Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
6, N
o
. 2
,
A
p
r
il
201
6, p
p
.
75
9
~
76
9
I
S
SN
: 208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i
2.8
596
7
59
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
A cognitive Approach for Evaluati
ng the Usability of Storage as
a Service in Cl
ou
d Comp
uting Environment
Sha
rmist
ha
Ro
y*
,
Pra
s
ant Kuma
r
Pa
tt
na
ik*, R
a
j
i
b Ma
ll*
*
* School of
Computer
Engineerin
g, KIIT University
, Ind
i
a
** Departmen
t
o
f
Computer Science
a
nd
Engineering, II
T Kharag
pur, India
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
J
u
l 15, 2015
Rev
i
sed
D
ec 13
, 20
15
Accepte
d Ja
n
4, 2016
Cloud computing is a sty
l
e of co
mputi
ng which thrives users requirements b
y
deliv
ering scalable, on-d
e
mand
and pa
y
-
per
-
u
s
e IT services. It offers
differen
t
service models, out of
which
S
t
orage
as
a S
e
rvic
e (S
t
aaS
) is
the
fundam
e
ntal
blo
c
k of Infr
astruct
u
re cl
oud th
at f
u
lfills user’s
excess dem
a
nd
of elas
ti
c com
puting res
ources
.
But cons
iderin
g the com
p
eti
t
i
v
e bus
ines
s
scenario
choosing the b
e
st
clou
d storag
e provid
e
r is
a diff
icult
task.
Thus,
usabilit
y is cons
idered to be th
e
ke
y
perform
an
c
e
indic
a
tor whic
h evalua
tes
the better cloud
storage b
a
sed on
user’s sa
tisfaction. This p
a
per
aims to focus
on the usability evalu
a
tion
of StaaS
provid
e
rs n
a
mely
Google d
r
ive, Dro
p
box and One d
r
ive. This p
a
per
propos
ed a fu
zzy
b
a
sed AHP model for
m
easuring user satisfa
ctio
n
.
Usabilit
y ev
alu
a
tion i
s
carried out bas
e
d on user
feedback throug
h Interview an
d Questionnaire method. Analy
s
is of u
s
er
feedback is don
e based on
the f
u
zzy
app
r
oach
in order to r
e
move vaguness
.
W
h
ereas
, AHP
m
odel is
us
ed for m
easuring satisfaction degree of th
e
differen
t
cloud
storage services
and it so
lves the problem of selecting b
e
st
cloud stor
age.
Keyword:
AH
P m
odel
Fuzzy
StaaS
Usab
ility
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Prasan
t Ku
m
a
r Pattn
aik
,
Sch
ool
o
f
C
o
m
put
e
r
E
ngi
neer
i
ng,
KI
IT Uni
v
er
sity
,
Pat
i
a
,
B
h
uba
ne
swar
, 75
1
0
2
4
,
In
di
a.
Em
a
il: p
a
tn
aik
p
r
asan
tfcs@k
iit.ac.in
,
p
a
t
n
aikp
rasan
t
@g
m
a
il
.co
m
1.
INTRODUCTION
Consi
d
eri
n
g the rece
nt ec
onomic scenario,
users
a
n
d en
terprises are m
a
i
n
ly driv
en
t
o
ward
s h
i
gh
ly
av
ailab
l
e, reli
ab
le and
co
st-effectiv
e infrastru
c
tu
re an
d all th
o
s
e n
e
cessities are fo
rci
n
g
t
o
wards the
im
pl
em
ent
a
t
i
o
n of cl
o
u
d
co
m
put
i
ng envi
r
onm
ent
.
In t
h
e
fi
el
d of cl
o
u
d
com
put
i
ng, St
aaS i
s
a dat
a
st
ora
g
e
m
o
d
e
l wh
ich
facilitates c
l
o
u
d
ap
p
licatio
n
t
o
scale b
e
yon
d
t
h
eir limited
in
frastru
ctu
r
e. Alt
h
oug
h, clou
d
sto
r
ag
e
is h
a
v
i
ng
no
defin
ite arch
itectu
r
e
o
r
set of cap
a
b
ilities,
bu
t
it allo
ws the en
orm
o
u
s
storag
e
o
f
u
s
er’s d
a
ta and
t
h
e i
n
f
o
rm
at
i
on wi
t
h
o
u
t
any
un
de
rl
y
i
ng
har
d
wa
re co
st
. S
o
, st
ora
g
e
has b
ecom
e
an essent
i
a
l
and i
n
t
e
g
r
al
part
of
e
v
e
r
y
hum
a
n being. Im
portance of
stora
g
e has reache
d
s
u
ch an ext
e
nt that the
idea of storing
data i
n
des
k
t
o
p/
PC
’s
has
becom
e
o
b
sol
e
t
e
.
To
re
m
ove such
co
nst
r
ai
nt
cl
ou
d
st
ora
g
e cam
e int
o
t
h
e pi
ct
u
r
e
whi
c
h
facilitates
m
obilit
y in retrievi
ng, shari
n
g and immedi
ate
access
to data. Huge dem
a
nd for getting
acc
ess
t
o
d
a
ta wh
er
e an
d wh
en p
e
op
le
w
a
n
t
, fo
r
c
e th
e in
tro
d
u
c
tion
of
cloud
stor
ag
e pr
ov
id
er
s.
In
th
e last few
years, nu
m
e
ro
u
s
clou
d
st
o
r
age p
r
o
v
i
d
e
rs came in
to
ex
isten
ce bu
t all th
at it req
u
i
res i
s
t
h
e best
st
ora
g
e
pr
ovi
der i
n
t
e
rm
s of st
or
age ca
paci
t
y
, desi
g
n
feat
ure
,
su
pp
ort
e
d e
n
vi
r
onm
ent
,
ec
on
om
i
c
scen
ari
o
and
secu
rity etc. Thu
s
,
u
s
ab
ility e
v
alu
a
tion
is
a
satisfacto
r
y measu
r
e
fo
r fi
n
d
in
g
o
u
t
th
e
b
e
st clo
ud
sto
r
ag
e
for t
h
e fu
lfill
m
e
n
t
o
f
u
s
er’s
req
u
i
remen
t
. Usab
ility is a
q
u
a
lity attrib
u
t
es
wh
ich m
easu
r
es ease of
u
s
e
o
f
an
y serv
ice o
r
p
r
od
u
c
t,
fu
l
f
ill u
s
er’s
d
e
man
d
s and
m
eet
u
s
er’s satisfact
io
n
.
Tod
a
y’s
wo
rl
d
supp
ort variety
of
free cl
o
ud
st
ora
g
e
pr
ovi
d
e
rs f
o
r
aut
o
m
a
t
i
c
upl
oa
d
of
f
i
l
e
s t
o
cl
ou
d s
t
ora
g
e, s
h
ari
n
g
of
fi
l
e
s acr
os
s t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
75
9 – 7
6
9
76
0
world, to syn
c
d
a
ta acro
ss
variou
s
d
e
v
i
ces
wh
ich
f
acilitat
e
s co
llaborative and
produ
ctiv
e
wo
rk
on
the web.
Th
e
p
a
st g
e
n
e
ratio
n
rest
ricts en
orm
o
u
s
d
a
ta sto
r
ag
e and
also d
on’t p
r
ov
id
e
th
e d
a
ta av
ailab
ility facilit
y a
c
ross
m
u
lt
i
p
l
e
devi
c
e
s w
h
i
c
h l
e
a
d
t
o
i
m
m
obi
li
t
y
. Thi
s
pape
r c
o
nsi
d
e
r
s t
h
ree
m
o
st
pop
ul
arl
y
used cl
o
ud s
t
ora
g
e
p
r
ov
id
ers n
a
mely Drop
bo
x,
Goog
le d
r
i
v
e an
d
On
e
d
r
i
v
e
wh
ich
all d
r
iv
es to
ward
s
p
r
ovid
i
n
g
st
o
r
ag
e facility
.
Bu
t to
rate th
e h
i
g
h
l
y u
s
ed
StaaS p
r
o
v
i
d
e
r, u
s
ab
ility ev
alu
a
tio
n
is co
nd
ucted
b
a
sed
o
n
u
s
er feedb
a
ck
an
d
for
an
alyzin
g
and
measu
r
ing
u
s
er satisf
actio
n AH
P and
fu
zzy ap
pro
ach is under
t
ak
en
.
The
rest
of
t
h
e
pa
per
i
s
or
ga
ni
zed
as
fol
l
o
ws:
Sect
i
o
n
2
descri
bes t
h
e
o
b
ject
i
v
e
of
t
h
e
p
r
o
b
l
e
m
.
In
Sectio
n 3 th
e
related
work
s
in
usab
ility evalu
a
tio
n and
ran
k
i
n
g
of cl
o
u
d
serv
ice provid
e
rs are m
e
n
tio
n
e
d.
Sect
i
on
4 an
d
5 de
scri
be
s t
h
e
AH
P m
odel
and
Fuzzy
a
p
p
r
oach
res
p
ect
i
v
el
y
.
The m
e
t
hodol
ogy
of e
v
al
uat
i
n
g
th
e usab
ility o
f
Clo
u
d
Sto
r
ag
e as a
Serv
ice is d
e
scrib
e
d
in Sectio
n
6
.
Sectio
n 7 con
c
lud
e
s th
e
work
.
2.
MOTI
VATI
O
N &
OBJE
CT
IVE
C
l
ou
d St
aaS
i
s
co
nsi
d
e
r
ed
t
o
be t
h
e
fut
u
re
of
t
h
e
pr
esent
day
t
r
a
d
i
t
i
onal
st
ora
g
e
de
vi
c
e
s (e.
g
.
Pe
n
d
r
i
v
e, Hard-d
isk
,
C
D
, DVD
etc.) wh
ere u
s
ers
will
b
e
ab
le
t
o
h
o
s
t and
st
o
r
e th
eir
d
a
ta as
well as ap
p
licatio
n
with
ou
t
reg
a
rd to
wh
ere th
e
d
a
ta is stored
an
d wh
at
un
de
rl
y
i
ng
har
d
war
e
i
s
re
qui
r
e
d
.
Dat
a
st
o
r
ed
i
n
cl
ou
d
envi
ronm
ent should be
a
v
ailable a
n
d can
be acces
se
d a
n
ytim
e from
anywhere
. St
orage se
rvice
provi
ders
sho
u
l
d
al
so g
u
a
rant
ee t
h
at
da
t
a
st
ored s
h
o
u
l
d
p
r
eser
ve t
h
e
i
n
t
e
gri
t
y
and c
onsi
s
t
e
ncy
fo
r im
pro
v
i
n
g cust
om
er
tru
s
t and
relatio
n
s
h
i
p
.
Th
erefo
r
e, u
s
ab
ility ev
alu
a
tion
is
essen
tial to
b
e
carried
ou
t fo
r
p
r
ov
id
ing
cu
st
o
m
ers
wi
t
h
t
h
e
be
st
c
l
ou
d St
aaS
.
A
pi
l
o
t
st
u
d
y
has
bee
n
c
o
n
d
u
ct
ed
to id
en
tify the v
a
riou
s
features
of cl
oud stora
g
e
services
. Stora
g
e capacity is one
of
the
key requirem
ents for Stora
g
e se
rvi
ce
pr
ovi
der
s
fol
l
o
we
d wi
t
h
fi
l
e
shari
n
g,
online
bac
k
up a
n
d a
r
chi
v
ing, easy
na
vigatio
n
while accessing and upl
o
adi
n
g files. Applications
desi
g
n
e
d
s
h
o
u
l
d
be
fl
exi
b
l
e
a
n
d
scal
abl
e
i
n
or
der
t
o
i
m
pro
v
e m
a
rket
val
u
e an
d
pr
o
duct
i
vi
t
y
.
3.
LITERATU
R
E
REVIE
W
Garg
et al. [1
]
in
th
eir wo
rk
p
r
op
o
s
ed
a nov
el fram
e
wo
rk wh
ich
m
easu
r
es th
e Qu
ality o
f
Serv
ice
(Q
oS
)
of cl
o
u
d
ser
v
i
ce p
r
ovi
d
e
rs a
n
d
al
so
co
m
put
e Servi
ce
M
easurem
ent
I
nde
xes
(SM
I
)
whi
c
h i
s
i
m
por
t
a
nt
fo
r com
p
ari
n
g and
ran
k
i
n
g di
ffe
rent
cl
o
ud s
e
rvi
ces.
Qo
S
are m
easures based o
n
seve
ral
at
t
r
i
but
es pr
o
pos
e
d
by
C
l
o
u
d
Ser
v
i
ce M
easu
r
em
ent
I
n
dex
C
o
ns
ort
i
u
m
C
S
M
I
C
[2]
,
w
h
i
c
h
are
fu
rt
he
r a
n
al
y
zed
usi
n
g
AH
P
m
odel
that helps i
n
ranki
ng
of the
cloud services. The a
u
th
ors a
l
so desi
gne
d t
h
e m
e
trics for each QoS attribute
s
whic
h help in
measuring the servi
ce level of each Cloud Service Prov
ide
r
(CSP). The proposed m
echanism
not
o
n
l
y
sol
v
es
t
h
e
pr
o
b
l
e
m
of sel
ect
i
ng t
h
e
b
e
st
C
SP
but
al
s
o
hel
p
s
t
h
e se
r
v
i
ce p
r
ovi
der t
o
i
m
pro
v
e t
h
e
Qo
S.
St
ora
g
e as a Servi
ce i
n
cl
ou
d com
put
i
ng e
nvi
ro
nm
ent
provi
des t
h
e m
e
chani
s
m
for re
pl
i
cat
i
on o
f
local data and to keep sy
nc
h
r
o
n
i
zat
i
on ac
r
o
ss di
ffe
rent
p
l
at
form
s for m
a
i
n
t
a
i
n
i
n
g co
n
s
i
s
t
e
ncy
of dat
a
. Fo
r
main
tain
in
g
sy
n
c
hro
n
i
zation
,
d
a
ta i
n
l
o
cal
file syste
m
s need
t
o
b
e
tr
ack
e
d r
e
gu
lar
l
y
an
d for
th
is pu
rpo
s
e
effect
i
v
e
an
d
ri
gi
d
fi
l
e
o
r
ga
ni
zat
i
ons
are
req
u
i
r
e
d
.
Th
us
, A
r
t
i
a
ga et
a
l
. i
n
20
1
3
[3]
ha
ve
pr
o
pos
ed a
m
echani
s
m
for i
m
provi
ng
t
h
e fi
l
e
sy
st
em
hi
era
r
chy
by
pr
o
v
i
d
i
n
g si
m
u
l
t
a
neo
u
s
vi
e
w
s
of t
h
e
fi
l
e
sy
st
em
organization, known
as
name space virt
ualization. The
m
e
thod
ology suggests the
requirem
ents and
architecture for virt
ualization. Nam
e
space
virtualizatio
n
helps i
n
im
proving the
flexi
b
ilit
y and
usa
b
ility of
cl
ou
d-
base
d se
rvi
ces.
Zhe
ng et
. al
[
4
]
i
n
t
h
ei
r w
o
rk
prese
n
t
e
d a
pers
o
n
al
i
zed
cl
ou
d ra
nki
ng
fram
e
wor
k
t
o
pre
d
i
c
t
t
h
e
ran
k
i
n
g
of
di
f
f
e
rent
cl
o
u
d
se
r
v
i
ces bas
e
d
o
n
Qo
S wi
t
h
o
u
t
t
h
e i
n
vo
cat
i
ons
of a
d
di
t
i
onal
s
e
rvi
ces.
The
ra
nki
ng
i
s
do
ne
wi
t
h
t
h
e hel
p
o
f
t
w
o
r
a
nki
ng
al
g
o
ri
t
h
m
s
nam
e
ly
C
l
oud
R
a
n
k
1
an
d
C
l
ou
d R
a
n
k
2
.
Ex
peri
m
e
nt
al
resul
t
sh
ows th
at th
e
p
r
op
o
s
ed
al
g
o
rith
m
p
e
rform
s
b
e
tter th
an
o
t
her rating
al
go
ri
th
m
s
.
Su
nda
res
w
ara
n
i
n
20
1
2
[
5
]
,
pr
op
ose
d
a
n
a
r
ch
i
t
ect
ure w
h
i
c
h
i
s
base
d
on
cl
o
u
d
b
r
o
k
e
r
f
o
r s
e
l
ect
i
on o
f
best
cl
o
u
d
ser
v
i
ces. The
p
r
op
ose
d
arc
h
itecture provi
des two m
echanism
s
-
o
n
e f
o
r i
n
de
xi
ng
t
h
e cl
ou
d s
e
rvi
c
e
pr
o
v
i
d
er
s a
n
d
anot
her
i
s
t
h
e
que
ry
al
g
o
ri
t
h
m
for ser
v
i
ce
sel
ect
i
on.
The
aut
h
ors
ha
ve i
n
t
r
od
uce
d
t
h
e
br
o
k
er
base
d ap
pr
oac
h
i
n
o
r
de
r t
o
r
e
duce t
h
e h
u
g
e
com
put
at
i
o
n
a
l
l
o
ad d
one
b
y
users o
f
si
m
i
l
a
r prefe
r
e
n
ce
for t
h
e
sel
ect
i
on of di
f
f
ere
n
t
cl
o
u
d
se
rvi
ces
f
r
o
m
a hu
g
e
poo
l of
r
e
so
ur
ces.
Pat
i
n
i
o
t
a
ki
s et
.
al
i
n
2
0
1
3
[
6
]
,
ha
ve a
d
dress
e
d a
fram
e
wor
k
f
o
r e
v
al
uat
i
n
g cl
o
u
d
se
rvi
c
es base
d
o
n
het
e
r
oge
ne
ous
m
odel
of servi
ce charact
eri
s
t
i
c
s and al
so
pr
op
ose
d
t
h
e m
e
t
r
i
c
s for
ran
k
i
n
g cl
ou
d ser
v
i
c
es o
n
th
e b
a
sis
of v
a
ryin
g
lev
e
l of
fu
zzin
e
ss. Th
e
m
o
d
e
l allo
ws fo
r a un
ified
meth
od
o
f
m
u
lti-o
b
j
ectiv
e assessm
e
n
t
of cl
o
ud se
rvi
ces usi
n
g AH
P
m
e
t
hod. M
o
re
o
v
er, t
h
e i
m
preci
se servi
ce charact
eri
s
t
i
cs and va
gue
user
pre
f
ere
n
ces a
r
e
analyzed
usi
n
g
fuzzy a
p
proa
ch.
A sel
ect
i
on
of
best
cl
ou
d se
rv
i
ce for s
p
eci
fi
c
appl
i
cat
i
on
p
u
r
p
o
se i
s
a di
f
f
i
c
ul
t
t
a
sk. T
hus
, Jaha
ni
i
n
20
1
4
[
7
]
ha
ve
pr
o
p
o
s
ed a
W_
SR
(
W
ei
g
h
t
Servi
ce R
a
n
k
)
m
e
t
hodol
o
g
y
fo
r cl
o
u
d
se
r
v
i
ce ra
nki
n
g
b
a
sed
on
QoS feat
ures
of
differe
n
t cloud se
rvices
.
Com
p
arison
with the ot
her a
p
proaches
shows that t
h
e propos
e
d
m
odel
i
s
m
o
re
fl
exi
b
l
e
a
n
d sc
al
abl
e
an
d ca
n
be
use
d
as t
h
e
best
se
rvi
ce
ra
nki
ng
al
g
o
ri
t
h
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A cogn
itive a
p
p
r
oa
ch
f
o
r evalu
a
ting
th
e
u
s
ab
ility o
f
S
t
o
r
age a
s
a S
e
rvice i
n
…
(Pras
a
nt Kumar
P
a
ttnai
k)
76
1
Rehm
an et al.
[8] in 2014 propose
d
a m
e
thod for se
lection of cl
oud services
by utilizin
g the history
of Q
o
S
of di
f
f
e
rent
ser
v
i
ce at
di
ffere
nt
peri
od
of t
i
m
e
and i
n
paral
l
e
l
t
h
e anal
y
s
i
s
of t
h
e resul
t
i
s
don
e usi
n
g
M
C
DM
al
gori
t
hm
. TOPSIS a
nd EL
EC
TR
E t
echni
q
u
es ar
e us
ed
as th
e MCDM alg
o
rithm
.
Fo
r b
e
tter an
alysis,
the use
r
pre
f
erences a
r
e also conside
r
ed
from
time
to
ti
me wh
ich
h
e
l
p
s in th
e
rankin
g
o
f
clou
d serv
ices
am
ong t
h
e av
ai
l
a
bl
e servi
ce
pr
ovi
ders
. T
h
e ra
nki
ng al
s
o
vari
es at
di
ffe
rent
t
i
m
e
peri
od
. T
h
us,
resul
t
s
obt
a
i
ned
at diffe
re
nt tim
e
pe
rio
d
s a
r
e c
o
m
b
ined to
provide t
h
e
ove
ral
l
rating of cl
oud se
rvices
.
Kum
a
r and M
o
ra
rjee i
n
[
9
]
h
a
ve desi
gne
d a
pers
on
al
i
zed s
t
ruct
u
r
al
fram
e
wo
rk
fo
r ra
n
k
i
ng
of cl
ou
d
servi
ces
base
d
on
Q
o
S. T
h
e
p
r
o
p
o
sed a
p
pr
o
ach t
a
r
g
et
s cl
o
ud a
p
pl
i
cat
i
on
whi
c
h re
qui
re
pers
o
n
al
ra
nki
ng
wi
t
h
t
h
e su
p
p
o
r
t
o
f
opt
i
m
al
servi
ces. The
fram
e
wo
rk
nam
e
d C
l
ou
d R
a
n
k
al
s
o
d
o
es
n’t
r
e
q
u
i
re any
i
n
voca
t
i
on o
f
real worl
d se
rvices.
Ro
y an
d
Patt
n
a
ik
[10
]
in
2
013
h
a
v
e
p
r
o
p
o
s
ed
u
s
ab
ility ev
alu
a
tio
n
tech
n
i
qu
es to
ev
alu
a
te th
e
u
s
ab
ility o
f
web
serv
ices
b
a
sed
o
n
sp
eci
fic u
s
ab
ility attr
ib
u
t
es.
Th
e
auth
ors id
en
tified
two
n
e
w u
s
ab
ility
facto
r
s n
a
m
e
ly
d
e
v
i
ce i
n
d
e
p
e
n
d
e
n
ce an
d
pro
v
i
si
o
n
for
p
hys
ically d
i
sab
l
ed
p
e
rso
n
i
n
ad
d
ition
to
t
h
e
cu
rren
t
attrib
u
t
es
wh
ich
en
rich
es th
e
q
u
a
lity of th
e
web serv
ices an
d in
crease the g
l
ob
al
v
a
lu
e of th
e produ
cts.
4.
AHP
M
O
DEL
A
N
D
ITS
CO
NSISTE
NC
Y
CHEC
KI
NG
Anal
y
t
i
c
Hi
era
r
chy
P
r
oces
s (
A
H
P
) m
odel
was de
vel
o
pe
d
by
Pro
f
. T
h
o
m
as L. Saat
y
[11]
w
h
i
c
h i
s
a
Mu
lti-Criteria Decision
Mak
i
n
g
(MC
D
M)
m
o
d
e
l fo
r co
m
p
lex
p
r
o
b
l
em
s. It an
alyses th
e t
h
e
p
r
oble
m
b
y
decom
posi
ng i
n
to a hie
r
arc
h
y of goal, criteri
a, sub-c
r
ite
ria and alternative
s
. It deri
ves the ratio scale weights
fr
om
pai
r
ed co
m
p
ari
s
on i
n
st
e
a
d of assi
gni
n
g
t
h
em
arbi
t
r
ari
l
y
. C
onsi
d
er t
h
e fol
l
owi
ng e
x
am
pl
e whi
c
h sho
w
s
th
e p
a
i
r
-wise co
m
p
ariso
n
m
a
t
r
ix
for
t
h
ree differe
n
t
criteria:
A =
Firstly th
e eighen
v
ector
o
f
t
h
e m
a
trix
will be
calcu
lated
wh
ich
will b
e
the weigh
t
m
a
tri
x
.
St
eps
fo
r
fi
n
d
i
n
g
o
u
t
wei
g
ht
of
m
a
t
r
i
x
A are
desc
ri
be
d
bel
o
w:
Su
m
o
f
each
co
lu
m
n
of th
e matrix
A
Divide
each va
lue across the c
o
lum
n
by its c
o
rres
ponding sum
which
will gene
rate a
ne
w m
a
trix B
Sum
of
eac
h ro
w whi
c
h gi
ves
(3
X 1
)
m
a
t
r
i
x
Div
i
d
e
th
e m
a
trix
b
y
th
e ord
e
r
o
f
t
h
e m
a
trix
to
g
e
t t
h
e
weigh
t
m
a
trix
(w)
Th
e resu
ltin
g
weigh
t
m
a
tix
i
s
:
W
A
=
Co
n
s
isten
c
y of th
e m
a
trix
can b
e
ch
eck
e
d usin
g th
e
fo
llo
wi
ng
step
s:
To c
o
m
put
e
max
fo
r a
p
a
irwise co
m
p
arison
matrix
th
e steps are m
e
n
tio
n
e
d
b
e
low:
Multiply each value
of t
h
e fi
rst colu
m
n
of
the m
a
trix (A) by the wei
g
ht of the
first ite
m
.
Si
m
ilarl
y
m
u
ltiply
each value of
the
se
cond c
o
lum
n
by weight of t
h
e
secong
item
a
n
d go
on
for the rest
of m
a
trix.
Th
en
su
m
th
e v
a
lu
e acro
ss t
h
e ro
ws of t
h
e resu
ltin
g m
a
trix
to
o
b
t
ain th
e ‘weigh
t su
m
’
matrix
(B
).
Divide
each va
lue of B
by the
corres
p
onding value
of t
h
e
weight m
a
trix (W
A
).
Com
pute
the
a
v
era
g
e of
the values whic
h be
com
e
s
ma
x
The
n
t
h
e C
o
ns
i
s
t
e
ncy
In
de
x (
C
I)
need
s t
o
b
e
cal
cul
a
t
e
d
us
i
ng C
I
= (
ma
x
– m
)
/ (m
-1),
whe
r
e m
is the
num
ber of
cri
t
eri
a
.
Here
C
I
of
t
h
e
m
a
t
i
x
A =
(3
.9
3-
3)/
2
=
0.
4
6
5
,
w
h
ere
m
ax
= 3.93
an
d no
. of
alter
n
ativ
es =
3
.
Next
C
o
nsistency Ra
tio
is measu
r
ed
,
wh
ich is th
e ratio
o
f
th
e con
s
isten
c
y in
d
e
x to
th
e
co
rresp
ond
i
ng
r
a
ndo
m
in
d
e
x
i
.
e CR= CI/RI.
There
f
ore,
C
R
of
t
h
e m
a
t
r
i
x
A
=
0.
4
65/
0.
58
=
0.
8
If the
value of CR is less than
0.
1 the
n
the m
a
trix is c
onsiste
nt
and it is acceptable otherwise the
judgem
ent has
to be
cha
n
ged.
Sin
ce, th
e
v
a
lue o
f
CR is
m
o
re th
an
0.1, h
e
nce th
e
m
a
trix
is in
-co
n
s
isten
t
. So
, th
e
p
a
ir-wise
m
a
trix
need
s t
o
be
cha
nge
d.
1 3 5
1
/
3 1
3
1
/
5 1
/
3 1
0.
63
0.
26
0.
31
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
75
9 – 7
6
9
76
2
5.
F
U
ZZY
A
P
PR
OA
CH
Cu
sto
m
er satisfactio
n is th
e on
ly k
e
y fo
r m
eas
u
r
in
g
t
h
e u
s
ab
ility o
f
an
y serv
ice
o
r
produ
ct.
In
terv
iew and
Qu
esti
o
n
n
a
ire
ap
pro
ach
is one o
f
th
e su
itab
l
e u
s
ab
ility ev
alu
a
tio
n
m
e
th
o
d
s to
g
e
t feed
b
a
ck
and
rat
i
ng of
any
p
r
o
d
u
ct
base
d o
n
di
ffe
rent
s
u
b
-
cri
t
e
ri
a. In o
u
r
ap
pr
oac
h
, rat
i
ng of
eac
h q
u
e
st
i
on
i
s
do
ne base
d
on
fi
ve
di
ffe
re
nt
l
i
n
g
u
i
s
t
i
c
va
ri
abl
e
s. B
u
t
,
i
n
or
de
r
to g
e
t t
h
e ap
propriate score
o
f
satisfactio
n
for each sub
-
criteria
fuzzy approach [12]
is ad
op
ted
t
o
co
nv
ert th
e ling
u
i
stic re
su
lt in
to
m
e
m
b
ersh
i
p
v
a
l
u
e.
Hen
c
e, th
e
fuzzy
set
t
h
eo
ry
i
s
ap
pl
i
e
d t
o
han
d
l
e
t
h
e
v
a
gue
ness
i
n
h
u
m
a
n ju
d
g
m
e
nt
. In
o
u
r
ap
p
r
oach
, eac
h l
i
n
gui
st
i
c
v
a
riab
le in
conv
erted
to TFN
[13
]
to ac
hieve
the a
p
propriate score.
6.
METHO
D
OL
OGY OF
US
ABILITY E
V
ALUATIO
N
The
pr
o
p
o
s
ed
F-A
H
P
ap
p
r
oa
ch ca
n
be s
u
b-
di
vi
de
d i
n
t
o
f
o
l
l
o
wi
n
g
st
e
p
s:
Id
en
tificatio
n of crite
ria a
n
d s
u
b-criteria
Calculation
of
m
e
m
b
ership value
of each s
u
b-criteria c
o
rresponding to
differe
nt alterna
tives base
d on
custom
er fee
d
back
using T
F
N
Pair-wise co
mp
ariso
n
of s
u
b-criteria and its
consistency checking
C
a
l
c
ul
at
i
on
of
wei
g
ht
m
a
t
r
i
x
fo
r eac
h c
r
i
t
e
ri
a usi
n
g
A
H
P m
odel
C
a
l
c
ul
at
i
on of
sat
i
s
fact
i
on de
gree
o
f
e
v
ery criterio
n
fo
r d
i
fferen
t altern
atives
6.
1. I
d
en
ti
fi
ca
ti
on
o
f
Cri
t
eri
a
and
Su
b-
Cri
t
eri
a
As st
o
r
a
g
e
has
bec
o
m
e
t
h
e int
e
g
r
al
val
u
e
of m
ode
rn l
i
ve
s an
d
peo
p
l
e
d
e
m
a
nds a l
a
rg
e am
ount
o
f
stora
g
e s
p
ace a
v
ailable a
n
ywhere a
n
ytim
e, so cloud
StaaS
is
an e
ssential re
qui
rem
e
nt for the s
u
stena
n
ce
of IT
i
n
d
u
st
ry
.
I
n
or
der
t
o
fi
n
d
o
u
t
app
r
op
ri
at
e
m
e
t
hod
ol
o
g
y
t
o
sel
ect
best
c
l
ou
d St
aa
S, a
pi
l
o
t
st
u
d
y
ha
s bee
n
con
d
u
ct
ed
by
expe
rt
s f
r
om
IT i
n
dust
r
y
t
o
i
d
ent
i
f
y
the criteria and sub-criteri
a s
u
itable for
prepari
ng
Questi
onnaires
based
on which use
r
fee
dback can
be
collected. User s
a
tisfaction de
gree on each
of the
criteria an
d
sub
-
criteria will also
h
e
lp
th
e serv
ice pro
v
i
d
e
r for im
p
r
o
v
i
ng
its
m
a
rk
et v
a
lu
e. Con
s
id
erin
g
all
t
hose
i
ssues
t
h
e f
o
l
l
o
wi
ng
cri
t
eri
a
an
d s
u
b
-
c
r
i
t
e
ri
a are t
a
ke
n
whi
c
h a
r
e m
e
nt
i
one
d
bel
o
w:
Figure
1. Hiera
r
chical stru
cture of selecting StaaS (St
o
ra
ge a
s
a Se
rvice
)
provi
der
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A cogn
itive a
p
p
r
oa
ch
f
o
r evalu
a
ting
th
e
u
s
ab
ility o
f
S
t
o
r
age a
s
a S
e
rvice i
n
…
(Pras
a
nt Kumar
P
a
ttnai
k)
76
3
6
.
2
.
Ca
lculation
of Membership
Va
lue
of E
a
ch
Sub-Cri
te
rion B
a
sed
on
Cus
t
omer
Fee
dbac
k
using T
F
N
The following
proce
d
ures are
unde
rt
ake
n
to
calculate the fuzzy
m
e
m
b
ersh
ip
v
a
l
u
e of th
e
su
b-criteria
base
d
on
cust
o
m
er feed
bac
k
.
6
.
2
.
1
Method
o
l
o
g
y
of Usab
ility ev
alu
a
tion
:
Usab
ility ev
al
u
a
tio
n is co
ndu
cted
b
a
sed
on
th
e
Qu
estion
n
a
i
r
e-b
a
sed
ev
alu
a
tion
,
wh
i
c
h
invo
lv
ed
resp
o
ndi
ng t
o
a st
anda
rd
q
u
e
s
t
i
o
n
n
ai
re
pre
p
ared t
h
r
o
u
g
h
a
pi
l
o
t
st
u
d
y
ba
sed
on t
h
e ab
o
v
e cri
t
e
ri
a a
n
d
su
b-
cri
t
e
ri
a. Each
quest
i
o
n
n
ai
re
i
s
associ
at
ed wi
t
h
5 di
f
f
e
r
ent
rat
i
n
gs b
a
sed o
n
som
e
l
i
ngui
st
i
c
var
i
abl
e
s.
Li
ng
ui
st
i
c
vari
abl
e
s are m
a
i
n
l
y
hum
an ju
dg
m
e
nt
s havi
ng
no sc
o
r
e b
u
t
s
o
m
e
word
s i
n
nat
u
ral
l
a
ng
ua
ge.
A
rating scale ca
n
be
pre
p
are
d
against eac
h li
nguistic va
riable for
getting the sc
ore
against each c
r
iterion as
m
e
nt
i
oned
i
n
T
a
bl
e 1
bel
o
w
.
Tabl
e
1. R
a
t
i
n
g scal
e
of
Li
n
g
u
i
s
t
i
c
vari
a
b
l
e
s
L
i
nguistic var
i
ables
Sy
m
bol
Value
Ver
y
Unsatisf
actor
y
VUS
2
Unsatisfactor
y US
4
Average
A
6
Satisfactor
y S
8
Very
Satisfactory
VS
10
Quest
i
on
nai
r
es
are
pre
p
are
d
base
d o
n
e
x
pe
rt
g
r
o
u
p
of
pe
opl
e
w
h
o
are
f
a
m
i
li
ar wi
t
h
cl
ou
d st
ora
g
e
servi
ce
pr
o
v
i
d
ers an
d cl
ou
d
com
put
i
ng e
n
v
i
ro
nm
ent
.
The
avera
g
e
of t
h
e score
corres
p
onding to eac
h s
u
b-
cri
t
e
ri
on
i
s
obt
ai
ned a
n
d
depi
ct
ed bel
o
w
i
n
Tabl
e
2.
Table
2.
Avera
g
e sc
ore
of s
ub-cr
iteria b
a
sed
o
n
Lingu
istic
scale
Cloud Stor
age
Pr
ovider
s
Storage Facilit
y
(C
1)
Desi
gn Feature
(C
2)
Supported
E
nvir
o
n
m
ent (
C
3)
Oth
e
r issu
es (C4
)
C11
C12
C13
C14
C21
C22
C23
C31
C32
C33
C41
C42
C43
Dr
op
Box
6.
3
8.
2 8.
0
9.
0 8.
6 8.
8 8.
3
9.
7
9.
8 8.
4 9.
16
9.
48
8.
3
Google
Dr
ive
8.
2
7.
4 5.
6
7.
0 7.
8 8.
0 7.
8
9.
0
7.
5 8.
2 8.
0 8.
4 7.
6
One-
dr
ive
9.
0
6.
0 7.
0
6.
1 7.
8 6.
3 5.
6
8.
6
6.
0 6.
3 6.
1 8.
0 6.
7
6.
2.
2 Co
nver
si
on of
L
i
ng
ui
st
i
c
Vari
abl
e
s t
o
T
F
N
Fuzzy
i
s
o
n
e
of t
h
e
p
o
we
rf
ul
an
d best
m
e
t
hods t
o
re
prese
n
t
l
i
n
g
u
i
s
t
i
c
vari
abl
e
s.
It
rem
oves
vague
n
ess of the hum
a
n judgm
e
nt by converting each
linguistic variable into
Triangular Fuzzy Num
b
er
(TF
N
)
[1
4]
usi
ng
Fuzzy
Set
The
o
ry
. TF
N i
s
a
m
e
m
b
ership
function
whi
c
h is associate
d
with t
h
ree
va
riables
(a, b, c)
whe
r
e
a & c
are the
end
values and b repre
s
ent the peak val
u
e. The m
e
m
b
ersh
ip
v
a
lu
e of TFN lies
bet
w
ee
n 0
a
n
d
1. The
f
u
zzy
T
F
N
i
s
re
prese
n
t
e
d bel
o
w:
µ
LV
(x)
=
(x
-a)
/ (
b
-a
),
a
≤
x
≤
b,
a
≠
b
(c-
x
)
/ (c
-b),
b
≤
x
≤
c, b
≠
c
(
1
)
0,
othe
rwise
Here
, µ
LV
(x) giv
e
s th
e m
e
mb
ersh
ip
v
a
lu
e
for the fuzzy
scale (a,b,c). Based on e
q
uation 1, eac
h
l
i
ngui
st
i
c
vari
a
b
l
e
i
s
c
o
n
v
ert
e
d t
o
TF
N
wi
t
h
t
h
e co
rre
sp
o
n
d
i
ng
scal
e gi
ven
bel
o
w i
n
Ta
bl
e 3.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
75
9 – 7
6
9
76
4
Tab
l
e
3
.
Li
ng
uistic v
a
r
i
ab
les
an
d cor
r
e
spon
din
g
TFN
L
i
nguistic var
i
able
T
F
N
M
e
m
b
er
ship Function
Ver
y
Unsatisf
actor
y (
VUS
)
(
0
,
0,
3)
µ
VUS
(x
)
=
0
,
x
< 0
1
/
3
(
3
-
x
)
, 0
≤
x
≤
3
Unsatisfactor
y
(US)
(
0
,
3,
5)
µ
US
(x
)
=
1
/
3
(
x
-
0
)
,
0
≤
x
≤
3
1
/
2
(
5
-
x
)
,
3
≤
x
≤
5
Average (
A
)
(2, 5, 8)
µ
A
(x
) =
1
/
3
(
x
-
2
)
,
2
≤
x
≤
5
1
/
3
(
8
-
x
)
,
5
≤
x
≤
8
Satisfactor
y
(
S
)
(
5
,
7,
10)
µ
S
(x
) =
1
/
2
(
x
-
5
)
,
5
≤
x
≤
7
1
/
3
(
1
0
-x
), 7
≤
x
≤
10
Very
Satisfactory
(VS)
(
7
,
10,
0)
µ
VS
(x
)
=
1
/
3
(
x
-
7
)
,
7
≤
x
≤
10
0
,
x
>
1
0
Fro
m
th
e tab
l
e, it can
b
e
no
ticed
th
at th
e sa
m
e
ran
g
e
li
es in
m
o
re th
an
on
e ling
u
i
stic scale. To
o
v
e
rco
m
e th
is issu
e, t
h
at v
a
l
u
e is tak
e
n wh
i
c
h
fits the b
e
st (Fo
r
e.g. if t
h
e score(x
)
is 8.1
th
en
it lies in
b
o
t
h
Sat
i
s
fact
ory
a
nd
Very
Sat
i
s
fact
ory
.
Fi
ndi
n
g
o
u
t
t
h
e
m
e
m
b
ershi
p
val
u
e for
x i
t
can be obse
r
ved
t
h
at
for
Satisfacto
r
y (S)
its v
a
lu
e
is
0
.
6
and
for Very Satisf
acto
r
y
(VS) its v
a
lu
e is 0.4.
So
t
h
e
resu
lt
will be
considere
d
a
s
Satisfactory (si
n
ce, 0.6>
0.4)).
Goi
ng
o
n
fi
n
d
i
ng t
h
e m
e
m
b
er
shi
p
val
u
e
of e
ach s
ub-criteri
on t
h
e cust
omer satisfaction
level can be
obtaine
d c
o
rres
p
onding to eac
h c
r
iterion
for each of
the
alte
rnativ
es
in t
h
e
form
of m
a
trix as shown bel
o
w:
T
h
e
av
e
r
ag
e
s
c
o
r
e ob
ta
in
ed
for
ev
e
r
y su
b-
c
r
ite
ri
on
i
n
Tabl
e
2 i
s
con
v
e
r
t
e
d t
o
TFN
usi
n
g
M
e
m
b
ershi
p
f
unct
i
o
n a
s
m
e
nt
i
oned
i
n
Tabl
e
3.
S
o
, t
h
e co
rr
esp
o
n
d
i
n
g T
F
N
of eac
h
su
b-
cri
t
e
ri
on
(
r
ep
re
sent
e
d
as C
) i
s
gi
ve
n
bel
o
w i
n
Ta
bl
e 4:
Tabl
e 4.
T
F
N
obt
ai
ne
d f
o
r
e
a
c
h
c
r
i
t
e
ri
o
n
ba
sed o
n
a
v
era
g
e
sco
r
e
Cloud Stor
age
Pr
ovider
s
Storage Facilit
y
(C
1)
Desi
gn Feature
(C
2)
Supported
E
nvir
o
n
m
ent (
C
3)
Oth
e
r issu
es (C4
)
C11
C12
C13
C14
C21
C22
C23
C31
C32
C33
C41
C42
C43
Dr
op Box
(
5
,
7,
10)
(5
,
7,
10)
(5
,
7,
10)
(7
,
10,
10)
(7
,
10,
10)
(7
,
10,
10)
(5
,
7,
10)
(7
,
10,
10)
(7
,
10,
10)
(5
,
7,
10)
(7
,
10,
10)
(7
,
10,
10)
(5
,
7,
10)
Google Dr
ive
(
5
,
7,
10)
(5
,
7,
10)
(2
,
5,
8)
(5
,
7,
10)
(5
,
7,
10)
(5
,
7,
10)
(5
,
7,
10)
(7
,
10,
10)
(5
,
7,
10)
(5
,
7,
10)
(5
,
7,
10)
(5
,
7,
10)
(5
,
7,
10)
One-drive (7,
10,
10)
(2
,
5,
8)
(5
,
7,
10)
(2
,
5,
8)
(5
,
7,
10)
(5
,
7,
10)
(2
,
5,
8)
(7
,
10,
10)
(2
,
5,
8)
(5
,
7,
10)
(2
,
5,
8)
(5
,
7,
10)
(5
,
7,
10)
6.
3. Pai
r
-
W
i
s
e
C
o
mp
ari
s
o
n
of
Sub
-
C
ri
teri
a and
i
t
s C
o
ns
i
s
tency
C
h
eck
i
n
g
Pair-wise co
m
p
ariso
n
m
a
trix
is essen
tial in o
r
d
e
r
to
establish
th
e im
p
o
r
tan
ce
o
f
each
su
b-criteri
o
n
with
resp
ect to th
e o
t
h
e
r. From th
e h
i
erarchical stru
ctu
r
e
o
f
th
e
p
r
op
osed
m
o
d
e
l, th
e prio
rities of th
e su
b-
criteria n
e
ed
to
b
e
estab
lished
.
Pai
r
-wise co
m
p
ar
i
s
on
s a
r
e d
one
base
d
on t
h
e A
H
P
t
echn
o
l
o
gy
.
Whe
n
com
p
ari
ng a
p
a
i
r
of s
u
b-c
r
i
t
e
ri
a a rat
i
o
o
f
i
m
port
a
nce i
s
est
a
bl
i
s
hed
base
d o
n
t
h
e
bel
o
w
st
anda
rd scal
e
gi
ve
n
bel
o
w by
Saat
y
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A cogn
itive a
p
p
r
oa
ch
f
o
r evalu
a
ting
th
e
u
s
ab
ility o
f
S
t
o
r
age a
s
a S
e
rvice i
n
…
(Pras
a
nt Kumar
P
a
ttnai
k)
76
5
Table 5. Scale of
relative
im
porta
nce of
on
e
criteria w.r.t t
o
o
t
h
e
r criteria
Value of
im
portance
Definition Explanation
1
E
qually
i
m
po
r
t
ant
T
w
o cr
iter
ia
are
of
the sa
m
e
i
m
port
a
nce
3
Slightly
i
m
portant
One criteria is sligh
tly
im
por
tant than the
other
base
d o
n
the
judgem
e
nt
5
Str
ongly
im
por
tant
One cr
iter
i
a is strongl
y
im
por
tant than the ot
her
bas
e
d on t
h
e
judgem
e
nt
7
Very
str
ongly
i
m
p
o
r
t
ant
One cr
iter
i
a is
ver
y
str
ongly
i
m
portant than
the other
cr
iter
ia
based on exper
i
m
e
nt
and judgem
e
nt
9
E
x
tr
em
ely im
por
tant
One cr
iter
i
a varies
ex
tre
m
el
y
with res
p
ect to the other c
r
iteria
2,
4,
6,
8
I
n
term
ediate value be
tween the judgem
e
nt
Values ar
e taken when co
m
p
r
o
m
i
se is
r
e
quier
ed
Reciprocal o
f
above n
on-
zer
o
values
If
criteri
a c
i
has
one of the above
non-
zer
o
value when co
m
p
ared with cj
, the
n
c
j
will
take the r
ecipr
ocal
of that value.
Pair-wise co
m
p
ariso
n
of th
e
su
b-criteria is estab
lish
e
d b
a
sed
o
n
th
e p
ilo
t stu
d
y
d
o
n
e
with
th
e
exp
e
rts
of
t
h
e i
n
or
der
t
o
est
a
bl
i
s
h
c
o
r
r
ect
an
d c
o
nsi
s
t
e
nt
val
u
e.
Fig
u
re
2
.
Pair-wise co
m
p
ariso
n
m
a
trix
for each
criterion
Fo
llowing
th
e
ab
ov
e m
e
th
o
d
for co
nsisten
c
y ch
eck
i
ng
o
f
pair-wise co
m
p
arison
m
a
trix
,
as d
i
scu
ssed
in
sectio
n
4, the Co
n
s
isten
c
y
Ratio
(CR) is an
alyzed
u
s
i
n
g AHP m
o
d
e
l.
Fo
llowing
are
th
e con
s
isten
c
y ratio
(C
R
)
of
t
h
e a
b
ove
m
e
nt
i
oned
pai
r
-wi
s
e
com
p
ari
s
on
f
o
r
eac
h c
r
i
t
e
ri
on:
Tabl
e
6. C
o
n
s
i
s
t
e
ncy
R
a
t
i
o
of
di
f
f
ere
n
t
cri
t
e
r
i
on i
n
AH
P m
odel
Criteria
Consistency
Ratio
Storage Facilit
y
(C
1)
0.41
Design Featur
e
(
C
2)
0.
10
Suppor
ted E
nvir
o
n
m
ent (
C
3)
0.
10
Other
I
ssues
(
C
4)
0
6
.
4
.
Ca
lculation of Weig
ht
Matrix
for E
a
ch Cri
t
eria
us
ing
AHP
Model
Accord
ing
to th
e
AHP m
o
d
e
l th
e
weigh
t
m
a
trix
for ev
ery criterio
n
is calcu
lated
b
e
low:
W
C1
=
W
C2
=
W
C3
=
W
C4
=
In
or
der t
o
fi
t
t
h
e val
u
e i
n
s
t
anda
rd
(1
-1
0
)
scal
e as defi
n
e
d f
o
r Li
ng
ui
s
t
i
c
vari
abl
e
ea
ch val
u
e i
s
m
u
lt
i
p
l
i
e
d by
10
. The
r
ef
o
r
e,
t
h
e wei
g
ht
ag
e of F
r
ee st
o
r
a
g
e (C
11
) = 4
.
2, Fi
l
e
si
ze res
t
ri
ct
i
on (C
12
)
= 2.
6,
Ex
tr
a
f
r
e
e storag
e
(
C
13
) =
1
.
8
,
Supp
or
ted f
i
le typ
e
(
C
14)
= 1
.
6
.
W
h
er
eas, th
e
w
e
igh
t
ag
e of Backu
p
and
A
r
ch
iv
i
n
g (
C
21
)
= 1.4,
N
a
vig
a
tio
n
(
C
22)
= 5
.
7
and
Aut
o
m
a
ti
c upl
oad
f
r
om
devi
c
e
= 2
.
9
.
0.
42
0.
26
0.
18
0.
16
0.
14
0.
57
0.
29
0.
57
0.
14
0.
29
0.
14
0.
43
0.
43
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
75
9 – 7
6
9
76
6
The
weight a
g
e of the c
o
rres
ponding s
u
b-c
r
iteria’
s
of
S
u
pp
o
r
t
e
d e
n
vi
r
o
nm
ent
(C
3)
ar
e O
p
erat
i
n
g
sy
st
em
(C
31
)
= 5.
7,
De
vi
ce i
nde
pe
nde
nt
(C
32
) =
1
.
4
,
Fi
l
e
shari
n
g
(C
33
)
= 2.
9.
Fin
a
lly, Oth
e
r
Issu
es (C
4
)
h
a
v
e
its resp
ective weigh
t
ag
e
of th
e su
b-criteria as Av
ailab
i
l
ity (C4
1
)
=
1.
4, a
n
d
bot
h C
onsi
s
t
e
ncy
(C
4
2
)
an
d R
e
l
i
a
bi
l
i
t
y
(C
43
) =
4.
3
.
The wei
ght matrix of each
criterion is al
so c
onverte
d to TFN usi
ng
Me
m
b
ershi
p
Function as
m
e
nt
i
oned
i
n
T
a
bl
e 3
as s
h
ow
n
bel
o
w i
n
Ta
b
l
e 7:
Tabl
e
7. T
F
N
obt
ai
ne
d
f
o
r
w
e
i
ght
m
a
t
r
i
x
o
f
di
f
f
ere
n
t
s
u
b
-
c
r
i
t
e
ri
on
TFN scale
f
o
r the
weight m
a
t
r
ix
of
each
criterion
Storage Facilit
y
(C
1)
Desi
gn Feature
(C
2)
Supported
E
nvir
o
n
m
ent (
C
3)
Oth
e
r issu
es (C4
)
W
C1
1
W
C1
2
W
C1
3
W
C1
4
W
C2
1
W
C2
2
W
C2
3
W
C3
1
W
C3
2
W
C3
3
W
C4
1
W
C4
2
W
C4
3
TFN scal
e
(2,
5,
8)
(0
,
3,
5)
(0
,
3,
5)
(0
,
3,
5)
(0
,
0,
3)
(2
,
5,
8)
(0
,
3,
5)
(2
,
5,
8)
(0
,
0,
3)
(0
,
3,
5)
(0
,
0,
3)
(2
,
5,
8)
(2
,
5,
8)
6.5.
Calculation of Satisfacti
on De
gree of Every
Criteri
on
for Di
ffere
nt
Alternative
s
Sat
i
s
fact
i
on de
gree f
o
r every
cri
t
e
ri
on i
s
ve
ry
essen
tial in
o
r
d
e
r to
fi
nd
ou
t th
e lack
in
gn
ess of th
e
service.
It also
m
easures the best
cloud storage provide
r
.
The satisfac
t
i
o
n de
gree
of ea
ch cri
t
e
ri
o
n
ca
n be
calculated as:
Satisfactio
n d
e
g
r
ee (Z
Ci
)
,
(i=
1,
2,
..,
n) =
(2)
Each s
u
b-crite
rion com
e
s with di
ffere
n
t wei
ght.
So, wei
g
ht of each
s
u
b-c
r
iterion is
defi
ned
by the
peak val
u
e i.e
C
P
fo
r TF
N
(C
1
, C
P
, C
2
)
Th
er
efo
r
e
w
e
igh
t
of
Ci
(
i
= 1,2,…n ) =
(
3
)
Fo
r Dro
pbo
x, t
h
e satisf
action
d
e
gr
ee
o
f
eac
h
criteria is calculated bel
o
w:
Wei
g
ht
o
f
C
1
f
o
r
D
r
o
p
b
o
x
i
s
cal
cul
a
t
e
d
usi
n
g e
quat
i
o
n
3
be
l
o
w:
Using
equ
a
tio
n 2
,
th
e
satisfactio
n
d
e
g
r
ee of Sto
r
ag
e facility (Z
C1
) =
The i
n
t
e
gral
v
a
l
u
es o
f
t
h
e l
i
n
g
u
i
s
t
i
c
va
ri
ab
l
e
s can
be
obt
ai
ned
usi
ng t
h
e o
p
t
i
m
i
zat
i
on t
echni
que
as
gi
ve
n bel
o
w:
(t
o
+t
p
+4t
m
)
/
4
(
4
)
whe
r
e,
t
o
= op
timistic v
a
lu
e, t
p
= pessi
m
i
st
i
c
val
u
e an
d t
m
=
m
o
st likely value.
So, t
h
e i
n
t
e
gra
l
val
u
es
of t
h
e
TFN
of t
h
e l
i
n
g
u
i
s
t
i
c
vari
a
b
l
e
s are cal
cul
a
t
e
d usi
ng e
q
ua
t
i
on 4 a
n
d
sho
w
n i
n
t
a
bl
e
8
bel
o
w:
Tab
l
e 8
.
In
teg
r
al
v
a
lu
e of d
i
fferen
t
lingu
istic
v
a
riab
les
L
i
nguistic
term
Ver
y
Unsatisf
actory
(VUS
)
Unsatisfactory
(US)
Average (
A
)
Satisf
actory
(S)
Ver
y
Satisf
actory
(VS)
I
n
tegr
al Value
0.
5
2.
83
5
7.
17
9.
5
Hence
,
i
n
t
e
gra
l
val
u
e
of
Z
C1
(calcu
lated
u
s
i
n
g equ
a
tion
4
)
for
Drop
Box = 8.0, wh
ich lies b
e
tween
‘
S
atisf
actor
y’
an
d ‘V
er
y
Satisf
acto
r
y’
. Tab
l
e 9 sh
ow
s th
e
in
tegral valu
e
of
eac
h criterion
for every
altern
ativ
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A cogn
itive a
p
p
r
oa
ch
f
o
r evalu
a
ting
th
e
u
s
ab
ility o
f
S
t
o
r
age a
s
a S
e
rvice i
n
…
(Pras
a
nt Kumar
P
a
ttnai
k)
76
7
Tabl
e 9. Int
e
gr
al
val
u
e o
f
di
ff
er
en
t criterion
for each
alternativ
e
Cloud Storage Providers
Storage Facilit
y
(C
1)
Design Feature
(C
2)
Supported E
nviron
m
e
n
t (C3)
Other issues (C4)
Dr
op Box
8.
0
9.
55
9.
55
9.
23
Google Dr
ive
7.
12
8.
21
9.
55
8.
17
One Dr
ive
7.
47
7.
5
9.
42
8.
07
So
, t
h
e
o
v
e
r
a
ll satisf
actio
n
deg
r
ee
of
Dr
op Bo
x
=
36
.3
3, G
oog
le Dr
iv
e = 33
.0
5
and
O
n
e
Dr
iv
e =
32
.4
6 a
s
s
h
o
w
n i
n
fi
gu
re
3
be
l
o
w.
Fi
gu
re 3.
Ov
erall Satisfactio
n d
e
g
r
ee of
d
i
fferen
t
altern
ativ
es
Hence
,
fr
om
the a
b
o
v
e t
a
bl
e i
t
si
gni
fi
es
t
h
at
Dr
o
p
B
o
x i
s
ha
vi
n
g
bet
t
e
r c
u
st
om
er sat
i
s
fact
i
on
f
o
r
each
of t
h
e crit
eria with
res
p
e
c
t to othe
r cl
oud stora
g
e provi
ders
. Hence
,
the rating of the
above cl
oud st
ora
g
e
p
r
ov
id
er is
D
r
op
B
o
x >
G
oog
le Dr
iv
e > On
e
D
r
i
v
e.
7.
PR
ACTI
C
A
L
IMP
O
RT
A
NCE
OF T
H
E WO
RK
Stora
g
e-a
s
-a
-Service is gai
n
ing m
o
re im
portance in
rece
nt ti
m
e
s because of its ability to provi
de
enorm
ous st
orage s
p
ace
on
pay-as
-
y
o
u
-
g
o
fas
h
i
o
n.
T
hus
, i
t
o
p
t
i
m
i
zes t
h
e
use
of
di
sk st
ora
g
e
spac
e
a
n
d
minimizes the back-e
nd stor
age costs. Ma
ny researc
h
wo
rk
s [1
5, 16
, 1
7
]
fo
cu
ses on
f
i
nd
ing
ou
t th
e b
e
st
asp
ect of th
e sto
r
ag
e serv
ices
in
term
s o
f
fi
nd
ing
o
u
t
th
e SLA
p
a
ram
e
ters to
m
easu
r
e the qu
ality o
f
th
e StaaS
providers. The
researc
h
work [15]
categoriz
ed the SL
A pa
ram
e
ters of st
o
r
ag
e serv
ices in
to
two
categ
ories
n
a
m
e
l
y
, triv
ial and
n
o
n
-
t
r
iv
i
a
l p
a
ram
e
ters b
a
sed
on
Service Lev
e
l
Objectiv
es (SLOs). Th
e tri
v
ial
SLA
p
a
ram
e
ter in
clu
d
e
s av
ailab
ility an
d th
e
non
-tri
v
i
al p
a
ra
meters
are fau
lt
to
leran
ce, p
e
rform
a
n
ce, d
i
saster
reco
very
, Sec
u
ri
t
y
, Gove
r
n
an
ce, Dat
a
Li
fe C
y
cl
e M
a
nagem
e
nt
and err
o
r rat
e
. M
o
re
ov
er, som
e
wor
k
[16]
foc
u
ses
o
n
s
o
m
e
param
e
t
e
rs fo
r m
easuri
n
g t
h
e p
e
r
f
o
r
m
a
nce of t
h
e st
o
r
a
g
e servi
ces
nam
e
l
y
Upl
o
a
d
/
D
o
w
nl
oad
sp
eed
s
at Differen
t
Tim
e
s, Up
l
o
ad
/
D
own
l
o
a
d
sp
eed
o
f
Differen
t
files an
d
C
P
U
u
tilizatio
n
wh
ich
may v
a
ry
b
a
sed
o
n
n
e
twork
co
nd
ition
,
file
ty
p
e
s an
d file size.
Th
e
research
work m
e
n
tio
ned
i
n
[17
]
fo
cu
ses
on
measu
r
ing
th
e
u
s
ab
ility o
f
software as a
serv
ice i.e.
web
s
i
t
es b
y
using
qu
estion
n
a
ire meth
od
an
d statistica
l
approach. The
perform
a
nces of the
we
b
s
ites are m
easu
r
ed
u
s
ing
nu
m
b
er
o
f
click
s
, task
co
m
p
letio
n
ti
me an
d
t
a
sk succe
ss r
a
t
e
.
W
h
e
r
eas
,
t
h
e q
u
est
i
o
n
n
ai
re m
e
t
hod
use
d
WAM
M
I
que
st
i
o
n
n
ai
re
set
t
o
m
easu
r
e t
h
e
subjective
opi
n
ion
of t
h
e users rega
rd
i
n
g th
e b
e
st aspect o
f
t
h
e web
s
ites in
term
s o
f
attractiv
en
ess,
co
n
t
ro
llab
ility, efficiency, learn
a
b
ility an
d
help
fu
ln
ess.
W
h
er
eas, i
n
our
p
r
op
osed
w
o
rk
w
e
id
en
tif
ied th
e se
v
e
ral attribu
t
es and
sub-
attrib
u
t
es wh
i
c
h
prov
ide
b
e
tter v
i
si
b
ilit
y in
term
s o
f
d
e
term
in
in
g
th
e usab
ility o
f
Storag
e-as-a-Serv
i
ce in clo
u
d
co
m
p
u
ting
en
v
i
ron
m
en
t. Th
e attribu
t
es id
en
tified
fo
r m
easu
r
in
g
the u
s
ab
ility o
f
clo
u
d
storag
e
serv
ice
p
r
ov
iders are
Sto
r
ag
e
Facilit
y, Desi
g
n
Feat
u
r
e, Sup
p
o
r
ted Env
i
ro
n
m
en
t
an
d Ot
h
e
r m
i
s
cellan
e
ou
s issues wh
ich are fu
rt
h
e
r
d
i
v
i
d
e
d
in
t
o
su
b-attribu
t
es na
m
e
ly fi
le sto
r
ag
e, file
size restriction, e
x
tra free st
or
ag
e, su
ppo
r
t
ed
f
ile typ
e
,
back
u
p
an
d ar
chi
v
i
n
g,
na
vi
g
a
t
i
on, a
u
t
o
m
a
t
i
c upl
oad f
r
o
m
devi
ce,
o
p
erat
i
ng sy
st
em
, dev
i
ce i
ndepe
n
d
en
t
,
fi
l
e
sh
ari
n
g, av
ailab
ility, co
n
s
isten
c
y, reliab
ility. Th
e attr
i
b
u
t
es are m
easu
r
ed
u
s
i
n
g
u
s
er feed
b
a
ck
m
ech
an
ism
wh
ich
is an
alysed
u
s
ing
a fu
zzy ap
p
r
oach
w
h
i
c
h
p
r
o
v
i
d
es
s
i
gni
fi
ca
nt
insi
ght to t
h
e
research
work.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 2, A
p
ri
l
20
16
:
75
9 – 7
6
9
76
8
8.
CO
NCL
USI
O
N
In t
h
i
s
pape
r,
we ha
ve u
s
ed t
h
e f
u
zzy
ap
pr
o
ach t
o
s
h
o
w
t
h
e si
gni
fi
ca
nce
of l
i
n
gui
st
i
c
va
ri
abl
e
s w
h
i
c
h
are used
f
o
r us
er
fe
ed
bac
k
. M
o
re
ove
r,
t
h
e AH
P
m
odel
i
s
em
pl
oy
ed fo
r
s
e
l
ect
i
on of
cl
o
u
d
st
o
r
ag
e pr
o
v
i
d
e
r
s
.
The c
o
m
b
i
n
at
i
o
n
o
f
bot
h
gi
ves a
p
r
om
i
s
ing
o
u
t
c
om
e. Thi
s
a
r
t
i
c
l
e
pa
ves a
way
t
o
i
n
cl
ude
f
o
ur
d
i
ffere
nt
criteria of cl
oud StaaS
whic
h a
r
e
furt
h
e
r
d
i
v
i
d
e
d in
to su
b-criteria. Tria
ngular Fuzzy Num
b
er is
used
for
g
e
n
e
rating
th
e weigh
t
m
a
trix
for d
i
fferen
t
criterio
n
an
d
u
s
ed f
o
r use
r
fe
edbac
k
m
a
t
r
i
x
. Sat
i
s
fact
i
on
d
e
gre
e
o
b
t
ain
e
d
for each
criteri
o
n
sign
ifies th
e
u
s
ab
i
lity o
f
th
e
cloud
storag
e and
i
t
also
g
i
v
e
s feed
b
a
ck
t
o
th
e
serv
ice
pr
o
v
i
d
er re
ga
r
d
i
n
g t
h
e best
aspect
as wel
l
as t
h
e draw
bac
k
of t
h
e cl
ou
d s
t
ora
g
e. Q
u
est
i
o
n
n
ai
re m
e
t
hod use
d
for g
e
tting
u
s
er feed
b
a
ck
serv
es as th
e
best u
s
ab
ility ev
alu
a
tion
techn
i
qu
e.
It en
ables th
e clo
u
d
serv
ice
p
r
ov
id
er t
o
imp
r
ov
e th
e serv
i
ce q
u
a
lity an
d
in
crease th
e m
a
rk
et v
a
lu
e. In
fu
t
u
re
scop
e
of
work
o
t
h
e
r u
s
ab
ility
crietria m
a
y b
e
con
s
id
ered
fo
r th
e im
p
r
ove
m
e
n
t
o
f
u
s
er satisfaction le
vel
of
cons
umers as
well as
cloud
st
ora
g
e pr
o
v
i
d
ers.
REFERE
NC
ES
[1]
S.
K.
Garg,
et
a
l
.
, “SMICloud: A Framework for Compar
ing and Ranking C
l
oud Service”, in
Fourth I
E
EE
International Co
nference on
Utility and
Cloud
Co
mputing
, 2011
, p
p
. 210-218
.
[2]
C.S.M.I.C
.
(CS
M
IC), “SMI Framework” URL h
ttp
://betawww.cloudcommons.com
/servicemeasur
ementindex
.
[3]
E. Art
i
aga
,
et al.
, “
B
ett
e
r Clo
ud Storage Usa
b
ilit
y
Through
Nam
e
Space V
i
rtual
i
z
a
tion
”
, i
n
6
th
IE
EE/
AC
M
International Co
nference on
Utility and
Cloud
Co
mputing
, 2013
, p
p
. 67-74
.
[4]
Z. Zheng
,
et
al
.,
“Qo
S
Ranking Prediction for C
l
oud Services”,
IEEE Transacti
ons on Parallel and Distributed
Sy
ste
m
s,
vol. 24
(
6
), pp
. 1213-122
2, 2013
.
[5]
S. Sundareswaran and A. Squicciarini,
D. Lin, “A Brokerage-B
ased Appro
ach for Cloud Service S
e
lection
”
, in
5
th
IEEE Internatio
nal Conf
erence
on Cloud Comp
uting
, 2012, pp.
558-565.
[6]
I. Patiniotakis,
et al.
, “Managing
Imprecise Criter
ia in
Cloud Serv
i
ce R
a
nking with
a Fuzzy
Multi-criter
i
a Decision
Making Method
”,
S
e
rvice-Orien
t
ed and
Cloud
C
o
mputing, LNCS
, vol. 8135
, pp
. 3
4
-48, 2013
.
[7]
X. Jahani,
et al.
, “W_
SR: A QoS Based Ranking Appro
ach for Cloud Computing
Service”,
Computer Engineerin
g
and Applications Journal,
vol. 3(
2), pp
. 55-62
, 20
14.
[8]
Z.U. Rehm
an
,
et al.
, “Parallel C
l
oud Service Selection
and R
a
nking Based on QoS Histor
y
”
,
In
ter
national Journal
of Parallel Prog
ramming
, vol. 4
2
, pp
. 820-852
,
2014.
[9]
G.P. Kumar and K. Morarjee,
“Ranking prediction for cloud serv
ices from the past usages”,
International Journ
a
l
of Computer S
c
iences and
Eng
i
n
eering,
vol. 2(9)
, pp. 22-25, 2014
.
[10]
S. Ro
y
and P.K. Pattnaik
,
“Some Popular Us
ability
Evalu
a
tion
Tec
hniqu
es for Websites”,
Advan
c
es in Intell
igent
Systems and Co
mputing
, vo
l. 24
7, pp
. 535-543
,
2013.
[11]
T.L
.
S
a
at
y,
“
H
ow to M
a
ke
a
De
cis
i
on:
The
Anal
yt
ic
Hiera
r
ch
y P
r
oces
s
,
”
Interfa
c
es
, vol. 24(6
)
, pp
. 19-43
, 1994
.
[12]
L.A. Zad
e
h, “Fuzzy
Sets”,
Infor
m
ation and Con
t
ro,
vol. 8(5)
, pp
. 338-353, 1965.
[13]
T.S. Liou and C.W.
Chen
,
“Subjective
appr
aisal of
service
quality
using
fuzzy
linguis
tic
evaluation”,
International
Jo
urnal of Q
uality and
Re
liability Managem
e
nt
,
vol. 23(8)
, pp
. 9
28-943, 2006
.
[14]
J.J. Cronin and
S.A. Tay
l
or, “Measuring serv
ic
e qua
l
ity
:
a
re
-e
xa
mi
nation and
extension”,
Journ
a
l of Marketin
g
,
vol. 56(7)
, pp
. 5
5–68, 1992
.
[15]
N. Ghosh and S.K. Ghosh, “An
Approach to Identif
y
a
nd Monitor SLA Parameters fo
r S
t
orage-a
s
-a-S
ervice C
l
ou
d
Deliver
y Mode
l
”
, in
First International Wor
k
shop on Management
and S
ecurity Technologies for Clou
d
Compuitng
, 201
2, pp
. 724-729
.
[16]
X.
Zhang,
et a
l
.,
“Performance Evalu
a
tion of
Online Backup
Cloud Storage”,
International Journal of Cloud
Applica
tions and
Computing,
vol. 3(3), pp. 20-33,
2013.
[17]
S.
Roy
et al
.,
“
A
quantit
ativ
e app
r
oach to ev
alua
t
e
the us
abil
it
y of
academ
i
c
webs
ites
bas
e
d on hum
an percept
i
on”
,
Egyptian
Journa
l of
Informatics
,
vol. 15(3)
, pp
. 1
59-167, 2014
.
BIOGRAP
HI
ES OF
AUTH
ORS
S
h
arm
i
s
t
ha Ro
y
rece
ived
the
B.
T
ech
and M
.
Te
ch
degrees
in Com
puter S
c
ienc
e
an
d Engin
eering
from
National In
stitute of Techno
log
y
Ag
artala
, in
2010 and 2012,
respect
ivel
y
and
pursuing her
PhD in the field of Cloud Co
m
puting and Usabilit
y
Measur
em
ent from
KI
IT Universi
t
y
.
Moreover, she h
a
s received Gold medal during
her M.Tech. Her research
area includes: Cloud
Usabilit
y,
Secu
ri
t
y
,
and Softwar
e
Engin
eering
Evaluation Warning : The document was created with Spire.PDF for Python.