chin
g
scheme
wi
ll also p
r
e
d
i
c
t network
availability and then th
e
better
oppo
rtunity to download th
e data
will be
cho
s
e
n
. The
expecte
d out
come f
r
om thi
s
re
se
arch i
s
to
give a ne
w hi
gh qu
ality experie
nce for
subscri
b
e
r
s
wheneve
r
an
d
whe
r
eve
r
the
y
want to a
c
cess
the data with no bufferi
ng
since it will predi
ct the
good wi
reless network t
hat available such as
Wi-Fi.
Thi
s
i
s
be
cau
s
e
pa
ckets a
r
e l
o
st
and
delay
e
d
whe
n
n
e
two
r
k con
g
e
s
tion occurs
that will
result in a
po
or u
s
e
r
expe
rience. Additio
nally
, studie
s
have sho
w
n
use
r
s have l
o
w tole
ran
c
e f
o
r
a poo
r video
experie
nce a
nd are qui
ck t
o
su
rf away
[9]. When ev
eryone i
s
si
m
u
ltaneo
usly tried
to access the same
conten
t that will l
e
ad to buffer, or delay
s,
most
of the
user
s will click away.
Refer to Fig
u
r
e 3 that hav
e been
sho
w
n belo
w
, with
in ten se
con
d
s, more tha
n
10 perce
nt o
f
audie
n
ces h
a
ve surf a
w
a
y
whe
n
the
conte
n
t that
being
surfed
by them
ne
ed b
u
ffering
and
delays.
Figure 3. Audience Attention spa
n
[4]
The idea of
this re
sea
r
ch
is to enable
caching dat
a conte
n
t for users in an
intelligent and
predi
ctive
wa
y to delive
r
t
he
req
u
e
s
ted
co
ntent
by downloadi
ng
it
duri
ng non
-pea
k
time a
n
d
sub
s
e
que
ntly maximize th
e use
r
’s exp
e
r
ien
c
e.
4. Conclusio
n
The in
crea
sin
g
traffic
dem
and i
n
wirel
e
ss spe
c
trum
every yea
r
gi
ves imp
a
ct to
most
of
the
net
wo
rk provide
r
s whi
c
h are stru
g
g
ling
to
have
eno
ugh
spe
c
trum
s to
m
eet the
growi
ng
deman
d. Th
e
network it
sel
f
is
a hyb
r
id
of landl
i
ne
co
nne
ction
s
tha
t
form th
e b
a
c
kbon
e n
e
twork
or pa
cket co
re and
wirel
e
ss p
o
rtion fo
r the last
mile [8]. Therefore, the pro
p
o
se
d intellige
n
t
ca
chin
g is very useful to allow on d
e
man
d
deliver
y to sup
port the h
uge growth of
network traffi
c.
Referen
ces
[1]
Ovide S. AT
&T/T
-Mobile D
eal:
Expl
ai
nin
g
W
i
reless Sp
ectru
m
.
T
he W
a
ll Street Journ
a
l.
2
011; 1.
[2]
La Monica PR.
W
i
reless gets
black
ed o
u
t too.
alt. cellular. v
e
rizo
n. 200
3: 1.
[3]
Miller
MJ.
T
h
e
Spectru
m
Cr
isi
s
: Deal
in
g W
i
th
the Grow
th in
W
i
reless
De
ma
nd
. PC M
agaz
i
ne, F
o
r
w
ar
d
T
h
inking. 20
11
: 1.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 5, May 2013 : 2739 – 274
6
2746
[4]
Ogle
J.
Stan
dard W
a
tch:
High Qu
ality
Mobil
e
Exp
e
ri
ence (HQME)
.
Complem
ent
ar
y
M
obi
le
Ecos
y
s
yt
em Report. 201
1.
[5]
T
a
lle
y
B. Artour Gate
w
a
y
an
d W
ebExpress
use smart cachin
g and so
li
d mana
geme
n
t to speed
u
p
w
i
rel
e
ss acces
s
. 1997.
[6]
Barish G a
nd
Obraczka
K. World Wid
e
Web
Cachi
ng: T
r
ends an
d T
e
chni
ques.
IEEE Comm
unications
Maga
z
i
ne, US
C Informati
on Scienc
es
Institute
. 200
0; 178-
181.
[7]
T
a
y
T
T
, F
eng Y and W
ije
ys
u
nder
a MN.
A Distribute
d
Internet Cach
in
g System.
20
00; 62
4-63
3.
[8]
Sula
iman S, Shamsu
ddi
n SM, Abraham A
and Sul
a
ima
n
S.
W
eb Caching a
nd Prefet
chin
g: W
hat,
W
h
y, and How
?
Internatio
na
l S
y
mp
osi
u
m on
Information T
e
chno
log
y
, IT
Si
m. 2008; 4: 1-8
.
[9]
Conneally
T
.
Sony an
d San
D
i
sk look to i
m
pr
ove vid
eo strea
m
s by pr
edicti
n
g w
hat you
’
l
l
w
ant to w
a
tch.
Be
ta
n
e
w
.
20
11: 1
.
Evaluation Warning : The document was created with Spire.PDF for Python.