In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
DOI : 10.5121/ijwsc.2016.7102 23
S
ERVICE
S
ELECTION
B
ASED
O
N
D
OMINANT
R
OLE
O
F
T
HE
C
HOREOGRAPHY
D . Ra i Shanka Pandey, Richa Pa hak
Bi la Ins i u e O Technology, Ranchi, India
ABSTRACT
Web se ices a e playing dominan ole on In e ne o e-business. The composi ions o hese se ices a e
used o mee business objec i es. The web se ice cho eog aphy desc ibes he ex e nal obse able beha io
o hese composi ions. Many composi ions may a ailable o same unc ionali y. These composi ions canno
be dis inguished on he basis o unc ional p ope ies. This Quali y o se ices (QoS) may help he use o
selec web se ices and o analyze composi ion o he web se ices. Web se ice cho eog aphy is going o
dic a e implemen a ion o wo k low. This wo k low consis s o se e al asks. Each ask is implemen ed by
web se ices. These se ices a e hos ed in la ge numbe s by di e en se ice p o ide s on di e en se ice
clus e s. The mapping o se ice and ask is di icul issue in un ime en i onmen . The in e ope abili y
be ween se ices is also a g ea p oblem. The selec ion o se ices is e y big issue.
In his pape we ha e p oposed a bio-inspi ed selec ion algo i hm based on dominan ole and p oposed a
disco e y in as uc u e. We ha e also used he clien beha io o imp o e he ailu e o he composi ion o
he se ice.
INDEX TERMS
Cho eog aphy, dominan Role, Web Se ice Selec ion
1 INTRODUCTION
Se e al business o ganiza ions a e in e ac ing hei business ac i i ies h ough he e ogeneous
so wa e applica ions. Many imes hey make in e ac wi hin o ganiza ion and some ime wi h
di e en o ganiza ion. The so wa e o di e en o ganiza ions may use di e en H/W esou ces
and so wa e pla o ms. The basic p oblem aced by indus ies is in e ope abili y. The SOA
se ice o ien ed a chi ec u e p o ides such solu ion. So wa e ac s as a componen o a se ice.
These se ices a e sel desc ibing, in e ope able and loosely coupled. These se ices a e hos ed
on di e en se ice egis ies. Any business ac i i y may con ain mo e han one sub ac i i y and
each ac i i y may implemen ed by a one web se ice. To model he se ice in e ac ions se e al
s anda ds a e p esen like web se ice cho eog aphy and o ches a ion. Web se ice cho eog aphy
de ines common ules o in e ac ions be ween se e al collabo a ing pa ne s o pe o m a ask.
Collabo a ing pa ne s may be web se ices o cho eog aphy. I consis s o se o cho eog aphy,
se o oles, ela ions be ween oles, ac i i ies o be pe o med, documen s o be exchanged and
eac ions o he s a e o collabo a ing pa ne s. Roles ep esen a subse o beha iou o web
se ices. WS-CDL is an XML based language o desc ibe web se ice cho eog aphy. This
language is pla e o m and p og amming language independen . Web se ice cho eog aphy is
used o model in e ac ion be ween di e en o ganiza ions while o ches a ion is used o model
in e ac ion o one o ganiza ion. In se ice o ien ed en i onmen h ee ac o s a e se ice p o ide ,
se ice eques o and se ice b oke . Many se ice p o ide s may publish simila se ices. In
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
24
such cases unc ional p ope ies o se ices a e no su icien o dis inguish among hese se ices.
In such cases non- unc ional equi emen s o use play deciding ole wi h unc ional p ope ies o
he se ice in se ice selec ion. The seman ic ma ch is also impo an ole in se ice selec ion o
minimize ailu e o se ice composi ion. The p esen se ice desc ip ions do no cap u es his
in o ma ion in he WSDL ile. In his scena io clien in oca ion beha iou plays dominan ole o
minimize ailu e o se ice. Gene ally hese se ices a e s o ed in cen alized en i onmen . Due
o emendous g ow h in web se ices his cen alized app oach seems imp ac ical. This
mo i a ed dis ibu ed en i onmen o s o e web se ices. So many people p e e o s o e se ice
using dis ibu ed en i onmen . This en i onmen may be pee o pee en i onmen . Se e al
esea ch e o s ha e been made o index hese web se ices. The pee o pee indexing uses
Cho d, CAN Pas y e c. app oaches. The a chi ec u e o p esen P2P indexing sys em is like da a
lookup sys em and implemen in e ne scale dis ibu ed has able. In pee o pee en i onmen
each pee has esponsibili y o sea ch eques o pee and i se ice is p esen hen p o ide o he
use o ans e que y o neighbou pee and so on. In na u e insec s a e using hei wonde ul
sense o sea ch ood sou ce. All hey a e using e y good me hods o in e ac ion and
communica ion in hei socie y. All hey a e using pa h his o y o ge he ood. In insec socie y
some insec s a e ad e ising he good pa h o all o he insec s. Like an socie y uses phe omone
o ad e ise good pa h and bees a e selec ed good pa ches whe e good nec a may p esen . In his
pape we ha e been p oposed a bio-inspi ed algo i hm o selec bes se ices. In his we also
p oposed a amewo k o se ice disco e y which is also bio-inspi ed. We ha e also gi en he
me hod o composi ion o web se ices using ini e s a e machine. In his composi ion we ha e
gi en he me hod o compu e quali y o se ice a ibu es alues. So many quali y o se ice
a ibu es a e p oposed by di e en au ho s. We ha e only conside ed eliabili y, a ailabili y and
la ency. We ha e le ne wo k ela ed quali y o se ice a ibu es because hey a e nei he in he
con ol o se ice p o ide o he se ice eques o . We ha e used web se ice me a model wi h
QoS as in [5] o use quali y o se ice a ibu es in web se ices.
2. RELATED WORK
P amodh N. e al.[13] ha e p oposed a combined op imiza ion and anking mechanism o
selec ing web se ice. In hei p oposed mechanism op imiza ion is done on he basis o an
phe omone deposi ion p ocess and anking is done on he basis o he QoS pa ame e e alua ion.
QoS pa ame e s selec ed by he au ho a e locali y o e e ence, execu ion- ime, access-coun ,
a ailabili y. Locali y o e e ence pa ame e is used o ind ou he mos ecen ly accessed se ice,
access-coun de ines ha how many imes a se ice is e ched. Execu ion ime is conside ed as
he ime aken by web se ice o p o ide se ice o he use . These pa ame e s upda e hemsel es
e e y ime a se ice is execu ed. Changsheng Zhang e al ha e p esen ed CASS algo i hm o
se ice selec ion. This algo i hm uses clus e ing ela ed sh inking p ocess o p o ide di ec ion o
mo emen o he an s. Zongkai YANG e al[14] ha e p esen ed a combina ion o Gene ic
algo i hm and An Colony algo i hm o compose web se ices in dynamic en i onmen . Sunil R
Dho e e al[15] ha e p oposed a seman ic compose ha inds he op imal leng h among he all
possible composi ion pa hs and ecommend he bes pa h. They ha e also p oposed a amewo k
using agen s o p o ide nego ia ion du ing se ice composi ion. They ha e used di ide and
conque me hodology o pa allel execu ion o composi e web se ices. Alexande T e al[16]
ha e p esen ed a modi ied an colony algo i hm o e alua e QoS o web se ice. Web se ices a e
han selec ed o be composed using he esul o he algo i hm. Comple e phenomenon is di ided
in o se ice selec ion phase, se ice p ocessing phase, G aph building phase and ACO
applica ion phase o p o ide bes ou pu . They p esen ed a bee-inspi ed op imiza ion me hod
which selec s mos op imal solu ion om web se ice composi ion. In his me hod hey enhance
he sea ching by using planning g aph and ma ix o seman ic links which helps in ge ing op imal
solu ion. They ha e also used QoS pa ame e s and seman ic ma ching o se ices o e alua e
whe he he solu ion is op imal o no . They p oposed a me hod o op imizing o ind
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
25
comp ehensi e quali y o composi e se ice. They ha e p oposed a ma hema ical model wi h he
help o QoS agg ega ion model and hey ha e gi en an enhanced a i icial bee colony algo i hm
(I-ABC). The I-ABC algo i hm is based on aboo and chaos pa e n and wi h he o esul o some
simula ion expe imen hey p o ed easibili y and e ec i eness o hei model oe web se ice
composi ion. They ha e p oposed app oxima ion me hod which is based on QoS-awa e se ice
selec ion (SSP) and A i icial Bee Colony algo i hm (ABC) in which hey enhance ABC wi h
g eedy sea ch algo i hm. In his pape hey ha e gi en an algo i hm o ob aining neighbou ing
solu ion and he algo i hm a e pu e online as canonical ABC. They ha e p esen ed ad ance ABC
echnique o local sea ch wi h mo e op imali y and highe easibili y a e. In his pape hey ha e
p oposed app oxima e-mapping and Van-Neumann s a egy o ge be e local sea ch algo i hm
and imp o ed pe o mance in sea ching.
3. OVERVIEW
Cho eog aphy and Social Insec s:
Web se ice cho eog aphy is going o implemen wo k low o composi e web se ices.
This cho og aphy has se o oles which implemen a web se ice. One cho eog aphy
may ha e mo e han one oles and one ole may be used by mo e han one ime in [2],
called as Dominan Role. This dominan ole may play impo an ole in se ice selec ion.
In his pape , we p opose a me a-heu is ic sea ch based on dominan ole o
cho eog aphy. Social insec ’s socie y wo ks like a dis ibu ed sys em in which each
membe is e y simple. Collec i ely hey ep esen a complex social s uc u e. Each
membe wo ks in e y discipline manne o mee he objec i e o hei socie y. All
indi iduals a e sel o ganized. They ha e excellen coo dina ion among hemsel es.
These insec s ha e beha iou s like o aging, di ision o labo , b ood so ing, and
coope a i e anspo , neighbou hood sea ch (local sea ch and global sea ch) o p o ide
ood o hei colonies. In his pape we p opose he algo i hm based on he sea ching
mechanism o an and bee.
An sea ching mechanism:
The mos o an species a e blind. They a e using chemical signa u e o mo e o sea ch loca ions
while humans and o he species a e using isual and acous ic mechanism. These chemical
signa u es a e called phe omones. Some an species a e using ail Phe omone o make pa h in a
g ound o sea ching ood. Some an uses his ail pa h o sea ch ood. These wo p ocesses a e
called as ail lying and ail ollowing phenomena. They a e using indi ec communica ion
ollowed by phe omone deposi ed by o he an . An walking om nes o ood deposi ed
phe omone and o med a ail. An can smell hese deposi s o selec he p obabilis ic pa h on he
basis o s ong smell.
Bee Sea ching Mechanism:
Bees can sea ch hei ood up o a e y long dis ance (14km) in mul iple di ec ions oge he o
ha es nec a o pollen om he lowe pa ches. A small numbe o bees cons an ly sea ch
lowe pa ches. They a e called scou bees. They a e mo ing in e e y di ec ion o sea ch
p obabili y o ood and ha es ed he ood and deposi ed in hi e. I any indi idual ind high
p o i able ood hen go o he dance loo in hi e and pe o m i ual dance called as waggle
dance. Th ough his dance scou bees communica ed he disco e ed lowe pa ches o he idle
bees; hey s a ed o exploi lowe pa ches. The du a ion o he dance decides he a ing o he
scou ood and mo e bees a e ec ui ed in ha es ing o he nec a o m he good a ed lowe
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
26
pa ches. Scou bees a e wo king as disco e y agen and as a ad e ise . Rec ui e may also
pe o m waggle dance o ec ui mo e idle bees o exploi lowe pa ches.
Cho eog aphy and web se ices:
Cho eog aphy desc ibes he in e ac ions be ween se ices. I ep esen s he wo k low o he
composi e se ice. I consis s o se o oles, ela ions and in e ac ions be ween wo oles. One
Role may implemen by mo e han one web se ices. An in e ac ion akes place be ween o oles
using he ela ionship. In his ela ionship one ole ac s as oRole and o he ole ac s as omRole.
In any cho eog aphy mo e han one ask may p esen and each ask
i
may be implemen ed by web
se ice w
ij
whe e i is he ask index and j is he index o web se ice.
Web Se ice and Quali y o Se ice A ibu es:
A. DAmb ogio [5] has p oposed a me a model o web se ice, which include quali y o
se ices(QoS). They sugges ed eliabili y, a ailabili y and access con ol a se ice le el, ne wo k
ela ed QoS a ibu es a po le el, message ela ed QoS a message le el and ope a ion la ency
and ope a ion demand a ope a ion le el. In all hese a ibu es we ha e used eliabili y,
a ailabili y and ope a ion la ency as quali y o se ice a ibu es. Ne wo k ela ed a ibu es a e
nei he in he con ol o se ice consume and se ice p o ide . Fu he hese a ibu es a e based
on o he pa ame e like eliabili y is based on ime be ween ailu e and expec ed ailu e;
a ailabili y is based on expec ed a ailabili y, ime o epai and expec ed ailu e. Ope a ion
la ency is dependen on se ice ime and u na ound ime. Ope a ion demand is compu ed om
a i al a e o he ope a ion eques ed by he clien . Message QoS is compu ed om he message
enc yp ion quali y. Ne wo k ela ed QoS is de i ed om packe loss, bi a e and delay be ween
wo packe s (ji e ). These QoS ha e p o ided by se ice p o ide du ing hos ing o he se ice.
We ha e p esen ed a ool which gene a es a WSDL ile wi h QoS as gi en in [4].
Web se ices desc ip ions a e s o ed in he o m o web se ice desc ip ion language (WSDL) ile
in he egis y including unc ional p ope ies (like inpu / ou pu / p econdi ions) a e s o ed. A.
DAmb ogio [5] has p oposed Q-WSDL me a model . We ha e also been also p oposed o s o e
WSDL de ails wi h QoS in he egis y.
In se ice selec ion quali y o se ice plays impo an ole. We model se o se ices like
node o a g aph da a s uc u e. One ole o cho eog aphy is going o implemen any one
se ice a one le el o a g aph and simila ly o he se ices a e selec ed. These se ice
selec ions make a pa h and decided by he non unc ional p ope y o he se ice. As pe
my p oposal his selec ion has been s a ed o m he dominan Role o he cho eog aphy
and dominan In e ac ion as p oposed [2]. The indi idual web se ice selec ion is
dependen on quali y o se ice alues o a ibu es.
Bees Algo i hm as in[18]:
1. Find Ini ial solu ion wi h andom app oach.
2. Apply i ness unc ion o ge solu ion o m popula ion
3. Fo m new popula ion
4. Selec a ea o ind he neighbou
5. De e mine he pa ch size and ad e ise i using waggle
Dance.
6. Rec ui bees o selec ed si es and apply i ness unc ion.
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
27
7. 7Selec ad e ise bee om each pa ch
8. 8 Use es bees o he same ask.
Figu e 1 Me a Model o Web Se ice wi h QoS.
An Algo i hm as in[15,16,17]:
1. Ini ialize solu ion andomly
2. Find he des ina ion
3. Deposi he phe omone du ing e u n o he colony
4. E alua ion o pa h based on phe omone quali y
5. Res an s ollow he same pa h
P oposed Me hodology
Bee and an sea ching algo i hms a e indi idually e y unique and s ong sea ching algo i hms,
he e we decided o me ge heses algo i hms o ind he bes op imal composi ion. In bee sea ching
algo i hm selec ing highes op imal pa ch based on waggle dance numbe and in an sea ching
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
28
algo i hm colony pa e n is he basic idea. The combina ion o bo h concep s has imp o ed he
sea ching mechanism. Ou algo i hm is inspi ed by bo h an and bee concep , which is gi en as
ollows:
Algo i hm Inspi ed by An & Bee:
1. Find Ini ial solu ion wi h andom app oach.
2. Apply i ness unc ion o ge solu ion o m popula ion
3. Fo m new popula ion
4. Selec a ea o ind he neighbou
5. De e mine he pa ch size and ad e ise i using waggle
Dance.
6. Find he des ina ion
7. Deposi he phe omone du ing a el om one node o ano he node.
8. E alua ion o pa h based on phe omone quali y
9. Res an s ollow he same
4. EXTENDED SERVICE ORIENTED ARCHITECTURE
Se ice disco e y a chi ec u e aces di icul y o in e ope abili y be ween clien eques and
dynamically disco e ed web se ices. WSDL iles a e used o ma ch seman ic de ails like
inpu /ou pu bu do no con ain de ails like in oca ion o de o he ope a ion.[6] p oposed
ex ended SOA a chi ec u e which esol es such kind o issues. They p oposed in e ac ion
p o ocol se ice ex ension (IPSE) o cap u e he o de o in oca ion o he clien .
Figu e 2 BIO-Inspi ed Ex ended Se ice O ien ed A chi ec u e.
They c ea ed his IPSE using clien his o y. These IPSEs educe he ailu e o web se ice
in oca ion. Simila ly we p opose using hese IPSE we cons uc p o ocol which ells abou
clus e wi h good IPSE se ices. We added one nume ic numbe in each IPSE which eco ds
he numbe o he success ul in oca ion o web se ices and in one clus e he summa ion o
all such numbe s ep esen s he clus e usabili y. We ha e been decided he good pa ch
(clus e ) on he basis o hese numbe s. We ha e been p oposed a new IPSE which includes
success ul in oca ion o web se ices and named as BIPSE (Bee inspi ed se ice ex ension
p o ocol). Ini ially he BIPSE is c ea ed using WSDL ile hos ed by se ice p o ide on
egis y wi h in oca ion numbe as ze o and o de o in oca ion o ope a ion as gi en in he
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
29
ile. As clien in okes se ice he o de o in oca ion is e ined o e a ime as gi en in[9] and
numbe is upda ed by each success ul in oca ion. We ha e been p oposed a waggledance
p o ocol which s o es in o ma ion abou clus e quali y. I akes numbe om all IPSEs
p esen in same clus e and added i numbe is g ea e han he o al numbe o web se ices in
ha clus e hen waggledance p o ocol s o es he di e ence be ween hese numbe s. The
clus e s a e a ed on he basis o waggledance p o ocol and good clus e has mo e
waggledance numbe han o he s. [9]
P oposed Algo i hm :
1. Clus e is selec ed on he basis o waggledance numbe
2. The i s web se ice is selec ed on he basis o dominan Role o he web se ice
cho eog aphy and quali y o se ice a ibu es o he web se ices.
3. The nex se ice selec ion is decided by he phe omone concep in an colony me hod.
He e phe omone means quali y o se ice a ibu es.
3.1 I mo e han one pa h exi s hen which ha e mo e phe omone ha pa h is selec ed.
4. Repea s ep 3 o ge he des ina ion.
Figu e 3 Clus e s wi h Web Se ice BIPSE.
Bio Inspi ed Fo mal Model o Web Se ice wi h QoS A ibu es:
We ha e modelled web se ice as a ini e s a e machine which desc ibes co ec execu ion o web
se ice. I consis s o a uple (∑,s0,s , ∆,F)
∑: ∑
A
*∑
q*
∑
BIPSE
whe e
∑
A
is
se o ac i i ies , ∑
q
is se o quali y a ibu es o web se ices and
∑
BIPSE
is
s o e in oca ion o de and numbe o success ul in oca ions.
∑
A
(o1,o2…on) and ∑
q
( el,a l,c,la ) whe e oi is i h ope a ion and el, a l, c and la a e
eliabili y, a ailabili y , cos and la ency espec i ely. ∑
BIPSE
(BIPSE-Bee inspi ed p o ocol se ice
ex ension o1.o2.o3.on, n) whe e oi is he i h ope a ion and chain ep esen s o de o in oca ion
and n is he o al numbe o success ul in oca ion.
so is ini ial s a e and s inal s a e
F: is se o inal s a e
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
30
∆:S*∑S I akes se o inpu ope a ions and quali y o se ice a ibu es om one s a e and a e
ansi ion ans e s o o he s a e. I e u ns a new s a e a e accep ing inpu ope a ions, quali y o
se ice and bee inspi ed p o ocol se ice ex ension which clien his o y o in oca ion o de and
numbe o success ul in oca ion.
Dominan Role:
I is a ole o a web se ice cho eog aphy which appea s maximum numbe o imes in he
in e ac ion as p oposed in [2].Role cap u es he abs ac beha iou o a web se ice. One ole may
implemen many se ices which ha e same unc ionali y. The dominan Role decides he sou ce
web se ice selec ion. In many cases i will be same as a ge web se ice de ined in[8]. They a e
de ined a ge web se ice is a se ice in oked by clien .
Web Se ice Selec ion:
Waggledance p o ocol and Dominan Role a e decided he a ge clus e in which i s web
se ice is selec ed based on BIPSE. The nex se ice selec ion is decided by he nex web se ice
in oca ion. So many web se ices may a ailable o his in oca ion. This is decided by he
quali y o se ice a ibu es o indi idual web se ice p o ided by he se ice p o ide which is
e ined o e a pe iod o ime. In his g aph we a e demons a ing he ela ionship be ween he
se ices and hei success ul in oca ions. In he dis ibu ed en i onmen hese se ices a e hos ed
in di e en egis y domains.
Figu e 4 G aph be ween se ice, In oca ion and hei Clus e s.
Se e al a chi ec u es [9, 11] ha e been p oposed o o ganize hese se ices in he o m o
clus e . Ou p oposal is based on he ac ha his o y o success ul in oca ion may help
o he nex ime s se ice selec ion. As pe ou p oposal we a e showing in he good
pa ches which a e based on he waggle numbe as in he case o bees sea ching
mechanism. As we ha e shown in he g aph some se ices may o e lap in he di e en
clus e .In his g aph double ci cle is ep esen ing he pa ch ha ing highes waggle
numbe .
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
31
Fo mal Model Fo Quali y o Se ice o Web Se ice Composi ion:
We ha e conside ed eliabili y, a ailabili y and la ency quali y se ice a ibu es. Se e al esea ch
e o s ha e conside ed simila ule o se ice composi ion. We ha e conside ed in composi ion
he eliabili y o composi e web se ice is p oduc o eliabili ies o indi idual web se ice
eliabili ies. The a ailabili y o composi e web se ice is also p oduc o a ailabili ies o
indi idual web se ice a ailabili y. The la ency o composi e web se ice is summa ion o la ency
alues o indi idual web se ices. The ules a e summa izes as below.
Table 1 QoS Composi ion Rules.
QoS
A ibu es
Reliabili
y
A ailabili
y
La en
cy
Cos
Sequence
P oduc
P oduc
Su
m
m
a ion
Add
Pa allel
Minimum
Minimum
Max
i
mum
Add
Condi ion
Loop
Same
Same
N*la e
ncy
alue
N*co
s
We a e gi ing ules o web se ice composi ion. The ini e s a e machine consis s o uple ,
F=(O,L,S,Q,ƛ, α) . Whe e O is se o ope a ions, L is se o cons ains, S is se o s a es, Q is se
o quali y o se ice a ibu es , ƛ is ansi ion unc ion om one s a e o ano he s a e and α is
ansi ion unc ion which cap u es quali y o se ice a ibu es. The composi ion ule is gi ing
below.
F1=(O1,L1,S1,Q1,ƛ1, α1) and F2=(O2,L2,S2,Q2,ƛ2, α2) a e wo web se ices FSMs and
esul an ini e s a e machine is gi ing below.
F=(O,L,S,Q,ƛ, α)
L=L1*L2
O=O1*O2
S=S1*S2
Gi en se o cons ain s c={(O,i1….in): O is connec ed o i1,…..in}
ƛ ={(A1,A2,S1,S2,B1,B2)ϵ ƛ ˅ : O is connec ed o (O,i1….in) ϵ (O ϵ B1UB2 i and only i ij ϵ
A1UA2 o all j) } Whe e A1 and A2 a e subse o L1 and L2 espec i ely and B1 and B2 a e
subse o O1 and O2.
α ˭α1. α2 ={q|qϵ α q =q 1*q 2 and qa=qa1*qa2 and ql=ql1+ql2 }
α ˭α1|| α2 ={q|qϵ α q =min(q 1,q 2) and qa=min(qa1,qa2) and ql=ql1+ql2 }
α ˭α1loop α2 ={q|qϵ α q =q 1* o *q 2 and qa=qa1o qa2 and ql=ql1+ql2 }
In abo e exp ession q ,qa and ql a e eliabili y, a ailabili y and la ency quali y o se ice
a ibu es.
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
38
se ices o simila domain. Clus e 1 is ha ing waggle dance (WD) numbe 145 as coun ed by
o al numbe o success ul in oca ion o se ices wi hin clus e 1. Waggle dance numbe o
clus e 2, 3 and 4 is 88, 155, and 6 espec i ely.
CONCLUSION
In his pape we ha e p oposed bio-inspi ed se ice selec ion algo i hm o a gi en web se ice
cho eog aphy. In his selec ion algo i hm we ha e p oposed new sea ch algo i hm. This is based
on na u al selec ion. We ha e used an and bees sea ching algo i hm oge he . We ha e used bees
algo i hm o inding good (pa ches) clus e s.. An algo i hm is used o inding good se ices
wi hin a clus e on he basis o hei QoS pa ame e s as an use o inding hei pa hs using he
phe omones o he se ice selec ion. In his s a egy we a e using dominan ole concep o he
cho eog aphy o s a ou se ice selec ion.
We y o ind new se ice selec ion echniques which will include new QoS pa ame e s
like Geog aphical A ini y and also based on usage beha iou o he web se ices.
R
EFERENCES
[1] Ra i Shanka Pandey and B. D. Chaudha y, "A cos model o pa icipa ing oles based on
cho eog aphy seman ics " IEEE, 2008 Asia-Paci ic Se ices Compu ing Con e ence(APSCC08),
pp.277--283 , 9--12 Dec. 2008, Tiwan , doi: 10.1109/APSCC.2008.117
[2] Ra i Shanka Pandey and B. D. Chaudha y, " A Me a-Model Based P oposal o QOS o WSCDL
cho eog aphy " in P oc. In . Mul i Con . o Enginee s and Compu e Scien is s (IMECS-10) pp. 677-
-683 , 17--19 Ma ch. 2010, Hong --Kong , ISBN: 978-988-17012-8-2
[3] Ra i Shanka Pandey and B. D. Chaudha y, " An Es ima ion o Min-Max o QoS A ibu es o a
Cho eog aphy" in P oc. Ad ances in Enginee ing, Science and Managemen (ICAESM12) pp.872--
878 , 30-31 Ma ch 2012, Nagapa inam, Tamil Nadu ,ISBN: 978-1-4673-0213-5
[4] Ra i Shanka Pandey " A Pla o m Fo Se ice P o ide To Ad e ise Quali y o Se ice Using
WSDL " In e na ional Jou nal o In o ma ion( Acquisi ion (IGIA)o Wo d Scien i ic) accep ed o
publica ion.
[5] A. DAmb ogio," A model-d i en wsdl ex ension o desc ibing he qos o web se ices", In P oc. o
IEEE In . Con . on Web Se ices(ICWS 06), Washing on, DC, USA,2006,p.p. 789-796
[6] G. Dena o, M. Pezz´ e, D. Tosi and Daniela Schilling, "Towa ds Sel -Adap i e Se ice-O ien ed
A chi ec u es", In P oc. o 2006 wo kshop on Tes ing, analysis, and e i ica ion o web se ices and
applica ions, July 17, 2006, Po land, Maine, USA, ACM , p.p. 10--16
[7] Le-Hung Vu, Man ed Hauswi h and Ka l Abe e ," Towa ds P2P-based Seman ic Web Se ice
Disco e y wi h QoS Suppo ", (Da a, In o ma ion, and P ocess In eg a ion wi h Seman ic Web
Se ices, pp.1--15
[8] San-Yih Hwang, Wen-Po Liao, Chien-Hsiang Lee," Web Se ices Selec ion in Suppo o Reliable
Web Se ice Cho eog aphy", In P oc. o 2010 IEEE In e na ional Con e ence on Web Se ices, p.p.
115--122
[9] Mohamed Sellami, Sami Ta a, Zaka ia Maama and B uno De ude, “A Recommende Sys em o
Web Se ices Disco e y in a Dis ibu ed Regis y En i onmen ,” In: 4 h In e na ional Con e ence on
In e ne and Web Applica ions and Se ices (ICIW’09), IEEE May 2009, pp. 418-423.
[10] Na allia Kokash, Aliaksand Bi ukou and Vincenzo D'And ea, “Web Se ice Disco e y Based on
Pas Use Expe ience,” In: 10 h In e na ional Con e ence on Business In o ma ion Sys ems (BIS’07),
Sp inge Ap il 2007, ol 4439, pp 95-107.
[11] F eddy Lecue, “Clus e ing Social Ne wo ks,” In: 5 h In e na ional Wo kshop (WAW’07), Sp inge
Decembe 2007, p.p 57-67.
[12] A. Bi ukou, E. Blanzie i, P. Gio gini, and N. Kokash, “Imp o ing web se ice disco e y wi h usage
da a,” In: So wa e IEEE Decembe 2007, ol 24, issue no. 6, p 47-54
In e na ional Jou nal on Web Se ice Compu ing (IJWSC), Vol.7, No.1, Ma ch 2016
39
[13] P amodh N , S ina h V , S i K ishna A ,” Op imiza ion and Ranking in Web Se ice Composi ion
using Pe o mance Index”, Vol 4 No 4 Aug-Sep 2012, In e na ional Jou nal o Enginee ing and
Technology (IJET),pp-208-213.
[14] Zongkai YANG, Chaowang SHANG†, Qing ang LIU, Chengling ZHAO,” A Dynamic Web Se ices
Composi ion Algo i hm Based on he Combina ion o An Colony Algo i hm and Gene ic Algo i hm”
Jou nal o Compu a ional In o ma ion Sys ems , ol 6, Aug-2010 , pp-2617-2622.
[15] Sunil R Dho e, P o . D M U Kha a ,” QoS Based Web Se ices Composi ion using An Colony
Op imiza ion: Mobile Agen App oach”, In e na ional Jou nal o Ad anced Resea ch in Compu e
and Communica ion Enginee ing,Vol. 1, Issue 7, Sep embe 2012, pp-520-527.
[16] Alexande T ,E. Ki ubaka an, “Op imal QoS based Web Se ice Cho eog aphy using An Colony
Op imiza ion”, In e na ional Jou nal o Compu e Applica ions ,Volume 102 – No.11, Sep embe
2014, pp-39-46.
[17] Changsheng Zhang, Hao Yin, and Bin Zhang , “A No el An Colony Op imiza ion Algo i hm o
La ge Scale QoS-Based Se ice Selec ion P oblem”. Disc e e Dynamics in Na u e and Socie y,
Volume 2013, 2013, A icle ID 815193, pp 9.
[18] R. Sagayam, K. Aklidansewa i, “Compa ison o An Colony and Bee Colony Op imiza ion o Spam
Hos De ec ion,” In: In e na ional Jou nal o Enginee ing Resea ch and De elopmen (IJERD’08),
No embe 2008, ol 4, issue 8, pp 26-32.
