Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
13
,
No.
3
,
Ma
rch
201
9
, p
p.
110
2
~
11
0
9
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v1
3
.i
3
.pp
110
2
-
1
1
09
1102
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
The ev
alu
atio
n
of AdBlo
ck
techni
qu
e implement
ation f
or
enterpri
se
n
etwork en
vironment
Mohd I
sk
an
d
ar
Bi
n
Sa
m
su
ddin
1
,
M
oham
ad
Yu
s
of
Da
r
us
2
, Sham
sul
J
El
ias
3
,
Ab
id
ah
Hj M
at Taib
4
,
N
orkhush
aini Aw
ang
5
, R
os
hidi
D
in
6
1,2,3,4,5
Facul
t
y
of C
om
pute
r
and
Mathe
m
at
i
ca
l
Sc
ie
nc
es
,
Univer
si
t
y
Technol
og
y
of
MA
RA
,
Malay
s
ia
6
School
of
Com
puti
ng,
Coll
ege
of
Arts a
nd
Science
s,
UU
M,
M
ala
y
si
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Oct
8
,
2018
Re
vised Dec
6
,
2018
Accepte
d Dec
18
, 201
8
Thi
s
pape
r
pre
se
nts
the
eva
lu
atio
n
of
AdBlock
technique
implem
ent
a
ti
on
for
ent
erp
r
ise
net
wo
rk
envi
ronm
ent
.
Thi
s
stud
y
has
p
rese
nte
d
th
e
impact
of
web
brows
ing
ac
ti
viti
es
where
it
is
the
m
ost
ac
ti
ve
traffic
w
her
e
is
co
nsum
ed
the
highe
st
inbound
bandwidt
h
usage
in
ent
erp
r
ise
netw
ork
envi
ronm
ent
.
W
e
can
conc
lud
e
that
DN
S
AdBlock
is
the
best
solu
ti
on
for
en
te
rpr
ise
net
work
envi
ronm
ent
in
te
rm
of
blo
c
king
adve
r
ti
se
m
ent
compare
to
extension
adbl
ock
.
Adbloc
k
te
chni
q
u
e
al
so
red
uce
net
work
dat
a
req
uest
b
y
compari
ng
front
-
end
solut
i
on
(brows
er
extension
AdBlock
)
at
c
li
en
t
web
brows
er
and
net
works
le
v
el
adbl
o
ck.
Th
e
par
amete
rs
such
as
HTT
P
req
uest
,
TCP
conne
c
ti
on
and
net
work
ba
ndwidth
are
b
eing
exa
m
ine
d
to
m
ea
sure
th
e
e
ffe
c
ti
ven
ess
of
bloc
king
onl
ine
adve
rt
isement.
Both
te
chn
ique
s
per
form
the
red
uction
of
tra
f
fic
s
and
bandwi
dth
utili
za
t
ion.
The
r
esult
show
s
tha
t
DN
S
AdBlock
is
th
e
m
ost
eff
ec
t
ive
i
n
bloc
kin
g
onl
i
ne
adv
ert
isemen
t
using
th
e
exa
m
ine
d
par
amete
rs.
DN
S
AdBlock
c
an
sus
ta
in
the
usage
of
we
b
brows
ing
ac
t
ivi
t
y
for
en
terprise
ne
twork
and
a
lso
genera
te
subs
ta
nt
ial
sa
ving
a
cro
ss
seve
ral
fon
ts.
Th
is
stud
y
has
ide
n
ti
fie
d
cur
ren
t
we
b
brows
ing
tre
nds
tra
ffic
i
n
ent
erp
r
ise
n
et
wo
rk
where
it
con
sum
ed
50
per
c
e
nt
in
ave
r
age.
T
h
is
num
ber
inc
re
ase
d
when
industri
es
are
m
oving
to
cl
oud
web
-
base
d
consum
pti
on.
How
eve
r,
indust
rie
s
such
as
educ
ational
sec
tor,
web
brows
ing
tr
aff
ic
is
one
of
connect
ivit
y
tha
t
enterpri
se
s
net
work
should
be
inv
esti
ng
to
support
openne
ss
and
he
av
y
tr
aff
i
c
fr
om
educ
a
ti
ona
l
user
s.
Ke
yw
or
ds:
DNS
Ad
Bl
oc
k
E
nter
pr
ise
net
work
W
eb
br
ow
si
ng
traff
ic
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed.
Corres
pond
in
g
Aut
h
or
:
Moh
am
ad
Y
usof
D
a
ru
s
,
Faculty
of Com
pu
te
r
an
d
Ma
them
a
ti
cal
Scie
nces,
Un
i
ver
sit
y Tec
hnology
of MARA,
Sh
a
h Alam
, M
al
ay
sia
.
Em
a
il
:
yuso
f
@
t
m
sk
.u
it
m
.ed
u.m
y
1.
INTROD
U
CTION
In
te
r
net
bro
ws
ing
is
bec
om
i
ng
esse
ntial
pa
rt
of
e
ver
yda
y
li
fe
of
te
n
as
it
of
te
n
us
e
d
to
gat
he
r
inf
or
m
at
ion
.
A
t
so
m
eho
w,
w
eb
co
ntents
de
li
ver
ed
to
th
e
end
us
er
bro
w
ser
with
on
li
ne
web
a
dv
e
rtisem
ents
(ad
s
).
Dail
y
onli
ne
act
ivit
y
of
t
hous
a
nds
of
us
e
rs
i
n
one
netw
ork
en
vi
ronm
ent
giv
e
an
im
pact
of
onli
ne
adv
e
rtise
m
ent
wh
ic
h
ca
n
i
nc
rease
th
e
tra
ff
i
c
as
well
as
i
ncr
ease
ba
ndw
idth
c
onsu
m
ption.
As
World
W
i
de
Web
(
WWW)
m
akes
it
m
o
re
intel
li
gen
t,
the
i
m
plantat
ion
of
onli
ne
web
a
ds
in
w
ebsite
is
on
e
of
the
m
ark
et
ing
st
ra
te
gies
[1
]
.
W
e
b
te
chnolo
gies
are
cur
re
ntly
m
ov
ing
f
r
om
W
e
b
2.0
to
Web
3.0,
onli
ne
ad
s
con
te
nts
a
re
be
com
ing
one
of
the
el
em
ents
an
d
it
play
a
big
r
ole
in
we
b
-
ec
o
-
syst
e
m
[2
]
.
A
new
se
m
antic
form
at
of
e
m
bed
di
ng
onli
ne
ads
w
hich
w
ould
enc
oura
ge
la
rg
e
publis
he
r
s
to
ad
d
them
t
o
their
web
sit
es
[3
]
.
Be
sides,
on
li
ne
we
b
ad
s
bec
a
m
e
the
basic
rev
e
nue
s
ourc
e
for
webpa
ge
publishe
rs,
w
her
e
t
hey
just
si
m
ply
cop
y
the
c
ode
pro
vid
e
d
by
th
e
ad
net
wor
k
pro
vid
e
r
the
n
pa
ste
it
into
their
HTML
file
.
More
ov
e
r,
with
We
b
4.0 a
nd
W
e
b 5.0; o
nline a
ds a
nd e
-
c
omm
erc
e inno
vatio
ns
a
re alrea
dy un
de
r way
[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
Th
e ev
aluati
on
o
f A
dBlo
c
k tec
hn
i
qu
e
im
plem
entatio
n
fo
r e
nt
erprise
...
(
Mo
hd I
sk
andar
Bin Sams
uddi
n)
1103
On
li
ne
a
ds
will
keep
on
gro
w
ing
a
s
sta
te
d
in
[
5]
re
port
s
hows
t
hat
the
53
%
re
venue
is
be
ing
m
ad
e
from
disp
la
y
e
m
bed
ded
f
or
m
at
.
Disp
la
y
e
m
bedde
d
f
or
m
ats
fo
r
on
li
ne
ad
s
wh
ic
h
in
cl
ud
es
banner,
au
dio
,
an
d
vid
e
o
that
dis
pl
ay
ed
at
webpa
ges.
A
stu
dy
ha
d
bee
n
done
wh
e
re
they
e
va
luate
500
URL
s
of
popula
r
w
ebsite
s
and
th
e
res
ults
sh
owe
d
that
63
%
of
them
d
isplay
ed
so
m
e
form
of
ads
[
6
]
.
Fr
om
us
er
pe
rsp
ect
ive
,
the
y
have
bad
bro
wsi
ng
exp
e
rience
bec
ause
of
the
e
xt
ra
c
od
e
that
be
ing
dow
nlo
a
de
d
a
nd
dis
play
ing
ads
aut
om
a
ti
cal
l
y
[
7
]
.
Be
sides,
i
t
do
es
not
prov
i
de
any
m
e
anin
gful
co
ntr
ol
for
us
e
r
[
8
]
.
In
ad
diti
on
,
it
can
no
net
heless
be
intr
us
i
ve
a
nd
re
present
a
sign
i
ficant
dr
ai
n
on
net
work
res
ources
w
h
il
e
w
eb
browsi
ng
[
9
]
.
W
he
n
a
we
bpage
is
loaded
on
web
browser,
it
will
do
wnload
al
l
the
i
nfor
m
at
ion
of
the
web
sit
e
and
al
so
on
li
ne
ads
.
Fu
rt
her
m
or
e,
t
he
proce
ss
of
f
et
ching,
re
nder
ing
a
nd
disp
la
yi
ng
the
w
eb
c
on
te
nts
re
qu
i
re
a
lot
of
traf
fic
bein
g
gen
e
rated
wh
i
ch
inclu
ded
e
xtra
co
de
f
or
ads
[2
]
.
When
num
ber
of
br
oa
dcasti
ng
increased
,
ne
twork
congesti
on
wi
ll
exist
due
to
exc
essive
a
m
ou
nt
of
broad
ca
st
m
essa
ges
occ
up
ie
d
the
ba
ndwidt
h
of
com
m
un
ic
at
ion
c
hannel
th
us
the
m
essage
co
uld
not
be
delive
red
pr
op
e
rly
due
t
o
pac
ket
loss
durin
g
transm
issi
on
[
1
0
].
Tra
ff
ic
c
onge
sti
on
create
s
m
or
e
ne
gativ
e
i
m
pacts
eff
e
ct
ing
our
daily
act
ivit
ie
s,
safe
ty
and
qu
al
it
y
of
li
fe,
thu
s
at
tract
m
any
di
ff
e
ren
t
s
ci
entifi
c
intere
sts
[1
1
].
Assu
m
e
that
the
da
il
y
on
li
ne
act
i
vity
of
thousa
nd
s
of
us
ers
in
on
e
ne
twork
e
nvir
onm
ent;
m
ean
t
hat
the
im
pact
of
onli
ne
a
ds
can
be
sig
nifi
cantl
y
increase t
he
tra
ff
ic
as
w
el
l as i
ncr
ease
b
a
ndw
idth c
onsu
m
ption.
Accor
ding
to
[
1
2
]
,
the
inc
rea
sed
ad
opti
on
of
cl
oud
ser
vic
es
has
great
ly
dr
i
ven
up
the
a
m
ou
nt
of
In
te
r
net
tra
ff
ic
.
A
stu
dy
ha
d
be
en
done
at
Si
m
on
Fr
ase
rs
U
niv
e
rsity
,
as
public
sect
or
suc
h
as
unive
rsiti
es
are
increasin
gly
m
ov
i
ng
to
onli
ne
le
arn
i
ng
platfo
rm
;
they
concern
on
t
he
nu
m
ber
s
of
daily
onli
ne
us
e
rs
m
igh
t
i
m
pact
dr
ai
ning
on
netw
ork
resou
rces
w
he
re
the
pr
ese
nt
of
we
b
inf
orm
at
ion
is
inclu
de
d
with
on
li
ne
ads
con
te
nts
[
9
]
.
I
n
[1
3
]
s
how
t
he
sta
ti
sti
c
us
ag
e
of
daily
band
width
f
or
e
du
c
at
ion
al
sect
or
i
s
am
on
g
t
he
hig
he
st.
W
it
h
the
gro
wing
nu
m
ber
s
of
onli
ne
a
d
s
bloom
ing
f
or
m
ult
iple
pur
poses,
A
dBloc
k
te
ch
nolo
gy
is
bein
g
introd
uced
with
com
e
in
var
i
et
y
of
instal
l
op
ti
ons
[
14
]
.
T
her
e
a
re
seve
r
al
Ad
Bl
oc
k
te
chnolo
gies
tha
t
have
been
dev
el
op
e
d
wh
ic
h
a
re
A
dBloc
k
a
pp
li
c
at
ion
for
sm
artphon
e
,
bro
ws
er
e
xtensi
on
Ad
Bl
ock,
pro
xi
es
ad
-
blo
c
ker an
d D
NS
A
dBloc
ker.
This
st
u
dy
c
on
cern
s
on
we
b
browsi
ng
act
iv
it
ie
s
in
enter
pri
se
netw
ork,
it
will
be
bette
r
t
o
lo
ok
ov
e
r
the
cu
rr
e
nt
we
b
bro
wsing
tra
ff
ic
c
onsu
m
ption
f
or
ente
rpri
se
net
work
en
vir
on
m
ent
[
7
]
.
Acc
ordin
g
t
o
[
14
],
they
ha
d
c
ompari
ng
the
we
bp
a
ge
l
oa
d
ti
m
e
with
a
nd
without
A
dBl
oc
k
im
ple
m
enta
ti
on
;
a
nd
they
able
t
o
detect
a
fe
w
hund
red
m
il
li
se
conds
of
delay
in
we
bpage
load
.
T
he
relat
i
on
s
hi
p
bet
wee
n
nu
m
ber
of
r
equ
e
sts
m
ade
an
d
TC
P
co
nnect
ion
m
ay
le
ad
to
m
or
e
ba
ndwi
dth
co
ns
um
ption
[2
]
.
T
he
i
ncr
e
ase
in
tra
ff
ic
a
nd
data
exch
a
nge ca
n
c
reate t
he
c
om
plica
ti
on
in
c
ongestio
n
co
ntr
ol
once the t
ran
s
m
issi
on
s o
cc
ur
red
betwee
n o
ne
u
se
r
and
m
ulti
ple
us
ers
[1
1
]
.
Th
us,
as
c
onse
qu
e
nc
es
it
im
pa
ct
s
t
o
th
e
inc
rease
of
net
wor
k
tra
f
fic
in
we
b
bro
wsing
if
onli
ne
ads
traff
ic
is
not
bein
g
c
ontroll
ed.
This
stu
dy
al
so
be
the
im
pr
ov
em
ent
of
pr
e
vious
res
earch
cond
ucted
by
[
9
]
w
he
re
the
y
hav
e
e
valu
a
te
d
bro
wse
r
e
xtensi
on
A
d
-
bl
ock
(AdBloc
k
Plus)
an
d
as
king
to
com
par
e the
re
su
lt
acr
os
s a
no
ther
a
d
-
blo
c
k
t
echn
i
qu
e
.
Pr
e
vious
Ad
Bl
ock
te
ch
nique
aim
s
to
bl
ock
ads
ac
r
os
s
t
he
cl
ie
nt
de
vices
wh
e
re
it
a
ppli
ed
as
pl
ug
-
i
n
extensi
on
s
in
web
brow
se
r
s
uch
as
Mozi
ll
a
Fire
fox,
G
oogle
Ch
ro
m
e
et
c.
H
oweve
r,
ha
ving
su
c
h
plug
-
i
ns
w
orki
ng
in
th
e
bac
kgr
ound
re
qu
i
res
e
xtr
a
act
ivit
ie
s
an
d
c
om
pu
ta
ti
onal
powe
r
[2
]
.
Acc
ordin
g
t
o
[
14
],
the
de
fau
lt
set
ti
ng
of
A
dBloc
k
Plu
s;
it
has
the
f
uture
of
al
l
ow
i
ng
ce
rtai
n
“acc
eptable
a
ds”
to
be
dis
play
on
t
he
web.
Be
side
s,
ha
ving
s
uc
h
as
browser
extensi
on
A
dB
lock
m
igh
t
br
i
ng
fe
wer
diff
ic
ulti
es
f
orm
th
e
per
s
pecti
ves
of
netw
ork
a
dm
i
nistrato
r
since
the
gro
wing
num
ber
of
cl
ie
nt
own
dev
ic
es
s
uch
a
s
la
ptop,
ta
blet,
and
sm
artpho
ne
are
co
nn
ect
e
d
wirelessl
y
ac
cess
to
the
net
work
[
9
]
.
A
st
ud
y
c
onduct
ed
by
[
1
5
]
has
e
va
luate
d
m
ul
ti
ple
Ad
Bl
ock
to
ols
ef
fec
ti
ven
ess
in
bl
oc
king
onli
ne
ads
duri
ng
dow
nlo
a
ding
popul
ar
we
bs
it
es
howev
e
r,
their
st
ud
y
sco
ped
down
to
browser
exte
nsi
on
te
ch
nique
an
d
it
did
no
t
dif
fer
e
ntiat
e
the
tra
ff
ic
as
will
be
cond
ucted
i
n
t
his
stu
dy.
B
esi
des,
so
m
e
research
has
done
a
stud
y
on
the
eff
ect
ive
ness
of
avail
a
ble
A
d
-
Bl
ock
;
for
exam
ple,
they
hav
e
co
nduc
te
d
a
te
st
on
m
os
t
po
pula
r
bro
wser
e
xtensi
on
“
Ad
Bl
oc
k
Plus”
an
d
well
-
kn
own
hard
war
e
s
olu
t
ion
“
A
dTr
a
p”
[
14
]
.
Howe
ver,
the
e
ff
ect
ive
ness
res
ult
do
e
s
not
disti
ng
uish
betwee
n
dif
fer
e
nt
ad
-
blo
c
k
te
ch
ni
qu
es
in
te
rm
of
ty
pe
of
traf
f
ic
red
uctio
n.
A
lt
ho
ug
h
m
any
researc
hes
m
a
y
hav
e
an
opi
ni
on
on
this, the
scienti
fic evi
den
ce
to pr
oof
s
uc
h o
pin
io
n har
dly exi
sts
.
This
pa
per
pre
sents
the
Eval
uation
Of
A
dB
lock
Tec
hn
i
que
Im
ple
m
entat
i
on
F
or
E
nter
pri
se
Networ
k
En
vironm
ent.
This
st
ud
y
se
le
ct
ed
two
ty
pes
of
ad
blo
c
k
te
ch
nique
t
o
be
e
valuate
w
hich
a
re
B
rows
e
r
Extensi
on
A
d
-
Bl
ock
a
nd
DNS
A
d
-
Bl
oc
k
w
hich
is
sam
e
m
et
hod
bei
ng
do
ne
by
[
14
]
to
com
par
e
an
d
ev
al
uate
the
eff
ect
i
ve
ne
ss
of
a
dbloc
k.
An
open
-
s
ourc
e
app
li
cat
io
n
w
hich
a
re
A
dBlo
ck
Plus
f
or
Br
owser
E
xten
sio
n
an
d
Pi
-
H
ole
f
or
D
NS
a
re
te
ste
d
an
d
a
naly
sed
.
I
n
order
to
cond
uct
com
par
iso
n
of
A
d
-
blo
c
k
e
ff
ect
iv
enes
s
,
this
stud
y
c
rea
te
d
three
net
w
ork
sce
nar
i
os
of
e
xperim
ent
wh
e
r
e
Sce
na
rio
1:
is
netw
ork
with
out
A
d
-
Bl
ock
,
Scena
rio
2
net
work
with
Br
owser
E
xten
sio
n
A
d
-
Bl
ock
a
nd
Sce
nar
io
3:
netw
ork
D
NS
Ad
-
Bl
oc
k.
Acc
ordin
g
to
[1
5
]
and
[
16
]
the
best
way
to
te
st
the
eff
ect
iveness
of
ad
blo
c
k
is
to
te
st
on
m
os
t
po
pul
ar
we
bs
it
e
s
.
I
n
each
te
st
scenario
;
there
a
re
three
areas
of
in
vest
igati
on
that
thi
s
stud
y
f
oc
us
ing
w
hich
are
“HTTP
Re
qu
e
st
==
GET”,
“TC
P
Connect
io
n”
a
nd
“B
an
dw
i
dth
Co
nsum
ption
”.
Ea
ch
a
rea
of
i
nv
e
sti
gatio
n
is
a
naly
sed
a
nd
res
ult
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.
13
, N
o.
3
,
Ma
rc
h 201
9
:
110
2
–
11
0
9
1104
is
com
par
ed
w
it
h
oth
e
r
sce
na
rio.
In
[
2]
has
create
d
a
se
qu
ence
of
dow
nl
oad
i
ng
we
bpag
e
where
it
s
hows
that
“HTTP
Re
que
st
==
GE
T”
i
s
the
tra
ff
ic
be
ing
us
e
d
t
o
f
et
ch
onli
ne
ad
s.
I
n
[2]
al
s
o
m
entioned
tha
t
TCP
connecti
ons
ar
e
need
e
d
f
or
onli
ne
ad
ver
ti
si
ng
pur
po
se
.
T
CP
is
a
con
ne
c
ti
on
-
or
ie
nted
tr
ans
port
la
ye
r
prot
oco
l
that
bein
g
us
e
d
to
car
ry
ongoin
g
data
ex
ch
ang
e
.
A
s
sta
te
d
by
[
2]
;
the
m
or
e
TCP
connecti
on
t
o
est
ablishe
d,
the
m
or
e
ba
ndwidth
c
onsu
m
ed.
T
her
e
fore,
by
co
m
par
in
g
th
e
resu
lt
,
it
will
po
ssi
ble
to
cal
culat
e
the
redu
ct
ion
of
net
wor
k
tra
ff
ic
an
d
ba
ndwidt
h
c
onsu
m
ption.
As
a
res
ult,
this
stud
y
able
to
pro
po
s
e
a
su
it
able
Ad
Bl
oc
k
te
chn
iq
ue
to
be
i
m
ple
m
ent
i
n
enter
pr
ise
ne
twork
.
Th
us
,
this
stud
y
at
t
e
m
pts
to
inv
e
sti
gate,
analy
se
an
d
evaluate
t
he
Ad
Bl
ock
te
ch
niques
for
e
nt
erprise
netw
ork
e
nv
i
ronm
ent
an
d
docum
ent
ed
t
he
fin
din
gs
in
reducti
on
of
ne
twork
traf
fic a
nd b
a
ndwi
dth
consum
ption
s
with
resp
ect
t
o web b
r
ow
si
ng
act
ivit
ie
s.
2.
RESEA
R
CH MET
HO
D
OL
OGY
In
this
resea
rc
h,
the
t
wo
m
ajor
ty
pes
of
va
riables
are
i
ndepe
ndent
a
nd
dep
e
ndent
va
riables
bee
n
identifie
d.
A
n
ind
e
pende
nt
va
riable
is
a
var
ia
ble
that
aff
ect
the
dep
e
nd
e
nt
var
ia
ble
w
hile
dep
e
nd
e
nt
vari
able
is t
he variable
a resea
rch
e
r
is
interest
ed
i
n. T
he
c
om
po
ne
nts
of
var
ia
bles
ar
e shown i
n Fi
gure
1
.
Figure
1
.
Ex
pe
rim
ental
var
ia
bl
es
In
this
resea
rc
h,
there
a
re
three
ind
e
pe
nd
e
nt
var
ia
bles
that
be
en
m
anipu
la
te
d
w
hich
are
netw
ork
without
A
dB
lo
ck
(
Scena
rio
1),
ne
tw
ork
wit
h
br
ow
se
r
e
xtension
AdB
loc
k
(Sc
ena
rio
2),
an
d
net
wor
k
with
DNS
A
dB
loc
k
(S
cena
rio
3)
.
Fr
om
al
l
inde
pende
nt
com
pone
nt,
there
are
three
dep
e
nd
a
nt
areas
th
at
be
en
inv
est
igate
w
hich
a
re
HTTP “
get” re
quest
, T
CP co
nnect
ion
and b
a
ndwi
dth co
nsum
ption
.
2.1.
Measure
ment F
ramew
ork
The
fr
am
ewor
k
th
at
bee
n
use
d
for
t
his
stu
dy
is
sim
il
ar
with
pr
e
vious
researc
h
c
onduct
ed
by
[1
5
]
wh
e
re
t
hey
use
d
it
to
in
vestigat
e
the
am
ou
nt
of
data
generate
d
by
a
dv
e
rtisem
ents
w
he
n
bro
wsing
as
sho
w
n
in
F
ig
ure
2
.
Figure
2
.
Ex
pe
rim
ental
design
a
ppr
oac
h
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
Th
e ev
aluati
on
o
f A
dBlo
c
k tec
hn
i
qu
e
im
plem
entatio
n
fo
r e
nt
erprise
...
(
Mo
hd I
sk
andar
Bin Sams
uddi
n)
1105
Acti
viti
es
co
nducted
in
t
his
e
xperim
ent
wh
e
n
we
us
e
d
net
w
ork
pac
ket d
at
a
to
i
nv
est
i
gate the
am
ou
nt
of
data
ge
ner
at
ed
by
a
dv
e
rtise
m
ents
wh
e
n
br
ow
si
ng.
T
his
exp
e
rim
ent
need
s
to
set
-
up
a
netw
ork
e
nv
ir
onm
ent
wh
e
re
it
will
be
te
ste
d
with
three
sce
na
rios
as
m
ention
be
f
or
e
.
W
e
b
br
owsing
em
ulati
on
is
condu
ct
e
d
in
eac
h
cl
ie
nt
PC
f
or
each
sce
na
rio.
Data
c
ollec
ti
on
of
the
we
b
browsi
ng
tra
ff
i
c
is
m
easur
ed
us
in
g
pac
ket
s
niff
i
ng
pro
gr
am
“W
i
re
sh
ar
k”.
F
ur
t
her disc
us
sio
n for
each ste
ps
a
re
exp
la
ine
d
i
n ne
xt s
ub
-
to
pi
cs.
2.2
.
Net
w
ork E
nv
ir
on
men
t
Set
-
Up
F
i
gu
r
e
3
s
ho
w
s
t
he
ne
t
w
or
k
e
nv
i
r
o
nm
e
nt
se
t
-
up
f
o
r
t
hi
s
s
t
ud
y
.
T
he
r
e
a
r
e
t
hr
e
e
t
y
pe
s
of
s
c
e
na
r
i
o
s
w
he
r
e
t
he
ne
t
w
or
k
c
on
f
i
gu
r
e
d
w
i
t
ho
ut
A
d
B
l
oc
k
f
or
S
c
e
n
a
r
i
o
1,
S
c
e
na
r
i
o
2
w
i
t
h
br
ow
s
e
r
e
xt
e
ns
i
on
a
nd
w
i
t
h
D
N
S
A
dB
l
oc
k
f
or
S
c
e
na
r
i
o
3.
Figure
3
.
Test
ing
arc
hitec
ture
T
o
s
im
ul
a
t
e
t
he
ne
t
w
or
k
e
n
vi
r
on
m
e
nt
,
a
n
i
s
ol
a
t
e
ne
t
w
or
k
i
t
s
e
t
up
us
i
ng
n
e
t
w
or
k
a
dd
r
e
s
s
t
r
a
ns
l
a
t
i
on
(
N
A
T
)
a
nd
t
hi
s
s
t
ud
y
s
e
t
-
u
p
a
ne
t
w
or
k
w
h
i
c
h
c
l
i
e
nt
P
C
s
c
on
ne
c
t
e
d
t
o
t
he
s
w
i
t
c
h
a
n
d
t
he
a
l
l
P
C
s
w
i
l
l
be
c
on
f
i
g
ur
e
d
a
s
s
t
a
ti
c
I
P
.
V
i
r
t
ua
l
m
ac
hi
ne
i
ns
t
al
l
s
w
i
t
h
D
e
bi
an
o
pe
r
a
t
i
ng
s
y
s
t
em
t
o
r
un
A
ut
oB
r
ow
s
e
a
t
A
p
a
c
he
s
e
r
ve
r
po
r
t
80
8
1.
N
e
t
w
or
k
t
r
a
f
f
i
c
i
s
c
a
pt
ur
e
d
us
i
ng
Wi
r
e
s
ha
r
k
s
of
t
w
a
r
e
.
I
n
S
c
e
na
r
i
o
1,
no
r
m
a
l
D
N
S
s
e
r
ve
r
i
s
im
pl
em
e
nt
e
d
a
s
D
N
S
f
or
w
a
r
de
r
t
o
s
am
e
up
s
t
r
e
am
D
N
S
(
G
o
og
l
e
D
N
S
)
t
o
gi
ve
a
f
a
i
r
a
nd
c
on
s
i
s
t
e
nt
m
e
a
s
ur
em
en
t
of
t
e
s
t
t
r
a
f
f
i
c
i
n
ne
t
w
or
k
w
i
t
ho
ut
A
d
B
l
oc
k
.
In
S
c
e
na
r
i
o
2,
A
dB
l
oc
k
P
l
us
i
s
im
pl
em
e
nt
ed
i
n
c
hr
om
e
w
e
b
b
r
o
w
s
e
r
a
nd
i
t
po
i
nt
t
o
t
he
s
am
e
D
N
S
a
s
i
n
S
c
e
na
r
i
o
1.
W
hi
l
e
i
n
S
c
e
na
r
io
3,
c
l
i
e
nt
D
N
S
is
po
i
nt
ed
t
o
P
i
-
ho
l
e
a
s
i
t
f
ut
ur
e
ha
ve
t
he
a
bi
l
i
ty
t
o
pe
r
f
or
m
D
N
S
f
or
w
a
r
de
r
t
og
e
t
he
r
w
i
t
h
A
ds
B
l
oc
ki
ng
.
2.
3
.
W
e
b
B
r
o
w
s
i
n
g
E
m
u
l
at
i
on
In
t
his
rese
ra
ch
,
we
a
da
pted
the
m
i
m
ic
su
r
fin
g
us
in
g
Au
t
oBrows
e
pro
gr
am
wr
it
te
n
by
[1
7
]
.
This
pro
gram
i
s
in
JavaSc
ript
and
runs
on
t
he
PCs
web
bro
wser
t
hat
retrie
ves
a
gi
ven
set
of
URLs.
T
he
re
are
fifty
(50)
li
sts
of
URLs
that
has
bee
n
ide
nt
ifie
d
that
nee
d
ed
to
be
l
oad
e
d
fro
m
Ma
la
ys
ia
and
ot
her
c
ountry
web
sit
e.
I
n
ord
er
to
set
up
a
con
sist
e
nt
con
c
urren
t
re
qu
e
st
from
c
li
ent
web
bro
w
ser
,
this
reser
ach
is
m
o
difie
d
by
add
i
ng
a
ne
w
f
un
ct
io
n
“ge
tt
i
m
e()
”
in
A
uto
Br
ow
se
a
nd
r
un
it
in
A
pache
serv
e
r.
A
n
e
xp
e
rim
ental
test
tim
e
is
set
at
Au
to
Browse
source
cod
e
by
m
od
ify
ing
the
pa
ra
m
et
ers
sh
ow
n
in
Table
1.
T
he
refor
e
,
al
l
PC
s
are
request
ed
f
or
Au
t
oBrows
e
and
retrieve
it
w
it
h
the
curre
nt
tim
e
in
Ap
ac
he
serv
e
r.
All
the
URLs
are
l
oaded
f
or
each
of the sc
e
nar
i
o.
Table
1.
Gett
im
e ()
Descr
i
ption
Para
m
eters
Descripti
o
n
s
Valu
e1
Ho
u
rs start
Valu
e
2
Minu
tes start
Valu
e
3
Seco
n
d
s start
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.
13
, N
o.
3
,
Ma
rc
h 201
9
:
110
2
–
11
0
9
1106
2.4
.
Tr
affic
Measurem
en
t
In
t
his
rese
rac
h
,
we
decide
d
to
us
e
the
pac
ke
t
sn
iffi
ng
pro
gr
am
W
i
reshar
k
to
ca
pture
al
l
traff
ic
from
each
sce
nar
i
o.
Since
al
l
dat
a
passes
t
hro
ugh
NAT,
W
i
resh
a
r
k
can
si
m
pl
y
m
easur
e
data
that
is
bein
g
exch
a
nged
onl
y
for
the
netw
ork.
I
n
sce
nar
i
o
1
an
d
2,
al
l
cl
ie
nt
PC
is
point
ed
to
l
ocal
DNS
192.1
68.0
.
70
.
Th
us
,
t
hose
tw
o
sce
nar
i
os
ha
ve
the
sam
e
network
flo
w
a
nd
forw
a
r
d
to
s
a
m
e
up
stre
am
DNS
se
rv
e
r
.
W
hile
i
n
scenari
o
3,
al
l
cl
ie
nt
PC
is
po
i
nt
ed
t
o
19
2.168.5.
70
loc
al
DNS
c
onfi
gure
d
with
Pi
-
H
ole
a
nd
t
he
sam
e
up
st
ream
DN
S
as
in
scena
rio
1
a
nd
2.
All
traff
ic
from
each
IP
a
ddress
is
inv
est
igate
d
in
W
i
res
hark
in
to
filt
ering
t
he
tra
ff
ic
. Wi
res
ha
rk
is fil
te
red
the t
raffic
acco
r
ding to
the
r
ules.
A
cco
r
ding
to
[
9
]
,
to
al
l
ow
t
hem
to
ha
ve
un
i
form
base
of
c
om
par
ison
acr
os
s
al
l
the
co
nt
ro
ls
a
nd
al
l
the
s
ub
se
que
nc
e
they
c
ondu
ct
ed
the
te
st
one
a
sin
gle
da
y.
The
refo
re,
t
his
re
searc
h
is
co
nduct
ed
te
s
t
s
for
al
l
scenari
os
w
it
hi
n
twenty
-
fou
r
(
24)
ho
ur
s
.
All
scenari
o
s
ha
ve t
o
fo
ll
ow pr
oce
dure b
el
ow
[1
7
]
in
or
de
r
to g
e
t t
h
e
reli
able res
ult
:
a)
DNS
qu
e
ries
point t
o sam
e u
pst
ream
D
NS
(
Goo
gleD
NS
)
.
b)
All we
b bro
ws
er r
e
quest
to
AutoBr
owse a
nd
start an
d
e
nd c
on
c
urre
ntly
.
c)
Cl
ear w
e
b bro
wser cache
b
e
f
or
e
ev
e
ry test
s
ta
rt.
d)
Each sce
na
rio m
us
t
be
co
nduc
te
d
a
3
ti
m
es t
est
.
2.5
.
Data A
n
aly
sis
Af
te
r
ca
ptu
ri
ng
tra
ff
ic
f
ro
m
al
l
scena
rios,
we
co
nducte
d
a
n
offli
ne
inv
e
sti
gation
on
t
he
tra
ff
ic
gen
e
rates
from
we
b
brow
si
ng
em
ulatio
n.
For
data
a
naly
sis,
this
ex
pe
rim
e
nt
filt
ered
on
t
he
I
P
a
ddress
of
t
he
cl
ie
nt
PC
and
Au
t
obr
ow
se
program
s.
This
r
esearch
m
easu
red
eac
h
sce
na
rio
at
le
ast
thr
ee
(3)
tim
es
t
o
get
aver
a
ge
res
ult.
Com
par
e
the
resu
lt
s
of
net
w
ork
traf
fic
fro
m
al
l
scenarios
and
get
the
nu
m
ber
of
HT
TP
get
request,
TCP c
onnecti
on a
nd
band
width co
nsum
ption
wh
il
e i
m
ple
m
ented
the
Ad
Bl
ock.
3.
RESU
LTS
Af
te
r
c
onduct
ed
the
e
xp
e
ri
m
ent,
we
ha
ve
the
outp
ut
durin
g
data
ga
therin
g
a
nd
t
est
ing
ph
ase
.
The
res
ult
was
ta
bu
la
te
d
in
t
he
form
of
ta
bl
es
an
d
gr
a
ph
and
is
discuss
e
d
la
te
r
i
n
t
his
t
op
ic
.
Fig
ure
4
sh
ows
gr
a
ph
that
de
m
on
strat
e
network
with
ou
t
Ad
Bl
ock
a
f
te
r
25
m
inu
te
s
of
we
b
bro
wsing
act
ivit
ie
s
is
captu
red
wh
ic
h
inclu
ding
we
b
co
ntent
and
on
li
ne
a
dverti
se
m
ent.
A
zi
gzagged
li
ne
s
and
rem
ai
ned
co
ns
ta
nts
li
ne
sho
w
after s
pik
e
o
cc
ur
s
.
Figure
4
.
Grap
h
netw
ork wit
hout a
dbloc
k: sc
enar
i
o
1
Figure
5
s
how
s
a
grap
h
t
hat
dem
on
strat
e
ne
twork
with
bro
wser
exte
ns
io
n
A
dBlock.
T
hi
s
scena
rio
is
dem
on
sta
te
where
at
en
d
dev
i
ce
is
dep
l
oyed
us
in
g
browser
extensi
on
A
dBlock
.
The
grap
h
s
hows
a
c
ha
nge
of
beh
a
vior
w
here t
her
e
are
less
zigzag
ge
d
li
ne
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
Th
e ev
aluati
on
o
f A
dBlo
c
k tec
hn
i
qu
e
im
plem
entatio
n
fo
r e
nt
erprise
...
(
Mo
hd I
sk
andar
Bin Sams
uddi
n)
1107
Figure
5
.
Gr
ap
h
netw
ork wit
h b
rowser e
xten
sion ad
blo
c
k: s
cenari
o
2
Figure
6
s
how
s
a g
ra
ph
t
hat r
e
pr
ese
nt n
et
wor
k
with
DNS AdBl
oc
k
w
her
e t
his scen
a
rio A
dBloc
k
was
dep
l
oyed
at
D
NS
le
vel
(
Pi
-
H
ole).
T
he
gr
a
ph
sho
ws
the
si
m
il
arities
of
s
pik
e
occure
d
a
s
in
scena
rio
1
and
2.
In
co
ntrast,
thi
s
grap
h
dem
on
strat
ed
a
c
ha
ng
e
in
be
ha
vio
r
w
he
re
al
m
os
t
none
of
zi
gz
agg
e
d
li
nes
s
howing
after
a
s
pik
e
oc
cur
e
d.
M
or
e
over
,
com
par
e
t
o
gra
ph
Fig
ure
4
an
d
5
;
this
gr
a
ph
dem
on
strat
es
afte
r
a
sp
ike
occurs a
nd it
r
e
ached a l
ow al
m
os
t zero
(
0) by
te
p
er s
ec
ond.
Figure
6
.
Grap
h
netw
ork
with
DNS
a
dbloc
k:
sc
ena
rio
3
3.1.
Wir
esha
r
k P
ac
ket An
aly
sis
In
this
analy
sis,
we
si
m
ulate
t
wo
e
nd
de
vice
s
wh
e
re
each
of
it
ru
nn
i
ng
Ubun
t
u
Linux
an
d
W
in
dow
s
op
e
rati
ng
syst
em
(O
S)
both
c
on
c
urre
ntly
req
ue
sti
ng
for
th
e
sam
e
URL
us
ing
G
oogle
Chrom
e
web
br
ow
s
e
r
.
Table
2
show
s
te
st
case
resu
lt
fo
r
60
seco
nd
.
This
the
exp
e
rim
ent,
we
fo
und
that
dif
fer
e
nt
OS
ha
ve
dif
fer
e
nt
nu
m
ber
of
HT
TP
Re
quest
=
=
"GET"
a
nd
TCP
Tra
ff
ic
.
Howe
ver,
in
DNS
A
dBloc
k,
HT
TP
Re
quest
==
"GET"
is
the
sa
m
e
in
bo
th
O
S.
It
is
i
m
po
rtant
to
m
ention
her
e
that
we
b
brows
er,
Ad
Bl
ock
a
nd
OS
ar
e
bein
g
set
up
as
by
de
f
ault
wh
e
re
ther
e
is
no
m
od
ific
at
ion
at
end
de
vices
an
d
the
m
et
ho
d
of
data
colle
ct
ion
are
bein
g
captu
red at
n
et
work le
vel.
Table
2.
T
est
Ca
ses and Tim
e Fr
am
e
Test Cases
Ti
m
e
Fr
a
m
e
Test 1
(
Netwo
rk w
ith
o
u
t AdB
lo
ck
)
(f
ra
m
e
.ti
m
e
>=
"
M
ay
19
,
2
0
1
8
23
:0
0
:
0
0
"
&
&
f
ra
m
e.ti
m
e <=
"M
a
y
19
,
2
0
1
8
23
:
0
1
:0
0
")
Test 2 (
Netwo
rk w
ith
Brows
er
Exten
sio
n
AdB
lo
ck
)
(f
ra
m
e
.ti
m
e
>=
"
M
ay
20
,
2
0
1
8
12
:4
5
:
0
0
"
&
&
f
ra
m
e.ti
m
e <=
"M
a
y
20
,
2
0
1
8
12
:
4
6
:0
0
")
Test 2 (
Netwo
rk
w
ith
DNS
Ad
Blo
ck
)
(f
ra
m
e
.ti
m
e
>=
"
M
ay
20
,
2
0
1
8
12
:4
5
:
0
0
"
&
&
f
ra
m
e.ti
m
e <=
"M
a
y
20
,
2
0
1
8
12
:
4
6
:0
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.
13
, N
o.
3
,
Ma
rc
h 201
9
:
110
2
–
11
0
9
1108
As
s
how
n
in
Table
3,
the
i
m
ple
m
entat
ion
of
Ad
Bl
ock
c
on
s
um
e
alm
os
t
half
of
traf
fi
c
in
netw
ork
without
A
dBlo
ck.
Howe
ver,
the
de
plo
ym
ent
of
di
ff
e
ren
t
Ad
Bl
ock
te
c
hniq
ue
ge
ne
rates
a
diff
e
re
nce
resu
lt
(H
T
TP
Re
qu
e
st
==
"GET"
and
TCP
T
raffic
I
nbound)
.
HTTP
Re
ques
t
==
"GET"
f
or
bro
wse
r
ex
te
ns
i
on
Ad
Bl
ock
ge
ne
rates
le
ss
re
qu
est
com
par
e
to
D
NS
A
dBlo
ck
wh
il
e
TCP
inbo
und
c
onnecti
on
f
or
br
ow
se
r
extensi
on AdB
lock ge
ner
at
es
m
or
e connecti
on co
m
par
e t
o DNS
Ad
Bl
oc
k.
Table
3.
Packe
t An
al
ysi
s
Area
of
I
n
v
estig
ati
o
n
W
ith
o
u
t
Ad
Blo
ck
Bro
wser
Exten
sio
n
AdB
lo
ck
DNS AdB
lo
ck
HTT
P
Req
u
est ==
"GE
T"
319
102
152
TCP T
raff
ic I
n
b
o
u
n
d
4
,57
5
2
,68
7
1
,91
8
Ban
d
wid
th
(
MB)
8
4
4
This
ha
pp
e
n
be
cause
of
TC
P
co
nnect
io
n
us
e
d
the
sel
ec
ti
ve
re
peat
protoc
ol
(
SRP)
with
posit
ive
ackno
wled
gm
e
nts
(
ACK
s)
a
nd
ti
m
e
-
ou
t.
Ea
ch
byte
sent
is
num
ber
ed
a
nd
bee
n
ac
knowle
dg
e
d.
A
num
ber
of
byte
s
an
d
AC
Ks
ca
n
be
sen
t
in
the
sam
e
pack
et
w
her
e
ind
ic
at
es
the
s
equ
e
nce
num
ber
of
t
he
ne
xt
byte
exp
ect
e
d
by
th
e
receive
r.
AC
Ks
ca
rr
yi
ng
se
qu
e
nce
num
ber
that
ca
rr
ie
d
a
ckno
wled
gm
ent.
If
the
re
is
pa
cket
lost, r
ecei
ver s
end a
duplica
te
A
CK
s fo
r
a s
ubse
qu
e
nt r
ecei
ved p
ac
ket.
4.
CONCL
US
I
O
N
This
resea
rch
outcom
e
sh
ow
s the
i
m
pact
of
web
b
r
owsin
g
act
ivit
ie
s.
The
act
ive
traff
ic
c
on
s
um
ed
the
highest
in
bound
ba
ndwi
dth
us
a
ge
i
n
e
nterprise
netw
ork
e
nv
i
ronm
ent.
W
it
h
the
nu
m
ber
of
daily
onli
ne
act
ivit
y
of
us
er
in
e
nter
pr
ise
netw
ork,
onli
ne
ad
ver
ti
sin
g
c
on
te
nts
m
igh
t
im
pact
hig
h
dat
a
dem
and
fro
m
web
browsi
ng
act
ivit
y.
As
pr
e
di
ct
ion
by
2018
,
80
%
of
we
b
co
ntents
are
delivere
d
in
m
edia
platform
[5
]
.
This
stu
dy
has
identifie
d
c
urr
ent
web
b
r
owsing
tre
nds
traf
fic
in
enterprise
netw
ork
w
her
e
it
con
su
m
ed
ar
ou
n
d
50
pe
rce
nt
in
aver
a
ge.
T
his
sta
ti
sti
c
is
cl
aime
d
f
ro
m
a
stud
y
c
onduct
e
d
by
Ma
la
ysi
an
Com
m
un
ic
at
i
on
s
a
nd
Mult
i
m
edia
Com
m
issi
on
(M
CM
C)
sta
te
d
t
hat
the
tre
nd
s
for
ye
ars
2017
,
87%
of
t
heir
respo
nd
e
nts
use
d
web
browser t
o retr
ie
ve
in
form
at
io
n
a
nd 67 pe
rce
nt u
sed
for f
orm
al
an
d
in
f
orm
al
f
or
stu
dy.
This
stu
dy
agree
s
that
bo
t
h
A
dBloc
k
te
chn
i
ques
pe
rfor
m
a
reducti
on
of
tr
aff
ic
an
d
ba
nd
width
us
a
ge.
Howe
ver,
the
best
so
l
utio
n
in
e
nter
pr
i
se
netw
ork
f
or
A
dBlock
i
ng
te
ch
nique
is
DN
S
A
dB
lock.
By
i
m
ple
m
entin
g
D
NS
A
dBl
oc
k
in
enter
pri
se
net
work
e
nvir
on
m
ent
can
su
sta
in
the
usa
ge
of
we
b
br
ow
si
ng
act
ivit
y
fo
r
ent
erprise
netw
ork
an
d
al
so
it
has
the
po
te
ntial
to
gen
e
rate
sub
sta
ntial
saving
acro
s
s
seve
ral
fonts.
Be
sides,
t
he
i
m
ple
m
entat
ion
of
browser
e
xtensi
on
A
dBlock
pr
ov
e
d
t
ha
t
the
process
of
disp
la
yi
ng
onli
ne
adv
e
rtise
m
ent
dr
ai
ns
a
sign
if
ic
ant
a
m
ou
nt
of
e
nergy
us
a
ge
.
Ther
e
f
or
e
,
wh
e
n
scal
ed
t
o
la
rg
e
netw
ork,
D
NS
Ad
Bl
ock is an
eff
ect
ive
so
l
ution t
o
c
ontr
ol of e
nd d
e
vices
f
or pro
ce
ss
of bl
ock
in
g o
nline
adv
e
rtise
m
ent.
Howe
ver,
bro
wser
e
xtensi
on
A
dBloc
k
ha
s
the
lim
it
ation
for
ente
rpris
e
networ
k.
As
nu
m
ber
of
dev
ic
e
c
onnect
ed
is
increas
ed
,
each
dev
ic
e
ne
ed
to
instal
le
d
Ad
Bl
oc
k.
A
di
scussion
ha
d
be
en
m
ade
by
Si
m
on
Fr
ase
r
Un
i
versi
ty
wh
ere
by
dep
loyi
ng
br
ow
se
r
exte
ns
ion
A
dBloc
k
pr
ese
nts
so
m
e
diff
ic
ulti
es
from
the
per
s
pecti
ve
of
a
networ
k
adm
inist
rator.
A
st
ud
y
that
has
be
en
co
nducte
d
by
[1
8
]
,
s
how
n
that
the
lim
i
ta
tio
n
in
data
colle
ct
io
n
w
her
e
resea
rc
ers
did
not
kn
ow
the
e
xact
nu
m
ber
of
us
e
rs
with
br
ow
s
er
exte
ns
io
n
e
nab
le
d.
In
this
rese
a
rc
he
r
stud
y
pr
ov
e
d
that
there
is
a
gap
in
TCP
c
onnecti
on
w
he
re
m
assive
web
act
ivit
y
is
ge
ner
at
e
d
since
it
us
ed
Web
Re
quest
AP
I
.
O
n
a
la
r
ge
netw
ork,
m
assive
am
ou
nt
of
ba
ndwidt
h
is
consum
ed
in
netw
ork
with
bro
wse
r
extensi
on
network
c
om
par
e
to
DN
S
Ad
Bl
ock.
Usi
ng
a
DNS
A
dBloc
k
act
s
as
DN
S
reso
l
ve
r
wh
e
n
de
vices
connecte
d
to
t
he
netw
ork
e
nvir
on
m
ent.
Fo
r
on
li
ne
ad
ve
rtise
m
ent,
DN
S
Ad
Bl
ock
c
on
tr
olled
at
netw
ork
le
vel.
Th
us
,
this
s
olv
e
d
t
he
dif
ficult
ie
s
of
a
net
work
a
dm
inist
rator
i
n
m
anag
ing
th
e
us
e
r
netw
ork
act
ivit
i
es.
ACKN
OWLE
DGE
MENTS
The
aut
hors
w
ou
l
d
li
ke
to
than
k
the
Fac
ul
ty
of
Co
m
pu
te
r
an
d
Ma
them
at
ic
al
Scie
nces
(F
S
KM)
Un
i
ver
sit
i Te
knol
og
i M
a
ra
(
UiTM)
for
s
po
ns
ori
ng t
his r
e
s
earch
.
REFERE
NCE
S
[1]
I.
Pawełosz
ek
,
“
W
eb
3.
0
Applic
a
ti
ons
in
Ente
rprise
Strateg
y
,
”
Publisher
of
t
he
Univer
sit
y
o
f
Ec
onom
ic
s
in
Katowic
e
,
vo
l.
2
34,
pp
.
129
–
139
,
2015
.
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
Th
e ev
aluati
on
o
f A
dBlo
c
k tec
hn
i
qu
e
im
plem
entatio
n
fo
r e
nt
erprise
...
(
Mo
hd I
sk
andar
Bin Sams
uddi
n)
1109
[2]
A.
Albasir,
et
al.
,
“
Expe
rimental
Stud
y
of
Ene
rg
y
and
Bandwidt
h
Costs
of
W
eb
A
dver
ti
sem
ent
s
on
S
m
art
phones
,
”
Pr
oce
ed
ings
of
the
6th
Inter
nati
onal
Conf
ere
nce
on
Mo
bil
e
Computing
,
Appl
i
cat
ions
and
Serv
ic
es
,
pp.
90
–
97
,
2014
.
[3]
E.
Thomas,
et
a
l.
,
“
Sem
ant
ic
A
dver
ti
sing
for
W
eb
3.
0
,
”
Lectu
re
Notes
in
Com
pute
r
Sci
ence
(In
cl
uding
Subs
erie
s
Le
c
ture Note
s in
Artificial
In
te
l
l
i
genc
e
and
L
ec
tu
re
Notes
in
B
ioinform
at
ic
s),
615
2
LNCS,
pp
.
96
–
105,
2010
.
[4]
M.
Te
kda
l,
et
a
l.
,
“
Deve
lopments
Of
W
eb
Te
c
hnologi
es
And
The
ir
R
efl
e
ct
ion
s
To
Edu
cation:
A
Com
par
at
iv
e
Stud
y
,
”
Journal O
f
Educ
a
ti
onal
And
Instruct
iona
l
Studies In
The
World
,
vol
/i
ss
ue
:
8(1), 2018.
[5]
D.
Silve
rm
an
,
“
I
AB
Inte
rne
t
adv
ert
ising
rev
enu
e
rep
ort
,
”
2017
f
i
rst
six
m
onths
r
esult
s.
Pri
ce
wa
t
erh
ouse
Cooper
s
LL
P,
2017
.
[6]
H.
Jeloda
r
,
et
a
l.
,
“
Provide
a
solu
ti
on
to
isolate
an
d
ide
nt
if
y
web
o
f
adve
r
ti
sing
Per
sian:
W
eb
anno
ying
,
”
Pro
c
edi
a
Computer
Scien
ce
,
v
ol
.
57
,
pp
.
4
11
-
417
,
2015
.
[7]
I.
Jalb
and
A.
Ol
te
anu
,
“
Custom
iz
ed
Ad
Bloc
k
ing
,
”
RoE
duN
et
Co
nfe
renc
e:
N
et
wo
rking
in
Edu
cat
i
on
and
R
ese
arch
.
[8]
M.
H.
Mughee
s
and
Z.
S.
Z
.
Qia
n,
“
Dete
cting
Anti
Ad
-
bloc
k
ers
in
the
W
il
d
,”
Pr
oce
ed
ings
on
Priv
ac
y
Enhancin
g
Technol
ogi
es
,
v
ol
.
3
,
pp
.
127
–
14
2
,
2017
.
[9]
A.
Parm
ar,
et al
.
,
“
AdBlock
Plus
Eff
icac
y
Stud
y
.
Te
chn
ic
a
l
R
epor
t,
”
Sim
on
Fraser
Univer
sit
y
,
201
5
.
[10]
Darus
M.
Y.,
et
al.
,
“
Optimizing
Congesti
on
Control
for
No
n
Safe
t
y
Mess
a
ges
in
VANE
Ts
U
sing
Ta
guc
h
i
Method
,”
in
Alf
red
R.
,
et
a
l.
,
“
Com
puta
ti
ona
l
Sc
ie
nc
e
and
T
ec
hn
olog
y
,”
ICCST
2
017.
Lec
ture
No
te
s
in
Elec
tri
cal
Engi
ne
ering
,
vol
.
488
,
2018
.
[11]
S.
M.
Hati
m
,
et
al.
,
“
VA
NETs
and
Inte
rne
t
of
T
hings
(
IoT
):
A
Discuss
ion,
”
Indone
s.
J.
Elec
tr.
Eng.
Comput.
Sc
i.
,
vol
/i
ss
ue:
12
(
1
)
,
pp.
218
–
224
,
20
18.
[12]
Ar
y
ak
a
Ne
tworks Inc
orpora
t
ed, “The
St
at
e
of the
W
AN
2017
,
”
Arka
y
a, San
Ma
te
o
,
CA,
US
A,
201
7.
[13]
Fortine
t
Inc
orpo
rat
ed
,
“
Fortinet
Thre
a
t
Repo
rt
Q
2
2017
,
”
Fortin
e
t,
Kif
er
Ro
ad
Su
nn
y
va
le,
CA,
U
SA
,
2017.
[14]
E.
L
.
Pos
t
and
C.
N.
Sekhar
an,
“
Com
par
at
ive
Stud
y
and
Eval
ua
ti
on
of
Onl
in
e
Ad
-
bloc
ker
s
,
”
Proce
ed
ing
2n
d
Inte
rnational
Co
nfe
renc
e
In
formation
S
ci
en
ce an
d
Sec
uri
ty (
ICISS)
,
2015
.
[15]
C.
E
.
W
il
ls
and
D.
C.
Uzunog
lu,
“
W
hat
ad
blo
ck
ers
are
(an
d
are
not)
doing
,”
Proce
ed
ings
-
4th
I
EE
E
Workshop
on
Hot
Topics
in
W
eb
Syst
ems and T
ec
hnologies,
H
otWe
b
2016
,
pp
.
72
–
77
,
2016
.
[16]
G.
Merz
dovnik,
et
al.
,
“
Bloc
k
Me
If
You
Can:
A
La
rge
Sc
al
e
Stud
y
of
Tracke
r
-
B
lo
cki
ng
Tool
s,
”
Pr
oce
ed
ings
of
the
2nd
IEEE European
Symposium
on
Sec
uri
ty and Privacy
(
IEE
E
E
uroS
&
P
)
,
2017
.
[17]
D.
B.
Van
and
A.
Pras,
“
The
c
osts
of
web
adv
er
ti
sem
ent
s
whil
e
m
obil
e
brows
i
ng
,
”
L
ecture
No
te
s
in
Com
pute
r
Scie
nc
e
(Inc
lud
i
ng
Subs
eri
es
Le
ct
ure
Notes
in
Artifi
c
ia
l
Int
el
l
i
genc
e
and
Le
c
tu
re
Notes
in
Bioin
for
m
at
ic
s),
747
9
LNCS,
pp.
412
–
422
,
201
2.
[18]
R.
J.
Sim
ons
an
d
A.
Pras,
“
The
Hidden
Ene
rg
y
Cost
of
W
eb
Advert
ising
,
”
Proce
ed
ings
of
the
Tw
el
ft
h
Tw
ent
e
Stude
nt
Confe
re
nce
on
Informati
on
Technol
og
y
,
2010
.
Evaluation Warning : The document was created with Spire.PDF for Python.