TELKOM
NIKA
, Vol. 11, No. 8, August 2013, pp. 45
0
5
~4
511
e-ISSN: 2087
-278X
4505
Re
cei
v
ed
Jan
uary 25, 201
3
;
Revi
sed
Ma
y 15, 201
3; Accepted Ma
y
25, 2013
An Optimal Semantic Networ
k-Based Approach for Web
Service Composition with QoS
Naji Has
a
n AH*, Gao Shu,
AL-G
abri M
a
lek, Jiang Z
i
-Long
Schoo
l of Com
puter Scie
nce,
W
uhan U
n
iver
sit
y
of T
e
chnol
og
y, W
uhan, 4
300
63 C
h
in
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: hasan
ye
1
985
@gmai
l
.com*
1
, gshu4
18@
16
3
.
com
2
,
ma
l
e
ky
e2
00
4@
g
m
a
i
l
.
co
m
3
, w
u
h
anjz
l
@1
63
.com
4
A
b
st
ra
ct
As the spread
of w
eb service
s
, the comp
osit
ion
of w
eb ser
v
ices bec
o
m
es
a hot topic on
both the
acad
e
m
ia
and
IT
doma
i
ns. T
he co
mp
ositi
on of se
rvic
es
dep
ends
on t
he se
ma
ntic d
e
scripti
ons, w
h
ic
h
descri
be the fu
nction
ality of the serv
ices, pro
v
ide
d
by the o
w
ners of se
rvices and o
n
the descri
p
tion of th
e
consu
m
er
’
s
re
quests. T
he co
mp
ositi
on a
l
so
dep
ends
on n
on-functi
on
al d
e
scripti
ons (Qo
S
). In this pap
er
,
w
e
present a
n
appr
oach
of w
eb serv
ic
es co
mp
ositi
on b
a
se
d on b
o
th se
mantic d
e
scripti
o
n
an
d QoS. T
h
e
prop
osed
a
ppr
oach
b
u
il
ds a
netw
o
rk of w
e
b serv
ices
by
match
m
aki
ng t
he s
e
mantic
c
once
p
ts i
n
OW
L-S,
usin
g the
o
u
tp
uts-inp
u
ts si
mil
a
rity betw
e
e
n
s
e
rvices
. T
h
en,
ma
ny co
mpos
e
d
ass
e
mbl
i
es c
an
be cr
eate
d
by
combi
n
in
g w
eb services in
the semantic
netw
o
rk
acc
o
rdin
g to
co
nsu
m
er
’
s
nee
ds.
T
hose c
o
mp
o
s
ed
asse
mb
lies w
i
l
l
be ra
nke
d
by
the QoS th
at provi
ded
by
th
e cons
u
m
er
a
nd the
suita
b
l
e
one
w
ill b
e
s
ent
back to th
e co
nsu
m
er. T
h
is
appr
oach
take
s the a
d
vanta
ges fro
m
P
e
ll
e
t
e DL R
eas
on
er, dept
h-first a
n
d
forw
ard chai
ni
ng a
l
gor
ith
m
s.
A protot
yp
e an
d a study
case
are pr
ese
n
ted
to ill
ustrate th
e efficie
n
cy of
ou
r
appr
oach.
Ke
y
w
ords
:
se
ma
ntic netw
o
rk
, w
eb service compos
ition,
Qo
S, simil
a
rity, forw
ard chain
i
ng
alg
o
rith
m
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1.
Introduc
tion
Service
-
o
r
ien
t
ed co
mputin
g (SO
C
)
uses
web
serv
ice
s
a
s
the
basi
c
co
nst
r
uct
s
to
develop inte
rope
ra
ble di
stribute
d
ap
plicatio
ns. Web
servi
c
e
s
are a
u
ton
o
mou
s
, platform-
indep
ende
nt and
l
o
o
s
ely couple
d
e
n
tities whi
c
h
ca
n
be d
e
scri
bed,
publi
s
h
ed, di
scovere
d
, an
d
comp
osed [1]
.
Web
Servi
c
e is the
key
concept in
Service
O
r
iente
d
Architectu
re (SOA), whi
c
h
is
a model for u
t
ilizing di
strib
u
ted ca
pabilit
ies that
may be und
er the
control of different o
w
ne
rship
domain
s
[1].
In SOA, se
rvice
co
nsu
m
ers
see
k
t
o
satisfy thei
r spe
c
ific
ne
eds by u
s
ing
the
capabilities offered as a servic
e by a
service provider. One ap
proach to reali
z
e SOA i
s
Web
s
e
rvic
es
[2].
Web
se
rvice i
s
a software
system
whi
c
h
des
ig
ned to
sup
port inte
roperable m
a
chine-to
-
machi
ne (ap
p
licatio
n-to
-a
pplication) int
e
ra
ction
s
ove
r
the intern
et. It is base
d
on eXtensi
b
le
Markup la
ng
uage
(XML) [
4
] that con
s
titutes the mai
n
techn
o
logy
of Web
serv
ice
s
. Ho
weve
r,
Web
se
rvice
s
a
r
e, u
s
uall
y
syntacticall
y des
cri
bed
with sta
nda
rds li
ke
(UDDI, SOAP, and
WSDL
). Universal
De
scrip
t
ion, Discove
r
y and Inte
gration (UDDI) [3] is a
virt
ual regist
ry that
store
s
an
d ex
poses inform
ation a
bout
Web
Servi
c
e
s
. Simple
Obj
e
ct A
c
cess P
r
otocol
(SOA
P) is
a protocol to
excha
nge
structur
ed i
n
formation in
a di
stribute
d
a
n
d
decentralize
d
environme
n
t. It
utilizes XML to demean extensib
le f
r
amework
of messages
wh
i
c
h provides a
constructed
messag
e that
can
be ex
ch
ange
d thro
ug
h a vari
ety of unde
rlying p
r
otocol
s. Th
e
proto
c
ol SOA
P
is ind
epe
nde
nt from a
n
y p
a
rticul
ar
prog
rammin
g
mo
del an
d from
any spe
c
ific
sema
ntics of
an
implementation [5].
Web Se
rvice
De
scription
Lang
uage
(WSDL) [3] p
r
ov
ides
a mod
e
l
and an XM
L
forma
t
for de
scribi
ng
We
b servi
c
e
s
. It se
parate
s
the
de
scri
pt
ion of the
ab
stract fun
c
tion
ality, offered
by
a servi
c
e, fro
m
the
co
ncre
te detai
l
s
of a
se
rvice d
e
scription
such a
s
h
o
w to
get
the servi
c
e
a
nd
where to get
it [4-7]. WSDL files
desc
ribe only t
he
syntactic inte
rf
ace
of We
b service
s
. Hen
c
e,
the pu
re
WS
DL i
s
not
en
ough
to
be
u
s
ed
for aut
o
m
atic We
b service
s
comp
osition:
S
e
m
antic
Annotation i
s
requi
red i
n
o
r
de
r to make
informatio
n a
c
cessibl
e
to autonom
ou
s
agent
s [8]. Qo
S
are al
so an
other type of formats u
s
e to
des
cri
be the
non-fu
nctio
n
of web services.
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 8, August 2013: 4505 –
4511
4506
The main p
u
rpo
s
e of th
is pap
er is
to present a
new ap
pro
a
ch of web
service
comp
ositio
n based on bot
h
Sema
ntic a
nnotation
s
us
ing o
n
tologi
e
s
a
nd
QoS. I
n
ou
r
pro
p
o
s
ed
approa
ch, a
sema
ntic n
e
twork
will be
built acco
rdin
g to the matchma
k
in
g am
ong inp
u
ts a
n
d
outputs of Web se
rvice
s
, whi
c
h supp
osed to
be de
scribed
sem
anti
c
ally via ontol
ogie
s
.
This pa
per i
s
organi
ze
d as follows. Section
2 is de
voted to the
work related
to Web
servi
c
e
com
positio
n. In Section 3,
we pre
s
e
n
t o
u
r p
r
op
osed
approa
ch, its a
r
chite
c
tu
re its
algorith
m
s a
nd strategie
s
. Section 4 pre
s
ent
s
a study case a
s
an exam
pl
e to prove the
efficien
cy of our app
ro
ach. Finally, Section 5 sum
m
ari
z
e
s
our
wo
rk,
findings a
nd
future wo
rks.
2.
Backg
round
And Rela
ted
Works
2.1. Definitio
n
s and Co
nc
epts
(1) Sema
ntic
Network: i
s
a
web
servi
c
e
netwo
rk
whi
c
h is built according to the
si
milarity
measure amo
ng the input
s and outp
u
ts o
f
web
se
rvice
s
usi
ng the semantic a
nno
tation.
(2) Se
mantic Annotation [
7
]: an annota
t
ion assign
s t
o
an entity, whi
c
h is i
n
the text, a
link to its se
m
antic de
scripti
on. A semanti
c
ann
otation refers to a
n
on
tology.
(3) Simila
rity
Measure [8]:
The rea
s
o
n
e
r
defines
four
levels of simil
a
rities b
e
twe
en two
con
c
e
p
ts A and B, and Similarity Measure can be o
n
e
of the following results:
A. Equivalen
c
e (Match): i
n
this me
asu
r
e the
con
c
e
p
ts A and B
are e
quivale
n
t, similar
c
o
nc
ep
ts
.
B. Subsum
ption (Su
b
sume
s): it me
an
s
that
the con
c
ept A is m
o
re gen
eral th
an the
c
o
nc
ep
t B;
C. Oppo
site
Subsu
m
ption
(Plug-in
):it mean
s
A c
o
n
c
ep
t is
s
ubs
u
m
ed
to
th
e
c
o
nc
ep
t
B,i,e.,
the con
c
ept B is mo
re gene
ral tha
n
A.
D. Differen
c
e
(Fail): Th
e co
nce
p
ts A and
B are differen
t
totally.
(4)
NFSL [9]: Non
-
fun
c
tion
al Specifi
c
ati
on Lan
gua
ge
which de
scri
be the Qo
s o
f
a web
serv
i
c
e.
OWL
-
S: an o
n
tology lang
u
age u
s
ed to d
e
scrib
e
we
b service
s
sema
ntically.
(5)
Re
asone
r [8, 9]: A mappin
g
en
gin
e
is
a reaso
ner li
ke th
e
rea
s
on
er i
n
Artificial
Intelligen
ce; it matche
s se
rvice adverti
se
ments wi
th
re
que
sts. The reasone
r provi
des a
sema
ntic
algorith
m
to match in
puts and output
s of Web
services d
u
rin
g
them at the matchm
aki
ng
p
r
oc
es
s
.
2.2. Relate
d Work
s
Web
Serv
i
c
e
Comp
ositio
n
(WS
C)
[10-
1
1
] is
re
qui
red
wh
en
a
clie
n
t
’s requ
est
ca
n not
be
fulfilled by a
singl
e pre-existing
web
se
rvices. In
su
ch a case, inte
grating
existi
ng web
servi
c
e
s
into one
co
mposite
servi
c
e m
a
y satisfy the req
u
e
s
t. Du
e to th
e impo
rtan
ce
of we
b
se
rvice
comp
ositio
n, there have b
e
en variou
s ap
proa
ch
es
to a
ddre
s
s the web se
rvice
co
mpositio
n.
In [12], Hassina Na
ce
r Ta
lantikite et al
, in
trodu
ce a
new a
pproa
ch of di
scovery and
comp
ositio
n based on se
mantic an
not
ation. An ont
ology rea
s
o
n
e
r ha
s bee
n use
d
in simil
a
rity
measure bet
wee
n
service
s
, and th
e compo
s
ition
re
sult
s are co
mpared a
c
co
rding to
simil
a
rity
time and me
mory sp
ace.
Colu
cci et al. in [13] introduce a form
ali
z
at
ion of mat
c
hin
g
, and propo
se prope
rties tha
t
should hold
i
n
a semanti
c
-based skill
m
a
tching
approach. They also
de
fi
n
e
a
n
al
gorithm
to
ra
nk
matches between skills
pro
fi
le
d
e
scri
p
tions and p
r
esent
an o
n
tology-b
ase
d
system which
embed
s a mo
di
fi
e
d
Ne
oCla
ssi
c re
asone
r implementin
g the ran
k
ing
algorith
m
s.
The QoS-aware WS
C problem is p
r
o
v
en to
be NP-hard [14], so wh
en the
problem
spa
c
e i
s
la
rg
e, it is usually
time con
s
u
m
ing to
de
rive
a QoS-optim
al se
rvice
co
mposite
sol
u
tion.
In [15] a Global method i
s
propo
se
d to solve
the Qo
S-aware
WS
C pro
b
lem, which mo
del
s the
QoS-a
w
a
r
e
WSC p
r
obl
e
m
as a MIP proble
m
and d
e
crea
se
s the time co
st to some extent.
Most of th
e
com
p
o
s
ition
wo
rks a
bov
e are
ma
in
ly to
c
o
mp
os
e s
e
r
v
ic
es
usin
g
th
e
sema
ntic or
QoS d
e
scri
ption a
nd
som
e
wo
rks u
s
ing
both d
e
scri
ptions. In
ou
r
work,
we p
r
e
s
e
n
t
an
o
p
timal web servi
c
e
s
comopo
sition
based on
se
mantic and
QoS
d
e
scri
ptions, whi
c
h
d
eals
with the i
nput
s an
d o
u
tput
s of
servi
c
e
s
and fo
rm
s a
netwo
rk, whi
c
h
can
give
many optio
ns to
obtain a suita
b
le se
rvice
co
mpositio
n, as
we c
a
n s
e
e in the following s
e
c
t
ions
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
An Optim
a
l Sem
antic Network-Ba
sed Ap
pr
oa
ch fo
r Web Service…
(Naji
Ha
san
AH)
4507
3.
An Optimal
Semantic Ne
t
w
o
r
k-b
ased
Appr
oa
ch for Web Serv
ice composi
t
ion
w
i
th QOS
In this se
ctio
n, we intro
d
u
c
e ou
r p
r
op
o
s
ed a
p
p
r
oa
ch
. We begi
n with the archite
c
ture
of
our
prototype
,
and the
n
we present the
matchm
aki
n
g and
discov
ery of web
service
alon
g
with
the idea of bu
ilding a sema
ntic network. Finally t
he co
mbining
se
rvice
s
acco
rdi
n
g to con
s
um
e
r
’s
requ
est
s
and
the ran
k
ing
strategy usi
ng
QoS are p
r
e
s
ented.
3.1. The Arc
h
itec
ture o
f
the Proto
t
y
p
e
Our p
r
ototyp
e con
s
ist
s
of five compo
nent
s, nam
el
y, Agent, DB, Semantic Network,
Composer and
Evaluator.
The architecture of ou
r prototype i
s
illustrated i
n
Figure 1,
The
Followi
ng is a
brief descript
i
on of our p
r
o
t
otype’s com
pone
nts:
Figure 1. The
Archite
c
ture
of the Prototype
(1) A
gent:
consume
r
(or req
u
e
s
ter)
provid
es
r
e
qu
e
s
ts
to
th
e Ag
e
n
t. T
h
e r
e
qu
es
t
inclu
d
e
s
input
s and
output
along
with Q
o
S para
m
eter
s. Later, the
Agent provid
es the
Comp
ose
r
and Evaluato
r
with tho
s
e d
e
scriptio
ns.
(2) Juddi
o
r
Datab
a
se: In
Data
ba
se
provider
s store
theirs’ se
rvi
c
e de
scriptio
ns usi
ng
OWL
-
S for se
mantic de
scri
ptions a
nd NFSL
for QoS (non
-fun
ction
a
l descri
p
tion
s).
(3) Sema
ntic
netwo
rk is th
e net
work
wh
ich
gen
erate
d
by mat
c
hing
inputs an
d o
u
t
puts of
servi
c
e
s
sto
r
ed in Datab
a
se u
s
ing
si
milarity
mea
s
ure (O
WL
DL re
asone
r), that includes a
matc
hing algorithm.
(
4
) Co
mp
os
er
is
c
o
mp
le
tely r
e
s
p
o
n
s
i
b
l
e
fo
r co
mp
osin
g
W
e
b se
rvic
e
s
.
T
h
e co
mp
os
er
fulfills the
re
spon
sibility of
Web
service
comp
os
itio
n
usin
g the
ge
n
e
rated
sema
ntic n
e
two
r
k
and
the req
u
e
s
ter’
s inp
u
t and o
u
tput. The co
mposer
ta
ke
s the advanta
g
e
s fro
m
both
Depth
-
First a
nd
Forward Ch
ai
ning alg
o
rith
ms, and then
gene
rate
s co
mposite
se
rvice
s
.
(
5
) Eva
l
u
a
t
or
r
a
nk
s
c
o
mpos
ite
s
e
r
v
ic
es
a
c
c
o
rd
in
g to
r
e
q
u
e
s
ter
’
s
Qo
S a
n
d
pr
es
en
ts
the
suitabl
e com
posite
servi
c
e
to the reque
ster.
3.2. Matchm
aking and Building a Semantic Ne
t
w
ork
In our ap
proach, Servi
c
es’
provid
ers a
r
e
su
pp
ose
d
to
de
scribe
thei
r se
rvice
s
sema
ntically
and m
ap th
ei
r inp
u
ts
and
output a
s
su
b
c
la
sses of
cla
s
ses (con
ce
p
t
s) i
n
a
comm
on
ontology calle
d Cu
rre
ncy. T
he process of
matc
hma
k
in
g and buil
d
in
g a Semantic netwo
rk, whi
c
h
build
s before any submitte
d requ
est
s
, and goe
s a
s
follows:
(1) Obtain all
servi
c
es’ inputs.
(2) Obtain all
servi
c
es’ outputs.
(3)
Mea
s
u
r
e t
he si
milarity
betwe
en the
con
c
e
p
ts
(out
put) of
a spe
c
ific se
rvice A
and the
con
c
ep
t
(input
) of other se
rvice
s
u
s
ing “Pell
e
t DL” Re
asone
r whi
c
h incl
ude
s the followin
g
matchma
k
i
ng
strategy:
(a) If the output of a servi
c
e A is “eq
u
i
v
alent
” to the input of se
rv
ice B, then create an
edge b
e
twe
e
n
the two se
rvices a
nd lab
e
l the edge b
y
similarity value ”1
”.
(b) If the out
put of a
se
rvice A i
s
“sub
sumed
”
by the
input of
se
rvice B, the
n
create an
edge bet
wee
n
the two se
rvices a
nd lab
e
l the edge b
y
similarity value ”2
”.
If the output of a service A is “ind
epe
nde
nt”, no
relatio
n
with, the input of servi
c
e
B, then
the simila
rity value is”0”, n
o
edge
cre
a
te
s between th
e two se
rvice
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 8, August 2013: 4505 –
4511
4508
3.3. Compos
ition Algorithm
The process
of building a
comp
osite
se
rvice fro
m
Semantic net
wo
rk go
es a
s
fol
l
ows:
(1) G
e
t req
u
e
s
ter’
s input (a
s Sour
ce
) an
d output (a
s
De
stination
)
.
(2) In semanti
c
network, we
look for all serv
ice
s
who
s
e inputs a
r
e similar with the
Source.
(3)
Cre
a
te a compo
s
ition fo
r each found
servi
c
e.
(4) F
o
r ea
ch
new
sou
r
ce which i
s
the inp
u
ts of se
rvice
s
:
Search a
gain
in the
sema
ntic net
wo
rk
all
se
rvice
s
who
s
e output
s
a
r
e simila
r with
the
new
sou
r
ce whose output a
r
e not already
treated.
(5) If all the
i
nputs are tre
a
ted o
r
n
o
m
o
re
so
urce
s,
then the
loo
p
will te
rmin
ate, else g
o
b
a
c
k to
step (4
).
(6) At the en
d, compo
s
e
r
gene
rate
s a list of comp
osi
t
e servi
c
es, e
a
ch
comp
osit
e servi
c
e sta
r
ts
from requ
est
e
r’s inp
u
t (so
u
rce)
and
en
ds
with
r
e
q
ues
te
r
’
s
ou
tp
u
t
(
d
es
tin
a
t
io
n)
. O
u
r s
t
ud
y ca
s
e
will describe
compositio
n algorithm clearly.
3.4. Ev
aluation
Our Evaluato
r
use
s
the foll
owin
g strate
g
y
to evaluate and ra
nk
com
posite
servi
c
e
s
.
(1)
Co
st: can
be cal
c
ul
ate
d
by getting
the sum
of
the value of
co
sts of serv
ice
s
that form a
comp
osite se
rvice.
(2) Som
e
QoS such as availabilit
y, can be calculat
ed by
getting
the m
u
ltiply of the avail
a
bility
value
of se
rv
ice
s
that co
n
s
ist a com
p
o
s
ite se
rvice. For example,
a
c
o
mp
os
ite s
e
r
v
ic
e
w
h
ich
inclu
d
e
s
thre
e web
servi
c
e
s
and thei
r availability values a
s
follows:
(a) WS1 (av
a
ilability)
=0.94,
WS2 (av
a
ilabilit
y) =0.80, and WS
3 (avail
ability) =0.79.
Then, the ava
ilability value of this co
m
p
o
s
ite se
rvice i
s
0.94*0.8*0.7
9
=0.5
94
(b) So
me Q
o
S such a
s
se
curity can b
e
cal
c
ulate
d
b
y
choo
sin
g
the lo
we
st value. For
example, a compo
s
ite se
rvice whi
c
h in
clud
es th
ree
web
servi
c
e
s
and their se
curity value
s
as
follows: WS1
(se
c
u
r
ity) =
high, WS2 (secu
r
ity)
= me
dium and
WS3 (se
c
u
r
ity) = medium.
The
se
curity value
of this comp
osite service is mediu
m
.
(3) Ea
ch
co
mposite
se
rvi
c
e
will be
compa
r
ed to
requ
este
r’
s
QoS and
got
ten a ra
nki
n
g as
follows
:
(a) T
he cost:
the co
st of co
mposite
se
rvice
shoul
d be
equal o
r
sm
al
ler than the
cost that
is provid
ed b
y
the consum
er.
(b) Other
QoS: such
as A
v
ail
ability, security
and response
time
of composite service
sho
u
ld be e
q
ual or la
rge
r
than it the co
st that is provided by the co
nsum
er.
Finally, the suitable co
mp
osite se
rvice will be
se
nt back to the reque
ster a
s
the optimal
comp
osite se
rvice.
4.
Stud
y
Case
In this sectio
n
,
we illustrate
and prove the e
fficiency of our ap
pro
a
ch
using a stud
y case.
The foll
owi
n
g
ca
se
study ill
ustrate
s
the t
e
ch
niqu
es an
d st
rategi
es d
i
scusse
d fo
r
usin
g Sem
a
n
t
ic
Network an
d QoS on servi
c
e comp
ositi
on. This
st
ud
y uses
cu
rre
n
cy exch
ang
e se
rvice
s
a
s
its
domain. T
he
requ
este
r i
s
Chin
ese tou
r
i
s
t who
w
ant
s to visit a
co
untry (Yem
e
n
), he
nee
ds to
excha
nge
Ch
inese curren
cy (RMB) to
Yemeni
cu
rre
n
cy (Yem
eni
Riyal, YR in
sho
r
t) a
nd
he
add
s som
e
Q
o
S as his
req
u
irem
ent.
Table 1. Web
Service
s
wit
h
their De
scri
ptions.
No
Ser
v
ice
Nam
e
Inpu
t
Ou
tpu
t
Cost
Securit
y
A
v
a
i
labili
t
y
Respo
n
se
time
1 RMB2
Y
R
RMB
Y
R
20
HIGH
0.90
2
2 RMB2EUR
O
RMB
EURO
10
HIGH
0.77
1
3 RMB2USD
RMB
USD
5
MEDIUM
0.85
2
4 EURO2
Y
R
EURO
Y
R
5
HIGH
0.76
1
5 EURO2SA
EURO
SA
2
MIN
0.89
3
6 SA2
Y
R
SA
Y
R
1
MIN
0.78
0.5
7 USD2
Y
R
USD
Y
R
1
HIGH
0.97
1
8 USD2SA
USD
SA
1
HIGH
0.69
1.5
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
An Optim
a
l Sem
antic Network-Ba
sed Ap
pr
oa
ch fo
r Web Service…
(Naji
Ha
san
AH)
4509
There are lots of cu
rren
cy
excha
nge
services, an
d ea
ch servi
c
e sh
ould be p
r
ovi
ded with
its Sema
ntic
descri
p
tion
s (OWL
-S)
and
non-fu
nctio
n
a
l
description
s (QoS
). Ou
r a
ppro
a
ch trie
s
to
help in findi
n
g
the suita
b
le
servi
c
e (or
compo
s
ite
service) a
nd off
e
r it to the re
que
ster. Ta
bl
e 1
sho
w
s som
e
web
servi
c
e
s
with their de
scriptio
ns.
And Figure 2 sho
w
s the de
scription of Q
o
S in an NFS
L
file.
Figure 2. An Example of QoS in NFSL
In an
NFSL fi
le, WSName
sho
w
s the
na
me of
web se
rvice, Co
stBy
US
D
sh
ow
s t
he c
o
st
of usin
g t
h
is
se
rvice,
Secu
rity sho
w
s
the
level of
se
cu
rity of the service,
AvailabilityByPercentag
e
sho
w
s th
e avail
abilit
y of the servi
c
e
by perce
ntage
, and
TimeBySeco
nds d
e
scri
be
s the re
quired
time for executing the se
rvice.
4.1. The Matchmaking an
d Semantic
Net
w
o
r
k
In matchm
aki
ng process,
we fo
cu
s on t
he “e
quivale
n
t
” and
“ind
ep
ende
nt” bet
ween the
outputs
and
i
nputs
of serv
ice
s
an
d ign
o
re
other
rel
a
tions. T
able
2 sho
w
s
so
me web
se
rvice
s
and the
relati
on bet
wee
n
them. In this t
able, t
he rows present the
inputs
of se
rvices, a
nd th
e
colum
n
s
pre
s
ent the outpu
ts of se
rvice
s
, each
ce
ll re
pre
s
ent
s the
relation
(si
m
il
arity) betwee
n
the input
s an
d output
s of the servic
es.
For exam
ple,
the output of
“RMB2E
URO
”
, in col
u
mn 2
,
is
equivalent to
the input of servi
c
e
s
“E
URO2Y
R
” a
n
d
“EURO2SA” in rows
4,5,
then they have
gotten “1”,
while the
r
e i
s
no relation
b
e
twee
n “RM
B
2EURO”
an
d “USD2SA
”
, then they h
a
ve
gotten “0
”.
Table 2. Web
Service
s
and
their Relatio
n
s after Mat
c
hmaki
ng
RMB2
YR
RMB2EUR
O
RMB2USD
EURO2
Y
R
EURO2SA
SA2
Y
R
USD2SA
RMB2
Y
R
0
0
0
0
0
0
0
RMB2EUR
O
0
0
0
1
1
0
0
RMB2USD
0
0
0
0
0
0
1
EURO2
Y
R
0
0
0
0
0
0
0
EURO2SA
0
0
0
0
0
1
0
SA2
Y
R
0
0
0
0
0
0
0
USD2
Y
R
0
0
0
0
0
0
0
USD2SA
0
0
0
0
0
1
0
Figure 3 sho
w
s th
e we
b service
s
a
nd t
heir
rela
tion
s in Semantic
Network. If there i
s
a
relation
between two
servi
c
e
s
, the o
u
tp
ut of se
rvice
A is eq
uivale
nt to the inp
u
t
of servi
c
e B
,
an
edge
will be created. Oth
e
rwise no ed
ge
bet
wee
n
two
services
will be created.
Figure 3. We
b Service
s
an
d their Rel
a
tions in Sem
a
n
t
ic Network
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 8, August 2013: 4505 –
4511
4510
4.2. Web Ser
v
ice Composition
The
com
p
o
s
e
r
combi
n
e
s
th
e web
se
rvice
s
in
Sema
ntic network
accordin
g to
re
q
ueste
r’s
input an
d out
put whi
c
h
are
sub
m
itted by
the age
nt.
It sets
up the
re
que
ster’
s
inp
u
t as the
source
point and the requester’
s output
as the destination point. Figure
4 illustrates the
result
of
combi
n
ing
Web se
rvices.
As we h
a
ve
see
n
, we
can
fulfill reque
ster’s
req
u
e
s
t usin
g any of the
following com
posite
servi
c
es:
(1) T
he atomi
c
se
rvice
No:
1:“RMB2Y
R”.
(2) T
he comp
osite service whi
c
h go
es th
roug
h: No: 3”
RMB2
USD”
No: 7
”
US
D2YR
”.
(3)
The
com
p
osite
servi
c
e
whi
c
h g
o
e
s
th
roug
h: No: 3
”
RMB2
USD”
N
o
: 8”
US
D2SA”
No: 6 “SA2Y
R
”.
(4) T
he
co
mposite
se
rv
ice
whi
c
h g
oes th
rou
gh:
No: 2”
RMB2EURO”
No: 4”
EURO
2YR
”
.
(5) T
he
co
mposite
se
rv
ice
whi
c
h g
oes th
rou
gh:
No: 2”
RMB2EURO”
No: 4”
EURO2SA”
No: 6 “SA
2
YR”.
Figure 4. The
Result of Co
mbining
Web
servi
c
e
s
4.3. Ranking
and Ev
aluating
After generating the com
posite
servi
c
es,
the Evaluator ran
ks
and evaluate
s
them
according to
the QoS which
submitt
ed by t
he Agent. The compo
s
ite services, which
are
gene
rated by
the comp
oser, with t
heir Q
o
S are liste
d in Figure 5.
Figure 5. The
result
s of the
study ca
se
At the end th
e Evaluator send
s the suit
able co
mp
o
s
i
t
e service to the req
u
e
s
ter,
which i
s
the comp
osit
e servi
c
e:”
RMB2USD-
>USD2YR “i
n ou
r study case.
5.
Conclusion
In this
pap
er,
we
pro
p
o
s
ed
an o
p
timal a
p
p
roa
c
h
for Web Se
rvice
Compo
s
ition
b
a
se
d o
n
sema
ntic d
e
scriptio
ns
and
on non
-fun
ct
ional de
sc
ri
ptions
(QoS
). In our
app
roa
c
h, a
sema
ntic
netwo
rk ha
s
been
esta
blished a
c
co
rdin
g to the
simi
l
a
rity mea
s
u
r
e of the in
put
s an
d out
put
s of
web servi
c
e
s
.
Lots
of com
posite se
rvices can be
b
u
ilt,
and
final
ly they were
ran
k
ed
accordin
g to
con
s
um
er’
s
QoS. A study case
sho
w
s that this
approach can
co
mpose se
rv
ices efficiently and
can p
r
e
s
ent the suita
b
le compo
s
ite se
rvice to the co
nsum
er a
c
cording to his re
quire
ment
s
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
An Optim
a
l Sem
antic Network-Ba
sed Ap
pr
oa
ch fo
r Web Service…
(Naji
Ha
san
AH)
4511
There will be
following works in future.
1) Movin
g
the
idea of
se
rvice di
scovery
and
servi
c
e
compo
s
ition u
s
ing
ontolo
g
y to clou
d
comp
uting, speci
a
lly the re
so
ur
ce
s’ clo
ud se
rv
ice
s
.
2) A furthe
r
developm
ent
of the disco
v
ery
and
co
mpositio
n proce
s
s with b
r
okers in
multi-cl
oud e
n
vironm
ent.
3) More study
cases a
nd te
sts shoul
d be
carried o
u
t.
Ackn
o
w
l
e
dg
ments
This work i
s
su
ppo
rted
b
y
Wuh
a
n
Internatio
nal
Coope
ration
P
r
oje
c
ts “Stud
y
ing an
d
Implementati
on on Se
m
antic-ba
s
ed
Visuali
z
ation
Service
s
P
ublishing a
n
d
Match
m
a
k
ing
Platform” (P
roject No. 200
9706
3426
9).
Referen
ces
[1]
AB Hass
ine,
S Matsub
ara,
T
Ishida
. A
Constra
i
nt-Bas
ed A
ppr
oach
to Hor
i
z
o
nt
al
W
eb Serv
ice
Co
mp
ositio
n
. Proc. F
i
fth Int'
l
Semantic W
eb
Conf. (ISW
C).
200
6; 130-
143.
[2]
Alons
o G, Cas
a
tiF
,
Kuno
H,
Machira
j
u V.
W
eb Se
rvic
es-
Conc
epts, Arc
h
itec
tures
an
d
Appl
icatio
ns.
Berlin V
e
rla
g
:
Sprin
ger
. 20
04
.
[3]
Chia-F
e
n
g
Li
n, Rue
y
-K
ai S
h
e
u
, Yue-S
han
Cha
ng, Sh
ya
n
-Ming Y
u
a
n
. A rela
xabl
e s
e
rvice se
lecti
o
n
alg
o
rithm for
QoS-base
d
w
eb servic
e co
mpositi
on.
Info
rmati
on a
nd S
o
ftw
are T
e
chn
o
lo
gy
. 201
1;
53(1
2
):13
70-1
381.
[4]
Hassin
a
N
a
cer
T
a
lantikite, Dj
amil A
i
ssan
i
,
Nacer
B
o
u
d
jl
id
a. Sema
ntic a
nnotati
ons f
o
r
W
eb servic
e
s
discov
e
r
y
and compos
ition.
C
o
mputer Stan
d
a
rds & Interfaces Journ
a
l
. 20
09; 31(6): 1
108
-111
7.
[5]
Jing L
i
u,
Xin
g
Guo Luo, Ba
i
Nan L
i
, Xi
ng
Ming Z
h
a
ng, F
an Z
h
a
n
g
.
An
Intellig
ent Jo
b Sche
dul
in
g
S
y
stem for W
eb Servic
e i
n
Clou
d
Com
p
uting.
T
E
LKO
M
NIKA
Indo
n
e
sia
n
Jour
nal
of Electrical
Engi
neer
in
g
. 2013; 1
1
(
6
): 389
-399.
[6]
Lia
n
Yong Qi,
W
an Chun D
ou, Xu
yun Z
h
ang, Jin Ju
n Che
n
. A QoS-a
w
ar
e Comp
o
s
ition meth
od
supp
orting acr
o
ss-pl
at form service inv
o
cati
on in clo
ud e
n
v
i
ronme
n
t.
Jour
nal of Co
mp
ute
r
and Syste
m
Scienc
es
. 201
1; 78(5): 13
16-
132
9.
[7
]
MP Pa
pa
zo
gl
ou
, P T
r
averso, S Dustdar,
F
Le
yman
n.
Servic
e-Orie
nt
ed C
o
mp
uting:
a R
e
searc
h
Roa
d
map.
Inte
rnatio
nal Jo
urn
a
l of Coo
per
ati
v
e Information
Systems
. 20
08
; 17(2): 223-2
5
5
.
[8]
M Gudgi
n,
M Had
l
e
y
, N
Mend
elso
hn,
J Mo
rea
u
. http://
w
w
w
.
w
3
.
o
rg/T
R/soap12
-part1/.
W
3
C
Recomme
nd
ati
on. 201
2; SO
AP Version 1.2
Part 1.
[9]
M Pap
a
zog
l
o
u
.
W
eb Servic
e
s
an
d SOA: P
r
incip
l
es an
d T
e
chnolog
y. C
ana
da: Pe
arso
n Ed
ucatio
n.
201
2.
[10]
Rajes
h
Karu
n
a
murth
y
, F
e
rh
at Khend
ek, Roch,
H Glith
o
. A novel ar
chitecture for W
eb servic
e
compos
ition.
J
ourn
a
l of Netw
ork and C
o
mp
uter Appl
icatio
ns
. 2012; 3
5
(2)
:
787-80
2.
[11]
Ri
yan
a
rto Sarn
o, Herma
w
a
n. Devel
o
p
i
ng
D
i
s
t
ributed
S
y
ste
m
w
i
th
Serv
ice Reso
urce
[12] Oriented
Arch
itecture.
T
E
LKOMNIKA
Indo
ne
sian Jo
urn
a
l of
Electrical E
ngi
neer
ing
. 2
0
1
2
; 10(2): 38
9-
399.
[13]
Schahr
am Dus
t
dar, W
o
lfgan
g
Schrein
e
r.
A surve
y
on
w
e
b
services com
p
ositio
n.
Interna
t
iona
l Jour
na
l
of W
eb and Gri
d
Services
. 20
05; 1(1): 1-30.
[14]
Sérgi
o
AB Cr
uz, Anton
i
o
MV Monteir
o
, Rafa
e
l
Sa
nt
os. Automate
d ge
osp
a
tial
W
eb Servic
es
compos
ition b
a
s
ed on g
e
o
dat
a qua
lit
y
r
e
q
u
ir
ements.
Co
mp
uters & Geosci
ences
. 20
12; 4
7
: 60-74.
[15]
T
homas H Corman, Char
les
E Leisers
on. In
tr
oductio
n
to Al
gorithms. MIT
Press, US. 200
9.
[16] W
S
DL1.1Sp
e
ci
cation.W
eb
Se
rvices
Descri
p
ti
on La
ng
uag
e (W
SDL) Versio
n1.1.
[17] W3CNote.
http://
w
w
w
.
w
3
.or
g
/
T
R/
w
s
d
l
, 201
2.
Evaluation Warning : The document was created with Spire.PDF for Python.