TELKOM
NIKA
, Vol.12, No
.4, Dece
mbe
r
2014, pp. 10
39~104
4
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v12i4.436
1039
Re
cei
v
ed Se
ptem
ber 12, 2014; Revi
se
d No
vem
ber
3, 2014; Acce
pted No
vem
b
er 20, 201
4
Optimization Research of the OLAP Query Technology
Based on P2P
Chun
feng Wang
Modern Edu
c
ation T
e
chno
l
o
g
y
Ce
nter,
Ya
nche
ng Institut
e of
T
e
chnol
og
y,
Yanch
e
n
g
224
051, Ch
in
a
e-mail: w
c
f@y
c
it.cn
A
b
st
r
a
ct
W
i
th the i
n
cre
a
s
ing
data
of th
e a
ppl
icatio
n s
ystem, the
fast
an
d e
fficient access to
the information
of supp
ort dec
i
s
ion-
mak
i
n
g
an
alysis h
a
s b
e
c
o
me
mor
e
an
d
mor
e
diffic
u
lt. At the sa
me ti
me, a
n
a
l
ysis of
the
data is
no l
o
nger
on a s
i
n
g
le s
e
rver or
a sin
g
le
ent
er
prise
data, b
u
t
on multi
p
le servers,
multip
l
e
dep
artments o
r
multi
p
le e
n
te
rprise dat
a. So, the
origin
al
OLAP technol
ogi
es have a
l
s
o
reveal
ed
ma
ny
shortco
m
i
ngs.
Althou
gh w
e
c
an us
e i
n
d
e
x te
chno
logy
opti
m
i
z
at
io
n
meth
od
to i
m
prov
e the
perfor
m
a
n
ce
in
a
certain exte
nt, but for the cont
inu
a
lly ex
pa
ndi
ng a
m
o
unt
of i
n
formatio
n
in d
a
ta w
a
reho
use
,
its performa
n
c
e
is still the pro
b
l
e
m
nee
de
d to be so
lv
ed. Usi
ng the
meth
od
of P2P net
w
o
rk technol
ogy a
nd OLAP stora
g
e
query
and
qu
ery meth
od, the
pap
er has c
o
n
s
tructed a
d
i
stributed P
2
P-OL
AP netw
o
rk mo
del a
nd th
e
mo
d
e
l
distrib
u
ted
stor
es the
d
a
ta
an
d ce
ntrali
z
e
d
mana
ges
the
no
de. N
e
xt, the
p
aper
h
a
s p
u
t fo
rw
ard the
stora
g
e
and
sh
arin
g sc
he
me
of
mu
ltid
imensi
o
n
a
l
dat
a, OL
AP
query
sche
m
e b
a
se
d o
n
co
ll
abor
ation
sup
port. T
h
e
ide
a
of the sc
h
e
me is th
at the
query
ana
lysis
is do
ne by t
h
e
coord
i
nati
on
a
nd co
op
eratio
n
of OLAP no
de
s.
F
i
nally, the p
a
p
e
r has show
n that the sche
m
e can effect
ivel
y impr
ove the
perfor
m
a
n
ce of
decisio
n an
aly
s
is
by the exp
e
ri
ment.
Ke
y
w
ords
:
P2
P, OLAP, query optimi
z
a
t
i
on,
mu
ltidi
m
ens
ion
a
l data set
1. Introduc
tion
With the
in
creasi
ngly fierce competition
of
ma
rket, the info
rmatio
n play
s a
m
o
re
an
d
more im
porta
nt role for th
e su
rvival an
d devel
op
me
nt of enterp
r
i
s
e
s
. At the same time, al
ong
with the
exte
nsive
appli
c
a
t
ion of d
a
tab
a
se
an
d
d
a
ta warehou
se
tech
nolo
g
ie
s, the e
n
terp
ri
se
informatio
n
system
with th
e a
c
cumulati
on of
time
wil
l
produ
ce
a l
a
rge
am
ount
of data
[1]. T
hat
how to
get u
s
eful de
ci
sion
i
n
formatio
n from th
e
compl
i
cated
data
e
n
vironm
ent a
nd h
o
w to m
a
ke
the rig
h
t an
alysis
and
d
e
ci
sion
-ma
k
in
g have b
e
come a
cruci
a
l link fo
r t
he survival
and
developm
ent
of enterp
r
ises. Onlin
e Analytical Pr
o
c
e
ssi
ng (
O
L
A
P
)
sy
st
em
can h
e
lp u
s
e
r
s t
o
analyze
the dimen
s
ion
a
l stru
cture
of comm
ercial
i
n
formatio
n ef
ficiently and
easily. It is f
a
st
softwa
r
e te
ch
nology of a
ccessing
and
a
nalyzin
g t
he
spe
c
ific
on-li
ne data fo
r
specifi
c
issu
es. It
tries to conve
r
t mass data i
n
data ware
h
ouse to us
efu
l
deci
s
ion info
rmation, so a
s
to reali
z
e th
e
data analy
s
is
and de
ci
sion,
then to help enterp
r
i
s
e
s
to achieve the
deci
s
io
n.
In rece
nt years, with the furthe
r re
sea
r
ch an
d appli
c
ation of OL
AP technolog
y, OLAP
techn
o
logy h
a
s m
ade
co
nsid
era
b
le d
e
velopme
n
t. The d
e
ci
sion
analysi
s
of
the traditio
n
a
l
client/serve
r
mode
an
d th
e wi
dely u
s
e
d
b
r
o
w
ser/se
rver
mod
e
d
e
ci
sion
can
provide
effe
ctive
sup
port for th
e quick de
ci
si
on analy
s
is a
nd trend a
nal
ysis of ma
ssi
ve data [2].
Ho
wever,
wit
h
the exp
a
n
s
ion of n
e
two
r
k
scale a
nd
the ente
r
p
r
ise data, the
existing
method
s hav
e been
sho
w
n some d
e
ficienci
e
s. Anal
ysis of the d
a
ta is no lon
ger on a
sin
g
le
serve
r
or a
singl
e enterp
r
ise d
a
ta, bu
t on multip
le serve
r
s, mul
t
iple depa
rtments or multi
p
le
enterp
r
i
s
e
da
ta. Espe
ciall
y
in the
cu
rrent P2
P n
e
twork tech
nol
ogy continu
e
s
, the
de
cisi
on
analysi
s
by
data co
ordin
a
tion
and
co
operation
sto
r
ed in a m
u
lti node ha
s b
e
com
e
po
ssi
b
le.
Different fro
m
the traditio
nal C/S mod
e
, P2P te
chn
o
logy ca
n organi
ze st
ru
cture by the
way of
netwo
rk
nod
e in the appl
ication laye
r,
which w
ill weaken
the se
rver
role
o
r
even ca
ncel the
serve
r
[3]. Th
e nod
e in
the
P2P syste
m
is b
o
th cli
ent
and
se
rver. I
n
an i
deal
P2
P system,
ea
ch
node
can
acq
u
ire th
e
sam
e
right
s a
nd
ob
ligation
s
, equ
ality excha
n
g
e
data
an
d p
r
ovide
servi
c
e
s
.
The
comm
uni
cation
of ea
ch nod
e is effective with
out
the control
o
f
serve
r
. Rela
tively spea
kin
g
,
the big
g
e
s
t a
d
vantage
of
P2P structu
r
e is that th
e servi
c
e
s
are
distrib
u
ted
to
ea
ch pee
r of
the
netwo
rk [4].
So, the P2P
netwo
rk can
provide
effe
ct
ive se
rvice
s
even
whe
n
o
ne n
ode
is fa
ilure
or ab
norm
a
l
in the whol
e netwo
rk. Distri
but
ed
st
orag
e ba
sed
on P2P is one of the
most
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 103
9 – 1044
1040
importa
nt ap
plicatio
n mod
e
ls an
d is very suitabl
e fo
r cub
e
sto
r
ag
e [5]. Therefore, in the P2P
netwo
rk
environment, ho
w
to respon
d to OLAP quer
y
and the e
s
tab
lishme
n
t of multidimen
sion
al
data more pe
rfectly ha
s be
come a
re
sea
r
ch fo
cu
s of many schola
r
s and exp
e
rts.
Based
on
the a
bove t
e
ch
nolo
g
y d
e
velopme
n
t
and th
e P2
P distri
buted
storage
con
s
tru
c
tion,
this pap
er ha
s co
nst
r
u
c
ted
an OLAP network mo
del
and formul
ated the relate
d
query
schem
e. The sch
e
m
e ca
n set data sha
r
in
g b
y
using the
multidimen
sio
nal data of O
L
AP
node
s, co
mpl
e
te que
ry an
alysis by coo
r
dinatio
n
and
coop
eratio
n, compl
e
te the dynamic j
o
in
and exit of the OLAP node
.
2. Definition
s of Rela
ted
Conc
epts
2.1 OLAP
The O
L
AP te
chn
o
logy i
s
d
e
sig
ned to
suppo
rt the
co
mplex an
alysis op
eration
and the
empha
si
s is
query an
alysi
s
dema
nd fo
r use
r
, and t
hen help
s
th
em to quickl
y
and accura
te
ly
gra
s
p the ov
erall situ
ation
,
market dem
and and d
e
velopme
n
t trend in their re
spe
c
tive are
a
s
, so
as
to mak
e
the right c
h
oices
[6].
OLAP ca
n h
e
lp users to
observe info
rmation from
multiple an
gles an
d a
s
pe
cts in the
usu
a
l way of
thinkin
g
. OLA
P
can hig
h
e
fficiency d
e
e
p
use
s
the hi
stori
c
al d
a
ta for se
rvice
s
.
Its
core
co
ncep
t of OLAP
is "dimen
si
on" [7]. OL
AP can
me
et the
anal
ysis
dem
and
of
multidimen
sio
nal
e
n
viron
m
ental rep
o
rts and que
rie
s
a
nd
b
e
calle
d
t
he
m
u
ltidime
n
sio
nal analy
s
is
tools.
2.2 P2P
P2P has a si
mple definitio
n: "P2P is a
kind
of ap
plication. It makes u
s
e of the
storag
e
spa
c
e, exe
c
u
t
ion cycl
e an
d the c
ontent
re
sou
r
ces i
d
l
ed in the
net
work". Betwe
en word
s, P2
P is
a distrib
u
ted
system lo
cat
ed in appli
c
a
t
ion la
yer an
d each node
can commu
nicate di
re
ctly by
routing p
r
oto
c
ol in the P2
P layer [8].
Each n
ode
with an obje
c
t databa
se (su
c
h a
s
file, MP3,
MPEG etc.) can que
ry the obje
c
t in the other
no
de
s b
y
logical conn
ection of P2P
layer.
2.3 Data Wa
r
e
hous
e
The definitio
n of data wa
reho
use ha
s
many ki
nd
s.
W.H.Inmo
n, the father of the data
wareho
use p
r
opo
sed
that d
a
ta warehou
se is dat
a
set
oriente
d
to
su
pport
man
a
g
e
ment
de
cisi
on-
makin
g
p
r
o
c
e
ss,
su
bje
c
t, integrate
d
, ch
ange
s
with ti
me in th
e "Building th
e
Data Ware
hou
se".
Data
wa
reh
o
u
se
allo
w th
e integ
r
ation
of vario
u
s
appli
c
ation
systems
and
provide
a
un
ified
sup
port data
b
a
se for th
e an
alysis of hi
sto
r
ical d
a
ta.
Data wareho
use
i
s
a se
m
antically con
s
ist
ent
data co
llection and stored
th
e
info
rmation
need
ed fo
r
d
e
ci
sion
ma
ki
ng [9]. It ha
s g
r
e
a
t si
gni
fican
c
e to
im
prove
the
efficien
cy of
da
ta
stora
ge an
d data pro
c
e
ssing ca
pability
.
Use
r
s
c
an
be more flexible in analy
s
is of data an
d
informatio
n a
nd ca
n find the valuable inf
o
rmatio
n,
then will brin
g hu
ge ben
efits to the enterp
r
ise.
3. OLAP Que
r
y
Technolog
y
Problems in the En
v
i
r
onment o
f
P2P
Whe
n
u
s
e
r
s
were
analyze
d when
u
s
in
g the
OLAP
system
will
i
nevitably inv
o
lves th
e
fact table an
d dimen
s
ion
table join and agg
reg
a
tion numb
e
r. A large num
ber of join a
nd
aggregatio
n operation
s
of fact table and dimen
s
i
on t
able will be i
nevitably involved, when u
s
e
r
analyzes the
data using
OLAP syste
m
. These a
c
ti
ons are always the op
eration
s
of very
con
s
umi
ng
system resource
s [10]. So
wh
en
user
sen
d
s a
que
ry ope
ratio
n
, perfo
rma
n
ce
of
OLAP syste
m
is not up t
o
the expe
cte
d
user re
spo
n
se
req
u
ire
m
ent. Although
we can u
s
e i
ndex
techn
o
logy o
p
timization m
e
thod to improve the
perf
o
rma
n
ce in a certai
n extent, but for the
contin
ually e
x
pandin
g
am
ount of i
n
formation in
da
ta wa
reh
o
u
s
e, its p
e
rfo
r
mance i
s
still the
probl
em ne
ed
ed to be solv
ed.
In the environ
ment of P2P,
the basi
c
of OLAP query i
s
multidimen
sional data set. In the
multidimen
sio
nal data set, the storage is
achi
eved
by the multi-dim
e
nsio
nal and
multi-level wa
y,
and the agg
regation records num
be
r of multiple gr
a
nularitie
s will
occupy GB, PB space, the
comp
uting ti
me is
also
ve
ry long [11].
Therefore,
we mu
st firstly solve the
sto
r
age effici
en
cy o
f
multidimen
sio
nal data
set, so a
s
to improve the
analy
s
is effici
en
cy of OLAP. Wh
ile the existin
g
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Optim
i
zation Re
sea
r
ch of the OLAP Qu
ery T
e
chnol
o
g
y Based on
P2P (Chu
nfe
ng Wa
ng)
1041
P2P que
ry a
nalysi
s
al
gorit
hm is compl
e
ted by
the
O
L
AP se
rver, t
he lo
ad
of O
L
AP se
rver is not
been effe
ctively redu
ce an
d the efficien
cy
of the algo
rithm is g
r
eatl
y
reduced.
The
re
sea
r
ch
of this
pap
e
r
will
take all
que
ry
service from
the o
r
iginal
se
rver
to ea
ch
node in P2P-OLAP netwo
rk for pro
c
e
s
si
ng, so as to
realize the cyber sour
ce sharin
g,
netwo
rk
load bal
an
cin
g
, and put fo
rwa
r
d the di
stributed q
uery algorithm b
a
se
d on the
multidimen
sio
nal
data set
s
. The research can improve the deci
s
i
on anal
ysis capability
of the whole system.
4. Optimizati
on Design o
f
the OL
AP Q
u
er
y
Scheme
4.1 Build of the Ne
t
w
o
r
k En
v
i
ronment
Each
nod
e in
the P2P
net
work i
s
both t
he
se
rvi
c
e provider and
se
rvice enjoy.
T
he
mai
n
obje
c
tive of
building
P2P-OLAP network mo
del i
s
to
achi
eve the
sha
r
in
g of
multidimen
sio
nal
data
sets ba
sed
on
sema
ntic level,
re
d
u
ce
the
l
oad
of OLAP
se
rver, an
d rapi
dly co
mplete
the
OLAP qu
ery
requ
est
analy
s
is of e
a
ch n
ode [1
2]. Th
e de
ployment
and
ope
ratio
n
flow chart
of
P2P-OLAP n
e
twork is
sho
w
n in Figu
re
1.
Figure 1. Dep
l
oyment and
operation
flo
w
ch
art of P2P-OLAP network
The followi
ng
is the three i
ndispen
sabl
e
steps of buil
d
ing P2P-O
L
AP network:
(1) Depl
oyme
nt
of
tracki
ng serve
r
The p
r
o
c
e
ss
of sha
r
in
g da
ta need
s a
tracking
se
rver. It is mainly
to assist
a n
ode in th
e
netwo
rk to
a
c
cess the inf
o
rmatio
n of o
t
her no
de
s a
nd coordinat
e the inform
a
t
ion betwe
en
the
different nod
e
s
at the sam
e
time.
The interactio
n betwe
en se
rver an
d the
node i
s
by the HTTP p
r
oto
c
ol. The n
o
d
e
regi
sters
the informati
on of the multidimen
sion
al data
set
s
, IP address
and po
rt to the se
rver. T
h
e
con
n
e
c
tion b
e
twee
n the n
ode
s is
esta
b
lishe
d a
c
cord
ing to the
reg
i
stration
information, an
d th
e
tracking
server will tell other nodes
about the registered inform
ation.
(2) Rele
ased Olap
file
The origi
nal n
ode will creat
e
.Olap
file
to
sh
ar
e the
rel
a
ted info
rmati
on of
multidi
m
ensi
onal
data sets
a
n
d
the add
re
ss of
tra
cki
ng se
rver.
T
he
syst
em
can
upl
oa
d the
file to th
e Web
serve
r
s.
The all nod
e
s
in network have the sa
me role a
n
d
collab
o
ratin
g
data analy
s
is a
nd de
ci
sion
mak
i
ng reques
t
in P2P-OLAP network
.
(3) Sha
r
in
g q
uery between
multidimen
si
onal data
set
s
Nod
e
gets .
O
lap file fro
m
the Web
serve
r
,
obtai
ns the ad
dre
ss of the tra
cki
ng serve
r
address from
.Olap
file,
and
re
giste
r
s inform
ation
to the t
r
a
cki
ng
se
rver a
c
cording
to th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 103
9 – 1044
1042
address. Fi
rstly, node o
b
tains re
gist
rat
i
on info
rmat
i
on from the
tracking
serv
er
stored i
n
the
other no
de
s. Secon
d
ly, node can b
u
ild the con
n
e
c
tio
n
to other no
des a
c
cordin
g to the address
locatio
n
information and
port inform
ation from othe
r node
s. Fin
a
lly, node ca
n compl
e
te the
query an
alysi
s
of multidim
ensi
onal dat
a
sets.
4.2 Distribu
ted Query
of Multidimensional Da
ta Se
ts
This p
ape
r h
a
s u
s
ed th
e CSMD-Tree
structu
r
e a
s
th
e storage m
o
de of multidi
m
ensi
onal
data set
s
in the P2P-O
L
AP network no
de ba
sed o
n
sema
ntic dim
ensi
on hie
r
a
r
chi
c
al chain.
The
multidimen
sio
nal an
alysi
s
o
f
P2P-OLAP
netwo
rk
ma
kes the
service proces
sing
to be di
strib
u
ted
to each nod
e, whi
c
h req
u
ire
s
a
se
rvice meth
od
correspon
ding
to compl
e
te
the que
ry a
nd
analysi
s
, so a
s
to reali
z
e the sha
r
e of multidimen
sion
al data sets b
a
se
d on the semantic level
in
the P2P-OLA
P
mode. Que
r
y optimizatio
n scheme i
s
shown in Figu
re 2.
Figure 2. Flow ch
art of qu
ery optimization schem
e
The pseud
o code of que
ry optim
izatio
n schem
e as foll
ows:
(1)
A node
(P) in
put que
ry an
alysis
statem
ents, t
hen co
nstru
c
t sema
ntic
hie
r
a
r
chical chai
n
se
t
s
(
di
) an
d extended
sema
ntic hierarchi
c
al
chain
sets
(
Di
) .
(2)
Acco
rdi
ng to
the semanti
c
hierarchi
c
al
chai
n
sets,
o
b
tain the l
e
n
g
ths
set
s
of t
he
sema
ntic
extended hi
erarchical ch
ain
|
Di
|.
(3)
Cal
c
ulation
of
the all len
g
th
s is
assig
ned
to |
Dij|
, the value exp
r
e
s
ses the
maxim
u
m length
of
sema
ntic extende
d hierarchical
chai
n se
ts.
(4)
Acco
rdi
ng to |
Dij|
and
Di
, o
b
tains the ext
ende
d hierarchical
chai
n
Di
j
.
(5)
Search the hi
era
r
chical ch
ain matchi
ng
with
Dij
from t
he every .Ola
p file on the Web Se
rver.
(6)
If successful, the node fro
m
the .Olap file will be add
ed to the list.
(7) If
|
Dij
| is NUL
L, remov
e
|
Di
j
| and
Dij
from s
e
t |
Di
| and set
Di
, the program
will jump to (2). If
|
Dij
| is not NULL an
d the
node (P) qu
ery re
sults w
ill be turned
into "not finish" state, the
scheme will
chang
e
Di
to
di.
(8)
If the query is su
cce
ssful
, the schem
e c
an outp
u
t the query result
s, and then exit the
prog
ram.
From th
e ab
ove, the que
ry optimizatio
n sc
hem
e in
this pa
per
can re
alize th
e match
query a
c
cordi
ng to the extende
d se
ma
ntic dime
ns
io
n hierarchy chain, ma
ke the nod
e sh
ared
with the form
of multidime
n
sio
nal data
set in
the n
e
twork a
nd
co
operated
with
other n
ode
s to
compl
e
te the OLAP query
analysi
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Optim
i
zation Re
sea
r
ch of the OLAP Qu
ery T
e
chnol
o
g
y Based on
P2P (Chu
nfe
ng Wa
ng)
1043
5. Experimental An
aly
s
is of OL
AP Q
u
er
y
Optimization Schem
e
The follo
win
g
test an
alyzes the
perfo
rmance of O
L
AP query o
p
timization f
r
om two
asp
e
ct
s. On
e
is th
e time
complexity; other is the
qu
ery a
nalysi
s
rate. Th
e ex
perim
ent b
u
il
d a
Web
se
rver,
a trackin
g
se
rver an
d the
P2P-
OLAP n
e
twork
store
d
two multidi
m
ensi
onal d
a
t
a
sets. T
he foll
owin
g test
wil
l
comp
are two schem
es: OLAP
que
ry optimizatio
n algorith
m
(d
e
noted
as S) an
d P2P pattern mat
c
hin
g
algo
rith
m (den
oted a
s
D).
(1) Multidim
e
n
sio
nal data
sets
sea
r
ch the co
rr
esp
o
n
d
ing .Olap file from the Web
serve
r
and
download
dat
a whe
n
the n
u
mbe
r
of nod
es is u
p
to
n, which all ne
ed a ce
rtain
co
st of time.
The
experim
ent assume
s that the numb
e
r of
node
s is in
a
certain
rang
e. The execu
t
ion time of two
algorith
m
s i
s
sho
w
n in Fig
u
re 3.
Figure 3. The
time load co
mpari
s
o
n
ch
a
r
t of two algorithms
From
the
abo
ve figure, we
can
kno
w
th
a
t
the d
a
ta tup
l
e is the
sha
r
e unit
in P
2
P-OLAP
netwo
rk
and
physi
co
chemi
c
al tre
a
tment
of multid
ime
n
sio
nal data
set can g
r
eat
ly improve th
e
acce
ss
spe
e
d
of network n
ode.
(2) T
he que
ry
analysi
s
rate
of two algorit
hms is
sho
w
n
in Figure 4.
Figure 4. The
OLAP query
time Comp
ari
s
on
cha
r
t of two alg
o
rithm
s
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 103
9 – 1044
1044
The a
nalysi
s
rate of S
is not m
u
ch
differen
c
e
with the
D
algorith
m
wh
en the
multidimen
sio
nal data sets is only inclu
de so
me server node
s o
r
there is fe
wer nod
es in t
he
netwo
rk.
But whe
n
th
e num
be
r of
nod
es in
cre
a
se
s, the
pe
rforma
nce of
S algo
rithm
is m
u
ch
better than
D algo
rithm. Of course, it is not
difficult to understand that this is be
cau
s
e
the
impleme
n
tation of
D algo
ri
thm is compl
e
ted o
n
th
e
server,
wh
en t
he n
u
mbe
r
of nod
es in
crea
se
s,
its perfo
rman
ce is limited t
o
the increa
se load of se
rver.
6. Conclusio
n
s
OLAP has b
e
c
ome a
re
sea
r
ch fo
cu
s of d
e
ci
sion
su
p
p
o
r
t. In recent y
ears, with the
study of
P2P netwo
rk techn
o
logy i
n
-de
p
th, P2P-OLAP mo
del
is mo
re
suit
able for th
e
deployme
nt a
nd
impleme
n
tation of
OLAP
query
sy
stem
. Ho
w to i
m
p
r
ove th
e O
L
AP analysi
s
qu
ery effici
ency
in
the P2P
network e
n
viron
m
ent ha
s be
co
me a
critical
probl
em. T
h
e
main
inn
o
vation of
this pa
per
lies in the foll
owin
g aspe
cts:
(1)
This p
ape
r h
a
s
con
s
tru
c
te
d the OLAP
netwo
rk
mod
e
l ba
sed
on t
he P2P environment. Th
e
model
distri
b
u
ted
store
s
t
he data
an
d
centra
lized m
anag
es th
e n
ode,
which
can obvio
usly
improve the e
fficiency of m
anag
ement.
(2)
This p
ape
r h
a
s p
u
t forward the di
stribu
ted
que
ry alg
o
rithm of mul
t
idimensi
onal
data set i
n
P2P-OLAP
netwo
rk. Th
e arithmeti
c
idea is tha
t
the query
analysi
s
is done by the
coo
r
din
a
tion
and coope
ra
tion of OLAP nodes, wh
ich can obvi
ously imp
r
ov
e the query
effic
i
enc
y
.
Ackn
o
w
l
e
dg
ments
This
work wa
s su
ppo
rted
by the 12th Five Y
ear Plan
ning Fou
ndat
ion of Jian
gsu Province
(Grant No.B-b/2013/0
1
/01
2
). It was al
so s
upp
orted
by the Modern Ed
ucation Te
chnol
o
g
y
Found
ation of
Jiang
su Provin
ce (Grant No.2013
-R247
73).
Referen
ces
[1]
Golfarel
li Matt
eo, Ma
ndre
o
li
F
ederic
a, Pe
nz
o W
ilm
a, R
i
zzi Stefa
no,
T
u
rricchia Elis
a.
A qu
er
y
reformul
a
tion f
r
amew
ork for P2P OLAP.
Procee
din
g
s of the 20
th
Italia
n S
y
mp
osi
u
m on
Advance
d
Datab
a
se S
y
stems, SEBD. 2012: 14
7-1
54.
[2]
Park, Nam Hun, Joo Kil
Hon
g
. Quer
y processin
g
on OLAP s
y
s
t
em
w
i
th cl
ou
d computi
n
g
envir
onme
n
t.
Internati
o
n
a
l Jo
urna
l of Multi
m
edi
a an
d Ubi
q
u
i
tous Eng
i
n
eeri
n
g
. 201
4; 9(5): 169-
174.
[3]
W
ang, Yi
F
e
i.
T
he
w
e
b
for
e
ig
n l
a
n
gua
ge
teach
i
ng
res
e
arch
base
d
o
n
P2P
tech
nol
og
y.
Applied
Mechanics and Materials
. 20
1
4
; 9(20): 13
2-1
36.
[4]
Z
eng, Deg
u
i, Geng, Yishu
a
n
g
. Content di
st
ributi
on mech
a
n
ism in mob
ile
P2P net
w
o
rk.
Journa
l of
Netw
orks
. 201
4; 9(5): 122
9-1
236.
[5]
Gómez, Letici
a
Irene, Gómez
,
Silvia A
lici
a
, Va
isma
n, Alej
a
ndro. Mo
del
in
g
and
qu
er
yin
g
contin
uo
us
fields
w
i
t
h
OLAP cubes.
Intern
ation
a
l Jo
urna
l of Data W
a
reh
ousi
ng a
nd Min
i
ng
. 20
13; 9(3):
22-45.
[6]
Kumar Arvind, Singh
Deepti, Sharma
Vineet.
Achi
e
v
ing
qu
ery o
p
timi
z
a
ti
on us
ing
spars
i
ty
m
a
nagem
e
nt in OLAP system
. Proce
edi
ng
s of the 201
4
Internat
io
nal Confer
ence on
Issues
an
d
Chal
le
nges i
n
Intelli
ge
nt Com
putin
g T
e
chniq
ues, ICICT
201
4. 2014: 7
97-8
01.
[7]
W
e
idn
e
r Martin, Dees Jon
a
than, San
ders
Peter.
F
a
st OLAP query ex
ecutio
n in mai
n
me
mory o
n
larg
e data
in
a
cluster
. Proc
eed
ings -
20
13
IEEE Internat
i
ona
l Co
nfere
n
c
e on B
i
g D
a
ta
. 2013:
518-
524.
[8]
Ciferri Cristi
na
, Ciferri Ricar
do, Gómez Leticia
, Schn
eid
e
r Markus, Vaisma
n Alej
an
dro, Z
i
mán
y
i
Esteban. C
u
b
e
alg
ebr
a: A gen
eric us
er
-c
entric mod
e
l
and
quer
y
l
ang
ua
ge for
OLAP cubes.
Internatio
na
l Journ
a
l of Data
W
a
reho
usin
g a
nd Min
i
ng
. 2
0
1
3
; 9(2): 39-65.
[9]
McCarthy
Mitzi, He Zhen
. Efficient
up
dates
for OLAP ran
ge q
uer
ies o
n
flash mem
o
r
y
.
Co
mpute
r
Journ
a
l
. 20
11; 54(1
1
): 177
3-1
789.
[10]
Z
hu Yu
e-An, Z
han
g Ya
n-So
n
g
, Z
hou
Xua
n
,
W
ang S
han.
C
o
lum
n
-ori
ente
d
qu
er
y e
x
ecuti
o
n e
ngi
ne f
o
r
OLAP based on triplet.
Rua
n
Jian Xu
e Bao/J
ourn
a
l of Softw
are
. 201
4; 25(4
)
: 753-76
7.
[11]
Cuzzocr
ea Alfr
edo, Gu
nop
ul
o
s
Dimitrios. A
decom
positi
o
n
frame
w
ork for
computi
ng
an
d qu
er
yin
g
multidim
ens
ion
a
l OLAP d
a
ta
cubes ov
er
prob
abi
listic re
latio
nal
data.
F
unda
menta I
n
formatica
e
.
201
4; 132(
2): 239-2
66.
[12]
Molin
a Carl
os, Prados-S
uárez
Bel
en, De Re
yes,
Mig
uel
Pra
dos, Pe
ña
Ya
ñ
e
z C
a
rmen.
Improvi
ng t
h
e
und
erstan
dab
ili
t
y
of OLAP qu
eries b
y
sem
a
ntic interpr
e
tati
ons.
Lectur
e
N
o
tes in Co
mpu
t
er Science
.
201
3; 813
2: 17
6-18
5.
Evaluation Warning : The document was created with Spire.PDF for Python.