I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
6
,
No
.
6
,
Dec
em
b
er
201
6
,
p
p
.
2
9
1
1
~
2
9
1
9
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
6i
6
.
1
0
5
5
5
2911
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
jo
u
r
n
a
l.c
o
m/o
n
lin
e/in
d
ex
.
p
h
p
/I
JE
C
E
Big
Data
and M
a
pReduce
Cha
llen
g
es, O
ppo
r
tunit
ie
s
a
nd Trends
S
a
chin Ar
u
n T
ha
ne
k
a
r
1
,
K
.
Su
bra
h
m
a
ny
a
m
2
,
A.
B
.
B
a
g
w
a
n
3
1,
2
De
p
a
rtem
e
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
E
n
g
in
e
e
rin
g
,
KL
Un
iv
e
rsity
,
V
a
d
d
e
sw
a
ra
m
,
G
u
n
tu
r,
In
d
ia
3
D
e
p
a
rte
m
e
n
t
o
f
Co
m
p
u
ter E
n
g
in
e
e
rin
g
,
RS
COE,
T
a
th
w
a
d
e
,
P
u
n
e
,
S
S
P
U,
I
n
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
r
2
1
,
2
0
1
6
R
ev
i
s
ed
Oct
5
,
2
0
1
6
A
cc
ep
ted
Oct
21
,
2
0
1
6
No
w
a
d
a
y
s
we
a
ll
a
re
su
rro
u
n
d
e
d
b
y
Big
d
a
ta.
T
h
e
term
„Big
Da
ta‟
it
se
lf
in
d
ica
tes
h
u
g
e
v
o
lu
m
e
,
h
ig
h
v
e
lo
c
it
y
,
v
a
riet
y
a
n
d
v
e
ra
c
it
y
i.
e
.
u
n
c
e
rtain
ty
o
f
d
a
ta
w
h
ich
g
a
v
e
rise
to
n
e
w
d
if
f
i
c
u
lt
ies
a
n
d
c
h
a
ll
e
n
g
e
s
.
Big
d
a
ta
g
e
n
e
ra
ted
m
a
y
b
e
stru
c
tu
re
d
d
a
ta,
S
e
m
i
S
tru
c
tu
re
d
d
a
ta
o
r
u
n
stru
c
t
u
re
d
d
a
ta.
F
o
r
e
x
isti
n
g
d
a
tab
a
se
a
n
d
s
y
ste
m
s
lo
t
o
f
d
iff
icu
lt
ies
a
re
th
e
re
to
p
ro
c
e
ss
,
a
n
a
ly
z
e
,
sto
re
a
n
d
m
a
n
a
g
e
su
c
h
a
Big
Da
ta.
T
h
e
Big
Da
ta
c
h
a
ll
e
n
g
e
s
a
re
P
r
o
tec
ti
o
n
,
Cu
ra
ti
o
n
,
Ca
p
tu
re
,
A
n
a
ly
sis,
S
e
a
rc
h
in
g
,
V
isu
a
li
z
a
ti
o
n
,
S
to
ra
g
e
,
T
ra
n
s
f
e
r
a
n
d
sh
a
rin
g
.
M
a
p
Re
d
u
c
e
is
a
f
ra
m
e
w
o
rk
u
sin
g
w
h
ich
w
e
c
a
n
w
rit
e
a
p
p
li
c
a
ti
o
n
s
to
p
ro
c
e
ss
h
u
g
e
a
m
o
u
n
t
o
f
d
a
ta,
in
p
a
ra
ll
e
l,
o
n
larg
e
c
lu
ste
rs
o
f
c
o
m
m
o
d
it
y
h
a
rd
w
a
re
in
a
re
li
a
b
le
m
a
n
n
e
r.
L
o
t
o
f
e
ff
o
rts
h
a
v
e
b
e
e
n
p
u
t
b
y
d
if
fe
re
n
t
re
se
a
r
c
h
e
rs
to
m
a
k
e
it
s
im
p
le,
e
a
s
y
,
e
ff
e
c
ti
v
e
a
n
d
e
ff
i
c
ie
n
t.
In
o
u
r
su
rv
e
y
p
a
p
e
r
w
e
e
m
p
h
a
siz
e
d
o
n
th
e
w
o
rk
in
g
o
f
M
a
p
Re
d
u
c
e
,
c
h
a
ll
e
n
g
e
s,
o
p
p
o
rt
u
n
it
ies
a
n
d
re
c
e
n
t
tren
d
s
so
th
a
t
re
se
a
rc
h
e
rs
c
a
n
th
in
k
o
n
f
u
rth
e
r
im
p
ro
v
e
m
e
n
t.
K
ey
w
o
r
d
:
B
ig
d
ata
C
lo
u
d
co
m
p
u
tin
g
Had
o
o
p
HDFS
Ma
p
r
ed
u
ce
Co
p
y
rig
h
t
©
2
0
1
6
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Sach
i
n
A
r
u
n
T
h
an
e
k
ar
,
Dep
ar
te
m
en
t o
f
C
o
m
p
u
ter
Sci
en
ce
an
d
E
n
g
i
n
ee
r
in
g
,
K
L
U
n
iv
er
s
it
y
,
Vad
d
es
w
ar
a
m
,
Gu
n
t
u
r
Dis
tr
ic
t,
An
d
h
r
a
P
r
ad
esh
,
I
n
d
ia.
E
m
ail: sac
h
i
n
.
t
h
an
e
k
ar
@
y
a
h
o
o
.
co
.
in
1.
I
NT
RO
D
UCT
I
O
N
No
w
ad
a
y
s
w
e
all
ar
e
s
u
r
r
o
u
n
d
ed
b
y
h
u
g
e
d
ata.
P
eo
p
le
u
p
lo
ad
/d
o
w
n
lo
ad
v
id
eo
s
,
au
d
io
s
,
i
m
ag
e
s
f
r
o
m
v
ar
iet
y
o
f
d
ev
ices.
Se
n
d
in
g
tex
t
m
e
s
s
a
g
es,
m
u
lti
m
ed
i
a
m
es
s
ag
e
s
,
u
p
d
ati
n
g
th
eir
Fa
ce
b
o
o
k
,
W
h
ats
A
p
p
,
T
w
it
ter
s
tat
u
s
,
co
m
m
en
t
s
,
o
n
l
i
n
e
s
h
o
p
p
in
g
,
o
n
li
n
e
ad
v
er
ti
s
i
n
g
etc.
g
en
er
ate
s
h
u
g
e
d
ata.
A
s
a
r
esu
l
t,
m
ac
h
i
n
es
h
av
e
to
g
e
n
er
ate
an
d
k
ee
p
h
u
g
e
d
ata
to
o
.
Du
e
to
t
h
is
ex
p
o
n
en
t
ial
g
r
o
w
t
h
o
f
d
ata
th
e
a
n
al
y
s
i
s
o
f
th
at
d
ata
b
ec
o
m
e
c
h
alle
n
g
i
n
g
an
d
d
if
f
ic
u
lt.
Fig
u
r
e
1
.
Fo
u
r
Vs o
f
B
ig
Data
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
6
,
No
.
6
,
Dec
em
b
er
201
6
:
2
9
1
1
–
2
919
2912
As
s
h
o
w
n
i
n
Fi
g
u
r
e
1
t
h
e
ter
m
„
B
ig
Data
‟
m
ea
n
s
h
u
g
e
v
o
l
u
m
e,
h
i
g
h
v
elo
cit
y
,
v
ar
iet
y
a
n
d
v
er
ac
it
y
i.e
.
u
n
ce
r
tai
n
t
y
o
f
d
ata.
T
h
is
b
ig
d
ata
is
i
n
cr
ea
s
i
n
g
tr
e
m
e
n
d
o
u
s
l
y
d
a
y
b
y
d
a
y
.
T
h
e
B
ig
d
ata
g
en
er
ated
m
a
y
b
e
s
tr
u
ct
u
r
ed
d
ata,
S
e
m
i
Stru
c
tu
r
ed
d
ata
o
r
u
n
s
tr
u
ctu
r
ed
d
ata.
E
x
is
ti
n
g
d
atab
ase
s
a
n
d
to
o
ls
a
r
e
n
o
t
g
o
o
d
en
o
u
g
h
to
p
r
o
ce
s
s
,
an
al
y
ze
,
s
to
r
e
an
d
m
an
a
g
e
s
u
c
h
a
B
ig
Data
e
f
f
ec
t
iv
el
y
a
n
d
ef
f
icie
n
tl
y
[1
-
3
]
.
2.
H
ADO
O
P
Had
o
o
p
is
an
o
p
en
-
s
o
u
r
ce
,
b
ig
d
ata
s
to
r
ag
e
an
d
h
ig
h
s
p
ee
d
d
ata
p
r
o
ce
s
s
in
g
s
o
f
t
w
ar
e
f
r
a
m
e
w
o
r
k
.
As
s
h
o
w
n
i
n
Fi
g
u
r
e
2
it
u
s
es
cl
u
s
ter
s
o
f
co
m
m
o
d
it
y
h
ar
d
w
ar
e
to
s
to
r
e
an
d
p
r
o
ce
s
s
b
ig
d
a
ta
in
a
d
i
s
tr
ib
u
ted
f
as
h
io
n
.
T
r
e
m
en
d
o
u
s
d
ata
s
to
r
ag
e,
p
r
o
ce
s
s
in
g
t
h
at
d
ata
w
it
h
h
ig
h
s
p
ee
d
ar
e
m
a
k
i
n
g
Had
o
o
p
m
o
r
e
s
u
itab
le
f
o
r
b
ig
d
ata
p
r
o
ce
s
s
in
g
[
4
]
.
Had
o
o
p
clu
s
ter
is
a
s
et
o
f
c
o
m
m
o
d
it
y
m
ac
h
in
e
s
i
n
v
o
lv
i
n
g
h
u
g
e
s
to
r
ag
e
ca
p
ab
ilit
ies,
n
et
w
o
r
k
e
d
to
g
eth
er
in
o
n
e
lo
ca
tio
n
i.e
.
clo
u
d
.
T
h
ese
clo
u
d
m
ac
h
i
n
es
ar
e
th
e
n
u
s
ed
f
o
r
Data
s
to
r
a
g
e
a
n
d
p
r
o
ce
s
s
in
g
.
Fro
m
i
n
d
iv
id
u
al
clie
n
t‟
s
u
s
er
ca
n
s
u
b
m
it
th
e
ir
j
o
b
s
to
clu
s
t
er
.
T
h
ese
clie
n
ts
m
a
y
b
e
p
r
es
en
t
at
s
o
m
e
r
e
m
o
te
lo
ca
tio
n
s
f
r
o
m
t
h
e
Had
o
o
p
clu
s
ter
.
Di
s
tr
ib
u
ted
f
ile
s
y
s
te
m
,
f
aster
p
r
o
ce
s
s
i
n
g
,
f
a
s
ter
d
ata
t
r
an
s
f
er
,
g
o
o
d
f
au
lt
to
ler
an
ce
m
ad
e
Had
o
o
p
v
er
y
e
f
f
icie
n
t
an
d
r
eliab
le.
Had
o
o
p
tr
an
s
f
er
s
co
d
e
to
d
ata
w
h
ich
i
s
tin
y
a
n
d
c
o
n
s
u
m
e
s
les
s
m
e
m
o
r
y
.
A
lo
n
g
w
it
h
d
ata
r
eq
u
ir
ed
th
is
ti
n
y
co
d
e
g
et
ex
ec
u
ted
th
er
e
it
s
el
f
.
As
d
ata
is
lo
ca
ll
y
av
ailab
le
o
n
t
h
at
m
ac
h
i
n
e
lo
t
o
f
ti
m
e,
co
m
p
u
ti
n
g
r
eso
u
r
ce
s
ar
e
s
av
ed
.
Fig
u
r
e
2
.
Had
o
o
p
C
lu
s
ter
I
n
o
r
d
er
to
p
r
o
v
id
e
b
etter
d
ata
av
ailab
ilit
y
an
d
f
au
l
t
to
ler
an
c
e
r
ep
licatio
n
o
f
d
ata
is
d
o
n
e.
User
n
ee
d
n
o
t
to
w
o
r
r
y
ab
o
u
t
p
ar
titi
o
n
i
n
g
t
h
e
d
ata,
d
ata
an
d
tas
k
as
s
i
g
n
m
e
n
t
to
n
o
d
es,
co
m
m
u
n
ica
ti
o
n
b
et
w
ee
n
n
o
d
es.
As Ha
d
o
o
p
h
an
d
les it a
ll,
u
s
er
ca
n
co
n
ce
n
tr
ate
o
n
d
ata
an
d
o
p
er
atio
n
s
o
n
th
at
d
ata.
2
.
1
.
I
m
po
rt
a
nt
F
ea
t
ures o
f
H
a
do
o
p
2
.
1
.
1
.
L
o
w
Co
st
As
Had
o
o
p
is
an
o
p
en
-
s
o
u
r
ce
f
r
a
m
e
w
o
r
k
,
it
is
f
r
ee
.
I
t
u
s
es
co
m
m
o
d
it
y
h
ar
d
w
ar
e
to
s
to
r
e
an
d
p
r
o
ce
s
s
h
u
g
e
d
ata.
Hen
ce
n
o
t
m
u
c
h
co
s
tl
y
.
2
.
1
.
2
.
H
ig
h Co
m
p
uting
P
o
w
er
Had
o
o
p
u
s
es
d
is
tr
ib
u
ted
co
m
p
u
tin
g
m
o
d
el.
Du
e
to
th
is
ta
s
k
ca
n
b
e
d
is
tr
ib
u
ted
a
m
o
n
g
s
t
d
if
f
er
en
t
n
o
d
es
an
d
ca
n
b
e
p
r
o
ce
s
s
ed
q
u
ick
l
y
.
C
lu
s
ter
h
a
v
e
th
o
u
s
a
n
d
s
o
f
n
o
d
es
w
h
ic
h
g
iv
e
s
h
i
g
h
c
o
m
p
u
ti
n
g
ca
p
ab
ilit
y
to
Had
o
o
p
.
2
.
1
.
3
.
Sca
la
bil
it
y
No
d
es
ca
n
b
e
ea
s
il
y
ad
d
ed
an
d
r
em
o
v
ed
.
E
v
e
n
f
ailed
n
o
d
e
s
ca
n
b
e
ea
s
il
y
id
en
t
if
ied
.
Fo
r
all
th
ese
ac
tiv
itie
s
v
er
y
litt
le
ad
m
i
n
is
tr
a
tio
n
is
r
eq
u
ir
ed
.
2
.
1
.
4
.
H
ug
e
a
nd
F
lex
ibl
e
Sto
r
a
g
e
Ma
s
s
i
v
e
d
ata
s
to
r
ag
e
i
s
av
a
ilab
le
d
u
e
to
th
o
u
s
an
d
s
o
f
n
o
d
es
in
t
h
e
cl
u
s
ter
.
I
t
s
u
p
p
o
r
ts
b
o
th
s
tr
u
ct
u
r
ed
an
d
u
n
s
tr
u
ctu
r
ed
d
ata.
No
p
r
ep
r
o
ce
s
s
in
g
i
s
r
eq
u
ir
e
d
o
n
d
ata
b
ef
o
r
e
s
to
r
in
g
it.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
B
ig
Da
ta
a
n
d
Ma
p
R
e
d
u
ce
C
h
a
llen
g
es,
Op
p
o
r
tu
n
ities
a
n
d
T
r
en
d
s
(
S
a
ch
in
A
r
u
n
Th
a
n
ek
a
r
)
2913
2
.
1
.
5
.
F
a
ult
T
o
l
er
a
nce
a
nd
Da
t
a
P
r
o
t
e
ct
io
n
I
f
an
y
n
o
d
e
f
a
ils
t
h
e
tas
k
s
in
h
an
d
ar
e
au
to
m
at
icall
y
r
ed
ir
ec
t
ed
to
o
th
er
n
o
d
es.
Mu
ltip
le
co
p
ies
o
f
all
d
ata
ar
e
au
to
m
atica
ll
y
s
to
r
ed
.
Du
e
to
t
h
is
ev
e
n
i
f
an
y
n
o
d
e
f
ail
s
t
h
at
d
ata
is
av
a
ilab
le
o
n
s
o
m
e
o
th
er
n
o
d
es
also
.
2
.
2
.
Co
m
pa
riso
n o
f
H
a
do
o
p w
i
t
h
T
ra
d
it
io
na
l R
DB
M
S
T
ab
le
1
is
s
h
o
w
i
n
g
th
e
d
i
f
f
er
en
ce
b
et
w
ee
n
tr
ad
itio
n
al
R
D
B
MS
an
d
Had
o
o
p
w
h
ic
h
i
n
d
i
ca
tes
th
a
t
tr
ad
itio
n
al
d
atab
ases
ar
e
n
o
t th
at
m
u
c
h
s
u
p
p
o
r
tiv
e
f
o
r
b
ig
d
at
a.
T
ab
le
1
.
Had
o
o
p
-
R
DB
MS
C
o
m
p
ar
i
s
io
n
S
r
.
N
o
.
H
a
d
o
o
p
R
D
B
M
S
01
H
a
d
o
o
p
s
t
o
r
e
s b
o
t
h
st
r
u
c
t
u
r
e
d
a
n
d
u
n
st
r
u
c
t
u
r
e
d
d
a
t
a
.
R
D
B
M
S
st
o
r
e
s d
a
t
a
i
n
a
s
t
r
u
c
t
u
r
a
l
w
a
y
.
02
S
Q
L
c
a
n
b
e
i
m
p
l
e
me
n
t
e
d
o
n
t
o
p
o
f
H
a
d
o
o
p
a
s
t
h
e
e
x
e
c
u
t
i
o
n
e
n
g
i
n
e
S
Q
L
(
st
r
u
c
t
u
r
e
d
q
u
e
r
y
l
a
n
g
u
a
g
e
)
i
s u
s
e
d
.
03
S
c
a
l
i
n
g
o
u
t
i
s
n
o
t
t
h
a
t
m
u
c
h
e
x
p
e
n
si
v
e
a
s ma
c
h
i
n
e
s c
a
n
b
e
a
d
d
e
d
o
r
r
e
mo
v
e
d
w
i
t
h
e
a
se
a
n
d
l
i
t
t
l
e
a
d
mi
n
i
st
r
a
t
i
o
n
.
S
c
a
l
i
n
g
u
p
(
u
p
g
r
a
d
a
t
i
o
n
)
i
s v
e
r
y
e
x
p
e
n
si
v
e
.
04
B
a
si
c
d
a
t
a
u
n
i
t
i
s
k
e
y
/
v
a
l
u
e
p
a
i
r
s.
B
a
si
c
d
a
t
a
u
n
i
t
i
s re
l
a
t
i
o
n
a
l
t
a
b
l
e
s.
05
W
i
t
h
M
a
p
R
e
d
u
c
e
w
e
c
a
n
u
se
s
c
r
i
p
t
s
a
n
d
c
o
d
e
s
t
o
t
e
l
l
a
c
t
u
a
l
s
t
e
p
s i
n
p
r
o
c
e
ssi
n
g
t
h
e
d
a
t
a
.
W
i
t
h
S
Q
L
w
e
c
a
n
st
a
t
e
e
x
p
e
c
t
e
d
r
e
su
l
t
a
n
d
d
a
t
a
b
a
se
e
n
g
i
n
e
d
e
r
i
v
e
s i
t
.
06
H
a
d
o
o
p
i
s
d
e
si
g
n
e
d
f
o
r
o
f
f
l
i
n
e
p
r
o
c
e
s
si
n
g
a
n
d
a
n
a
l
y
si
s o
f
l
a
r
g
e
-
sca
l
e
d
a
t
a
.
R
D
B
M
S
i
s
d
e
si
g
n
e
d
f
o
r
o
n
l
i
n
e
t
r
a
n
s
a
c
t
i
o
n
s.
2
.
3
.
H
a
do
o
p Sy
s
t
e
m
P
rinci
ples
2
.
3
.
1
.
Sca
lin
g
O
ut
I
n
T
r
ad
itio
n
al
R
DB
MS
it
is
q
u
ite
d
i
f
f
ic
u
lt to
ad
d
m
o
r
e
h
ar
d
w
ar
e,
s
o
f
t
w
ar
e
r
eso
u
r
ce
s
i.e
.
s
ca
le
u
p
.
I
n
Had
o
o
p
th
is
ca
n
b
e
ea
s
il
y
d
o
n
e
i.e
.
s
ca
le
d
o
w
n
.
2
.
3
.
2
.
T
ra
ns
f
er
co
de
t
o
da
t
a
I
n
R
DB
MS
g
en
er
all
y
d
ata
is
m
o
v
ed
to
co
d
e
an
d
r
esu
lts
ar
e
s
to
r
ed
b
ac
k
.
A
s
d
ata
is
m
o
v
i
n
g
t
h
er
e
i
s
al
w
a
y
s
a
s
ec
u
r
it
y
th
r
ea
t.
I
n
H
ad
o
o
p
s
m
al
l
co
d
e
is
m
o
v
ed
to
d
ata
an
d
it
i
s
e
x
ec
u
ted
t
h
er
e
its
el
f
.
T
h
u
s
d
ata
i
s
lo
ca
l.
T
h
u
s
Had
o
o
p
c
o
r
r
elate
s
p
r
ep
r
o
ce
s
s
o
r
s
an
d
s
to
r
ag
e.
2
.
3
.
3
.
F
a
ult
T
o
lera
nce
Had
o
o
p
is
d
esig
n
ed
to
co
p
e
u
p
w
it
h
n
o
d
e
f
ail
u
r
es.
As lar
g
e
n
u
m
b
er
o
f
m
ac
h
i
n
es
ar
e
th
er
e,
a
n
o
d
e
f
ail
u
r
e
is
v
er
y
co
m
m
o
n
p
r
o
b
lem
.
2
.
3
.
4
.
Abs
t
ra
ct
io
n o
f
Co
m
ple
x
it
ies
Had
o
o
p
p
r
o
v
id
es p
r
o
p
er
in
ter
f
ac
es b
et
w
ee
n
co
m
p
o
n
e
n
ts
f
o
r
p
r
o
p
er
w
o
r
k
in
g
.
2
.
3
.
5
.
Da
t
a
pro
t
ec
t
i
o
n a
nd
Co
n
s
is
t
enc
y
Had
o
o
p
h
an
d
les s
y
s
te
m
lev
el
ch
alle
n
g
e
s
as it s
u
p
p
o
r
ts
d
ata
co
n
s
is
ten
c
y
.
2
.
4
.
B
uil
din
g
B
lo
cks
o
f
H
a
do
o
p
As
s
h
o
w
n
in
Fi
g
u
r
e
3
a
s
et
o
f
r
esid
en
t
p
r
o
g
r
a
m
s
i.e
.
d
ae
m
o
n
s
ar
e
r
u
n
n
in
g
i
n
Had
o
o
p
.
T
h
ese
d
ae
m
o
n
s
m
a
y
b
e
r
u
n
n
i
n
g
o
n
t
h
e
s
a
m
e
s
er
v
er
o
r
o
n
t
h
e
d
i
f
f
er
en
t
s
er
v
er
s
i
n
th
e
n
et
w
o
r
k
.
A
ll
t
h
ese
d
ae
m
o
n
s
h
av
e
s
o
m
e
s
p
ec
i
f
ic
f
u
n
ctio
n
ali
t
y
as
s
ig
n
ed
to
th
e
m
.
L
e
t u
s
s
ee
th
ese
d
ae
m
o
n
s
,
Fig
u
r
e
3
.
Had
o
o
p
C
lu
s
ter
T
o
p
o
lo
g
y
2
.
4
.
1
.
Seco
nd
a
ry
Na
m
eNo
de
T
h
e
Seco
n
d
ar
y
Na
m
eNo
d
e
(
SNN)
m
o
n
ito
r
s
t
h
e
s
tate
o
f
t
h
e
cl
u
s
ter
H
DFS.
E
ac
h
cl
u
s
t
er
h
as
o
n
e
SNN
w
h
ic
h
r
es
id
es
o
n
it
s
o
w
n
m
ac
h
i
n
e
al
s
o
.
On
t
h
e
s
a
m
e
s
er
v
er
a
n
y
o
th
er
Data
No
d
e
o
r
T
ask
T
r
ac
k
er
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
6
,
No
.
6
,
Dec
em
b
er
201
6
:
2
9
1
1
–
2
919
2914
d
ae
m
o
n
s
ca
n
n
o
t
b
e
r
u
n
.
Na
m
eNo
d
e
also
p
r
o
v
id
es
s
n
ap
s
h
o
ts
o
f
th
e
HDF
S
m
e
tad
ata
at
r
eg
u
lar
in
ter
v
al
s
t
o
SNN.
2
.
4
.
2
.
Na
m
eNo
de
I
t
is
th
e
Ma
s
ter
n
o
d
e
in
HD
FS
.
I
t
p
r
o
v
id
es
in
s
tr
u
ct
io
n
s
to
s
la
v
e
(
Data
No
d
e)
f
o
r
in
p
u
t
o
u
tp
u
t
task
s
o
f
lo
w
le
v
el.
T
h
e
Na
m
e
No
d
e
k
e
ep
s
tr
ac
k
o
f
f
ile
s
b
r
o
k
en
d
o
wn
in
to
f
ile
b
lo
c
k
s
,
n
o
d
es
s
to
r
in
g
t
h
e
s
e
b
lo
ck
s
a
n
d
t
h
e
o
v
er
all
f
u
n
ctio
n
alit
y
o
f
th
e
d
is
tr
ib
u
ted
f
ile
s
y
s
te
m
.
Na
m
eNo
d
e
is
t
h
e
o
n
l
y
s
i
n
g
l
e
p
o
in
t
o
f
f
a
ilu
r
e
co
m
p
o
n
e
n
t i
n
HD
FS
.
2
.
4
.
3
.
Da
t
a
No
de
Na
m
eNo
d
e
tells
clien
t
th
e
b
lo
ck
ad
d
r
ess
es
in
Data
No
d
es.
T
h
u
s
clien
t
ca
n
d
ir
ec
tl
y
co
m
m
u
n
ica
te
to
Data
No
d
eto
p
r
o
ce
s
s
th
e
lo
ca
l
f
ile
s
i
n
s
id
e
t
h
o
s
e
b
lo
ck
s
.
Fo
r
r
ep
licatio
n
o
f
d
ata
o
n
e
Data
No
d
e
m
a
y
co
m
m
u
n
icate
w
it
h
o
t
h
er
Dat
aNo
d
e
d
ir
ec
tly
.
Data
No
d
es
c
o
n
tin
u
all
y
p
r
o
v
id
es
i
n
f
o
r
m
ati
o
n
to
Na
m
eNo
d
e
r
eg
ar
d
in
g
lo
ca
l
ch
a
n
g
es.
Data
No
d
e
also
r
ec
eiv
es
in
s
tr
u
c
tio
n
s
f
o
r
cr
ea
tio
n
o
r
m
o
v
e
m
e
n
t
o
r
d
e
letio
n
o
f
b
lo
ck
s
f
r
o
m
t
h
e
lo
ca
l d
is
k
.
2
.
4
.
4
.
J
o
bT
ra
ck
er
T
h
e
J
o
b
T
r
ac
k
er
d
eter
m
in
e
s
t
h
e
ex
ec
u
t
io
n
p
lan
.
I
t
d
eter
m
i
n
e
s
f
ile
s
to
p
r
o
ce
s
s
,
n
o
d
e
ass
i
g
n
m
en
ts
f
o
r
d
if
f
er
e
n
t
ta
s
k
s
,
tas
k
s
m
o
n
ito
r
i
n
g
e
tc.
T
h
er
e
is
o
n
l
y
o
n
e
J
o
b
T
r
ac
k
er
d
ae
m
o
n
p
er
Had
o
o
p
clu
s
ter
.
I
t
r
u
n
s
o
n
a
s
er
v
er
as a
m
a
s
ter
n
o
d
e
o
f
th
e
clu
s
ter
.
2
.
4
.
5
.
T
a
s
kT
ra
c
k
er
I
n
d
iv
id
u
a
l
tas
k
s
a
s
s
i
g
n
ed
b
y
J
o
b
T
r
ac
k
er
ar
e
ex
ec
u
ted
b
y
T
ask
T
r
ac
k
er
.
T
h
er
e
is
a
s
in
g
le
T
ask
T
r
ac
k
er
p
e
r
s
lav
e
n
o
d
e.
T
ask
T
r
ac
k
er
m
a
y
h
a
n
d
le
m
u
l
tip
le
task
s
p
ar
allell
y
b
y
u
s
i
n
g
m
u
ltip
le
J
VM
s
.
T
ask
T
r
ac
k
er
co
n
s
tan
tl
y
co
m
m
u
n
icate
s
w
i
th
th
e
J
o
b
T
r
ac
k
er
.
W
ith
i
n
a
s
p
ec
if
ied
a
m
o
u
n
t
o
f
ti
m
e
i
f
t
h
e
T
ask
T
r
ac
k
er
f
ails
to
r
esp
o
n
d
to
J
o
b
T
r
ac
k
er
th
en
it
i
s
ass
u
m
ed
t
h
at
t
h
e
T
ask
T
r
ac
k
er
h
a
s
cr
ash
ed
.
C
o
r
r
esp
o
n
d
in
g
ta
s
k
s
ar
e
r
esu
b
m
itted
to
o
th
er
n
o
d
es i
n
th
e
cl
u
s
ter
.
T
h
e
in
ter
ac
tio
n
b
et
w
ee
n
J
o
b
T
r
ac
k
er
an
d
T
ask
T
r
ac
k
er
is
s
h
o
w
n
b
y
Fi
g
u
r
e
4
.
Fig
u
r
e
4
.
J
o
b
T
r
ac
k
er
an
d
T
ask
T
r
ac
k
er
I
n
ter
ac
tio
n
2
.
5
.
H
a
do
o
p L
i
m
it
a
t
io
n
Had
o
o
p
ca
n
p
er
f
o
r
m
o
n
l
y
b
atch
p
r
o
ce
s
s
i
n
g
a
n
d
s
eq
u
en
t
ial
ac
ce
s
s
.
Seq
u
e
n
tial
ac
ce
s
s
is
ti
m
e
co
n
s
u
m
i
n
g
.
So
a
n
e
w
tec
h
n
iq
u
e
is
n
ee
d
ed
to
g
et
r
id
o
f
th
is
p
r
o
b
lem
.
2
.
6
.
H
a
do
o
p Distr
ibu
t
ed
F
ile
Sy
s
t
e
m
(
H
DF
S)
HDFS
ca
n
s
to
r
e
v
er
y
lar
g
e
f
il
es.
I
t
s
u
p
p
o
r
ts
s
tr
ea
m
in
g
d
ata
ac
ce
s
s
p
atter
n
s
.
HDF
S
r
u
n
s
o
n
clu
s
ter
s
o
n
co
m
m
o
d
it
y
h
ar
d
w
ar
e.
HDF
S h
as
f
o
llo
w
i
n
g
i
m
p
o
r
tan
t c
h
a
r
ac
ter
is
tics
,
a.
Hig
h
l
y
f
a
u
lt
-
to
ler
an
t
b.
Hig
h
th
r
o
u
g
h
p
u
t
c.
Su
p
p
o
r
ts
ap
p
licatio
n
w
ith
m
a
s
s
iv
e
d
ata
s
et
s
d.
Stre
a
m
i
n
g
d
ata
ac
ce
s
s
e.
E
asil
y
b
u
ilt o
n
co
m
m
o
d
it
y
h
ar
d
w
ar
e.
I
n
HDFS
a
f
ile
is
c
h
o
p
p
ed
in
to
6
4
M
B
/1
2
8
MB
ch
u
n
k
s
an
d
th
en
s
to
r
ed
k
n
o
w
n
as
b
lo
ck
s
.
As
s
h
o
w
n
in
Fi
g
u
r
e
5
HDFS
cl
u
s
ter
h
as
t
w
o
t
y
p
e
s
o
f
n
o
d
e
–
Ma
s
ter
(
Na
m
eNo
d
e)
an
d
Sla
v
e
(
Data
No
d
e)
.
Nam
eNo
d
e
m
an
a
g
e
s
t
h
e
n
a
m
e
s
p
ac
e
o
f
t
h
e
f
iles
y
s
te
m
.
I
t
m
a
in
ta
in
s
t
h
e
f
ile
s
y
s
te
m
tr
ee
.
T
h
e
m
e
t
ad
ata
co
n
tain
s
t
h
e
in
f
o
r
m
atio
n
ab
o
u
t
al
l
t
h
e
d
ir
e
cto
r
ies
an
d
f
ile
s
i
n
t
h
e
tr
ee
is
also
s
to
r
ed
.
T
h
is
i
n
f
o
r
m
atio
n
i
s
s
to
r
ed
co
n
s
tan
t
l
y
o
n
th
e
lo
ca
l d
is
k
in
t
h
e
f
o
r
m
o
f
t
w
o
f
ile
s
: t
h
e
n
a
m
e
s
p
ac
e
i
m
a
g
e
an
d
th
e
ed
it lo
g
.
T
h
r
o
u
g
h
th
e
co
m
m
u
n
icatio
n
w
it
h
th
e
Na
m
e
n
o
d
e
an
d
Data
n
o
d
es
a
cl
ien
t
ca
n
g
et
t
h
e
ac
ce
s
s
o
f
t
h
e
f
iles
y
s
te
m
.
T
h
e
u
s
er
co
d
e
is
u
n
a
w
ar
e
ab
o
u
t
w
h
ic
h
Na
m
en
o
d
e
an
d
Data
n
o
d
e
ar
e
f
u
n
ctio
n
.
On
l
y
a
f
ter
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
B
ig
Da
ta
a
n
d
Ma
p
R
e
d
u
ce
C
h
a
llen
g
es,
Op
p
o
r
tu
n
ities
a
n
d
T
r
en
d
s
(
S
a
ch
in
A
r
u
n
Th
a
n
ek
a
r
)
2915
in
s
tr
u
ctio
n
s
f
r
o
m
Na
m
eNo
d
e,
Data
n
o
d
e
s
s
to
r
e
an
d
r
etr
iev
e
b
lo
ck
s
.
A
t
th
e
s
a
m
e
ti
m
e,
t
h
e
y
ar
e
p
r
o
v
id
in
g
s
to
r
ag
e
u
p
d
ates to
Na
m
eNo
d
e.
Fig
u
r
e
5
.
HDFS
A
r
ch
i
tectu
r
e
3.
M
AP
RE
DUCE
Hu
g
e
a
m
o
u
n
t
o
f
d
ata
ca
n
b
e
ea
s
il
y
,
e
f
f
icie
n
tl
y
p
r
o
ce
s
s
ed
b
y
Ma
p
R
ed
u
ce
w
it
h
g
r
ea
t
p
ar
allelis
m
.
Mo
r
eo
v
er
,
th
ese
ap
p
licatio
n
s
ca
n
r
u
n
o
n
cl
u
s
ter
s
o
f
co
m
m
o
d
it
y
h
ar
d
w
ar
e
w
h
ic
h
m
a
k
es i
t
s
u
itab
le
f
o
r
s
ca
l
in
g
.
Ma
p
R
ed
u
ce
i
s
b
ased
o
n
j
av
a.
T
h
e
Ma
p
R
ed
u
ce
a
lg
o
r
it
h
m
c
o
n
tain
s
Ma
p
ta
s
k
an
d
R
ed
u
ce
tas
k
.
T
h
e
g
e
n
er
al
Ma
p
R
ed
u
ce
d
ataf
lo
w
i
s
as
s
h
o
w
n
in
F
ig
u
r
e
6
.
I
n
Ma
p
task
in
d
iv
id
u
al
ele
m
en
ts
ar
e
b
r
o
k
en
d
o
w
n
in
to
tu
p
les
also
k
n
o
w
n
as
k
e
y
/
v
al
u
e
p
air
s
.
R
ed
u
ce
tas
k
f
u
r
th
er
ta
k
e
s
t
h
e
s
e
i
n
ter
m
ed
iate
t
u
p
les
a
s
a
n
i
n
p
u
t.
T
h
en
R
ed
u
ce
task
co
m
b
i
n
es
i
t
in
to
a
s
m
all
er
s
et
o
f
tu
p
les.
R
ed
u
ce
tas
k
ca
n
b
e
s
tar
ted
o
n
l
y
a
f
ter
th
e
co
m
p
let
io
n
o
f
Ma
p
task
[
5
-
8
].
Fig
u
r
e
6
.
T
h
e
Gen
er
al
Ma
p
r
ed
u
ce
Data
f
lo
w
3
.
1
.
M
a
p Re
du
ce
co
re
f
un
ct
io
ns
a.
I
n
p
u
t r
ea
d
er
Div
id
es i
n
p
u
t i
n
to
s
m
all
p
ar
ts
/ b
lo
ck
s
.
T
h
ese
b
lo
ck
s
th
e
n
g
e
t a
s
s
i
g
n
ed
to
a
Ma
p
f
u
n
ctio
n
.
b.
Ma
p
f
u
n
ctio
n
I
n
d
iv
id
u
a
l e
le
m
e
n
t
s
ar
e
b
r
o
k
en
d
o
w
n
in
to
tu
p
les al
s
o
k
n
o
w
n
as k
e
y
/v
al
u
e
p
air
s
.
c.
Sh
u
f
f
le
a
n
d
So
r
t
P
ar
titi
o
n
f
u
n
ctio
n
W
ith
th
e
g
i
v
en
k
e
y
a
n
d
n
u
m
b
er
o
f
r
ed
u
ce
r
s
it f
i
n
d
s
th
e
co
r
r
ec
t r
ed
u
ce
r
.
C
o
m
p
ar
e
f
u
n
c
tio
n
Ma
p
in
ter
m
ed
iate
o
u
tp
u
ts
ar
e
s
o
r
ted
ac
co
r
d
in
g
to
th
is
co
m
p
a
r
e
f
u
n
ctio
n
.
d.
R
ed
u
ce
f
u
n
ctio
n
C
o
m
b
i
n
es i
n
ter
m
ed
iate
t
u
p
les
in
to
a
s
m
aller
s
et
o
f
t
u
p
les a
n
d
g
iv
e
s
it to
o
u
p
u
t.
e.
Ou
tp
u
t
w
r
i
ter
Giv
es
f
ile
o
u
tp
u
t.
L
et
u
s
u
n
d
er
s
ta
n
d
Ma
p
R
ed
u
ce
w
o
r
k
i
n
g
w
it
h
an
e
x
a
m
p
le,
Fil
e1
: "
Hi
Sru
s
h
ti Hi
S
h
r
u
ti"
Fil
e2
: "
B
y
e
Sru
s
h
t
i B
y
e
S
h
r
u
ti
"
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
6
,
No
.
6
,
Dec
em
b
er
201
6
:
2
9
1
1
–
2
919
2916
Nu
m
b
er
o
f
o
cc
u
r
r
en
ce
s
o
f
ea
c
h
w
o
r
d
ac
r
o
s
s
d
if
f
er
en
t f
i
les ar
e
to
b
e
co
u
n
ted
.
T
h
r
ee
o
p
er
atio
n
s
w
i
ll b
e
th
er
e
as f
o
llo
w
s
,
Map
Ma
p
1
Ma
p
2
<
Hi,
1
>
<B
y
e,
1
>
<Sr
u
s
h
ti,
1
>
<Sr
u
s
h
ti,
1
>
<
Hi,
1
>
<B
y
e,
1
>
<Sh
r
u
ti,
1
>
<Sh
r
u
ti,
1
>
C
o
m
b
i
n
e
C
o
m
b
i
n
e
Ma
p
1
C
o
m
b
i
n
e
Ma
p
2
<Sr
u
s
h
ti,
1
>
<Sr
u
s
h
ti,
1
>
<Sh
r
u
ti,
1
>
<Sh
r
u
ti,
1
>
<
Hi,
2
>
<B
y
e,
2
>
R
ed
u
ce
<Sr
u
s
h
ti,
2
>
<Sh
r
u
ti,
2
>
<B
y
e,
2
>
<
Hi,
2
>
3
.
2
.
Nu
m
ber
o
f
M
a
pp
er
s
a
nd
Reducer
s
Am
o
u
n
t
o
f
d
ata
an
d
th
e
b
lo
ck
s
ize
d
ec
id
es
t
h
e
n
u
m
b
er
o
f
Ma
p
s
.
Had
o
o
p
A
P
I
w
it
h
t
h
e
s
etNu
m
Ma
p
T
ask
s
(
i
n
t
)
m
et
h
o
d
p
r
o
v
id
es
th
e
c
u
r
r
en
t
n
u
m
b
er
o
f
m
ap
p
er
s
i
n
t
h
e
s
y
s
te
m
.
A
n
u
m
b
er
s
o
f
R
ed
u
ce
r
s
ar
e
d
ir
ec
tl
y
r
elate
d
to
th
e
Ma
p
p
er
's
i
n
p
u
t.
A
s
p
er
s
p
ec
if
ica
tio
n
i
t
w
il
l
b
e
ex
ec
u
ted
.
Ma
p
R
ed
u
ce
co
m
m
a
n
d
„
-
D
m
ap
r
ed
.
r
e
d
u
ce
‟
ca
n
s
et
t
h
e
n
u
m
b
e
r
o
f
R
ed
u
ce
r
s
at
r
u
n
ti
m
e
as
w
ell.
„
co
n
f
.
s
etNu
m
R
ed
u
ce
T
ask
s
(
in
t)
‟
is
t
h
e
m
et
h
o
d
th
r
o
u
g
h
w
h
ic
h
p
r
o
g
r
a
m
m
er
s
ca
n
s
et
it
w
i
th
co
d
i
n
g
.
3
.
3
.
F
a
ilu
re
H
a
nd
lin
g
in M
a
p Re
du
ce
Ma
ch
i
n
e
f
ail
u
r
e
h
a
n
d
lin
g
is
v
er
y
i
m
p
o
r
tan
t
asp
ec
t
o
f
Ma
p
R
ed
u
ce
as
it
u
s
es
h
u
n
d
r
ed
s
o
r
th
o
u
s
a
n
d
s
o
f
co
m
m
o
d
it
y
m
ac
h
in
e
s
.
T
h
er
e
ar
e
t
w
o
t
y
p
es
o
f
b
asic
f
ail
u
r
es a
s
Ma
s
ter
n
o
d
e
f
ail
u
r
e
o
r
W
o
r
k
er
n
o
d
e
f
ai
lu
r
e.
I
f
Ma
s
ter
n
o
d
e
f
ail
s
,
t
h
en
all
Ma
p
R
ed
u
ce
tas
k
i
s
a
b
o
r
ted
.
T
h
e
w
h
o
le
ta
s
k
is
to
b
e
a
s
s
i
g
n
ed
to
a
n
e
w
Ma
s
ter
n
o
d
e
an
d
ag
ain
i
t h
a
s
to
b
e
r
ed
o
n
e.
Ma
s
ter
co
n
s
ta
n
tl
y
c
h
ec
k
s
t
h
e
w
o
r
k
er
s
tatu
s
i
n
o
r
d
er
to
ch
ec
k
f
a
ilu
r
e.
I
f
w
o
r
k
er
d
o
es
n
o
t
r
esp
o
n
d
to
m
aster
i
n
ti
m
e,
t
h
en
it
is
m
ar
k
ed
as
a
f
ailed
.
I
f
m
ap
tas
k
w
o
r
k
er
f
ail
s
,
t
h
en
w
i
th
n
o
co
n
s
id
er
atio
n
o
f
an
y
m
ap
task
s
s
tate
i
.
e.
w
h
eth
er
it
is
i
n
p
r
o
g
r
ess
/
co
m
p
leted
etc.
w
o
r
k
er
s
ar
e
r
eset
to
th
eir
in
itial
id
le
s
tate.
T
h
e
task
th
en
w
ill
b
e
a
s
s
i
g
n
ed
to
o
th
er
id
le
w
o
r
k
er
.
I
f
r
ed
u
ce
ta
s
k
f
a
ils
a
n
id
le
w
o
r
k
er
i
s
ch
o
s
en
f
o
r
r
ea
s
s
ig
n
m
e
n
t
o
f
th
e
tas
k
ir
r
esp
ec
tiv
e
o
f
an
y
tas
k
s
tate.
3
.
4
.
Da
t
a
Sto
ra
g
e
a
nd
Replica
t
io
n in M
a
p Re
du
ce
I
n
Ma
p
R
ed
u
ce
co
m
p
leted
r
ed
u
ce
task
s
o
u
tp
u
t
is
s
to
r
ed
in
g
lo
b
al
f
ile
s
y
s
te
m
.
T
h
u
s
r
e
-
e
x
ec
u
tio
n
o
f
co
m
p
leted
r
ed
u
ce
tas
k
s
is
n
o
t
r
eq
u
ir
ed
.
L
o
ca
l
d
is
k
s
ar
e
u
s
ed
to
s
to
r
e
th
e
r
esu
lts
o
f
m
a
p
task
s
.
I
n
ca
s
e
o
f
f
ail
u
r
e
it c
an
b
e
r
e
-
e
x
ec
u
ted
f
r
o
m
lo
ca
l d
is
k
s
.
3
.
5
.
M
a
pReduce
Cha
lleng
es
Fo
llo
w
i
n
g
ar
e
th
e
li
m
itatio
n
s
o
f
Ma
p
R
ed
u
ce
id
en
t
if
ied
[
9
-
1
3
],
1)
No
r
ed
u
ce
ca
n
b
eg
in
u
n
til all
m
ap
s
ar
e
co
m
p
lete
2)
Ma
p
r
ed
u
ce
r
ed
u
ce
task
s
tar
ts
o
n
l
y
a
f
ter
f
i
n
i
s
h
i
n
g
o
f
t
h
e
all
m
ap
tas
k
s
.
3)
Ma
s
ter
m
u
s
t c
o
m
m
u
n
icate
lo
c
atio
n
s
o
f
i
n
ter
m
ed
iate
f
i
les.
4)
Af
ter
e
v
er
y
m
ap
tas
k
lo
t
o
f
in
ter
m
ed
iate
d
ata
i
s
g
e
n
er
ated
a
n
d
it
is
to
b
e
s
to
r
ed
a
n
d
also
t
o
b
e
in
f
o
r
m
ed
to
o
th
er
s
.
5)
T
ask
s
s
ch
ed
u
led
b
ased
o
n
lo
ca
tio
n
o
f
d
ata.
6)
L
o
t o
f
co
m
p
u
tatio
n
i
s
r
eq
u
ir
ed
to
p
r
o
v
id
e
d
ata
l
o
ca
tio
n
an
d
th
en
to
allo
ca
te
r
eso
u
r
ce
s
o
n
th
at
lo
ca
tio
n
.
7)
B
ef
o
r
e
r
ed
u
ce
f
in
i
s
h
e
s
if
m
ap
w
o
r
k
er
f
ails
,
ta
s
k
m
u
s
t b
e
co
m
p
letel
y
r
er
u
n
8)
I
f
m
a
s
ter
f
ails
th
e
n
t
h
e
w
h
o
le
Ma
p
R
ed
u
ce
tas
k
g
et
ab
o
r
ted
,
an
d
it
h
as
to
b
e
r
ed
o
n
e
af
ter
ass
i
g
n
in
g
n
e
w
m
aster
n
o
d
e.
9)
I
n
ter
m
ed
iate
d
ata
10)
L
o
ts
o
f
i
n
ter
m
ed
iate
d
ata
is
g
e
n
er
ated
.
Af
ter
u
s
e
it is
d
estro
y
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
B
ig
Da
ta
a
n
d
Ma
p
R
e
d
u
ce
C
h
a
llen
g
es,
Op
p
o
r
tu
n
ities
a
n
d
T
r
en
d
s
(
S
a
ch
in
A
r
u
n
Th
a
n
ek
a
r
)
2917
11)
Hete
r
o
g
en
eo
u
s
d
ata
Data
is
co
m
i
n
g
f
r
o
m
d
i
f
f
er
en
t
s
o
u
r
ce
s
an
d
d
i
f
f
er
e
n
t f
o
r
m
ats.
4.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
I
n
th
i
s
s
ec
tio
n
w
e
w
ill d
is
c
u
s
s
th
e
w
o
r
k
d
o
n
e
b
y
d
if
f
er
e
n
t r
es
ea
r
ch
er
s
o
n
d
if
f
er
en
t c
h
alle
n
g
es,
4
.
1
.
Cha
lleng
e
I
:
No
Reduce
c
a
n
B
eg
in Unt
il a
ll
M
a
ps
a
re
C
o
m
p
let
e
I
n
Ma
p
R
ed
u
ce
,
a
r
ed
u
ce
r
ca
n
n
o
t
s
tar
t
it
s
p
r
o
ce
s
s
in
g
till
t
h
e
co
m
p
letio
n
o
f
all
t
h
e
m
ap
p
in
g
tas
k
s
.
T
h
e
m
aj
o
r
d
r
aw
b
ac
k
o
f
t
h
is
te
ch
n
iq
u
e
is
t
h
at
r
ed
u
ce
r
s
h
a
v
e
to
w
ait
u
n
n
ec
es
s
ar
il
y
.
I
n
o
th
er
s
en
s
e
it
is
n
o
t
a
n
ef
f
ec
tiv
e
a
n
d
ef
f
icie
n
t
u
s
e
o
f
r
eso
u
r
ce
s
.
A
b
d
e
l
R
a
h
m
a
n
E
ls
a
y
ed
et
al.
,
[
11
]
d
o
n
e
in
v
e
s
ti
g
atio
n
o
n
Ma
p
R
ed
u
ce
r
esear
ch
tr
en
d
s
,
an
d
cu
r
r
en
t
r
esear
ch
ef
f
o
r
ts
.
T
h
ey
s
u
g
g
e
s
ted
th
at
n
e
w
a
lg
o
r
it
h
m
s
ca
n
b
e
d
ev
elo
p
ed
o
r
f
r
a
m
e
w
o
r
k
c
an
b
e
m
o
d
i
f
ied
in
o
r
d
er
t
o
im
p
r
o
v
e
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
Ma
p
R
ed
u
ce
.
Dh
o
le
P
o
o
n
a
m
et
al.
,
[
14
]
p
r
o
p
o
s
ed
a
s
o
lu
tio
n
f
o
r
th
i
s
p
r
o
b
l
e
m
.
I
n
t
h
eir
w
o
r
k
p
ip
eli
n
ed
m
ap
r
ed
u
ce
m
ap
p
er
ca
n
s
e
n
d
its
o
u
tp
u
t d
ir
ec
tl
y
to
r
ed
u
ce
r
as
a
n
i
n
p
u
t.
T
h
u
s
co
m
p
letio
n
ti
m
e,
s
y
s
te
m
u
tili
za
tio
n
f
o
r
b
atc
h
j
o
b
s
ar
e
im
p
r
o
v
ed
.
4
.
2
.
Cha
lleng
e
I
I
:
M
a
s
t
er
m
us
t
C
o
mm
un
ica
t
e
L
o
c
a
t
io
ns
o
f
I
n
t
er
m
ed
ia
t
e
F
iles
Dian
a
Mo
i
s
e
et
a
l.
,
[
5
]
p
r
o
p
o
s
ed
th
e
u
s
e
o
f
B
lo
b
Seer
d
ata
m
a
n
ag
e
m
e
n
t
s
er
v
ice
f
o
r
s
to
r
in
g
in
ter
m
ed
iate
r
esu
lts
.
I
t
i
s
a
f
a
u
lt
-
to
ler
an
t,
co
n
cu
r
r
en
c
y
o
p
ti
m
ized
d
ata
s
to
r
ag
e
la
y
er
.
T
h
u
s
it
i
s
an
a
lter
n
ati
v
e
f
o
r
lo
ca
l
s
to
r
ag
e
o
f
th
e
m
ap
p
er
s
.
T
h
u
s
th
e
i
n
ter
m
ed
iate
d
ata
ca
n
b
e
m
a
in
ta
in
ed
s
ep
ar
atel
y
an
d
later
o
n
it
ca
n
b
e
u
s
ed
ag
ai
n
.
4
.
3
.
Cha
lleng
e
I
I
I
:
T
a
s
ks
Sche
du
led B
a
s
e
d o
n
Lo
ca
t
io
n o
f
Da
ta
Nila
m
Kad
ale
et
al.
,
[1
5
]
s
tated
th
at
in
Ma
p
r
ed
u
ce
f
r
a
m
e
w
o
r
k
d
if
f
er
en
t
ta
s
k
s
ch
ed
u
li
n
g
m
e
th
o
d
s
ar
e
u
s
ed
to
s
ch
ed
u
le
th
e
ta
s
k
.
Su
r
v
e
y
o
f
v
ar
io
u
s
tas
k
s
c
h
ed
u
li
n
g
m
eth
o
d
s
o
f
Ma
p
r
ed
u
ce
f
r
a
m
e
w
o
r
k
i
s
d
o
n
e.
J
u
n
L
i
u
e
t
al.
,
[1
6
]
in
tr
o
d
u
ce
d
d
y
n
a
m
ic
p
r
io
r
it
y
s
c
h
ed
u
lin
g
a
n
d
r
ea
l
-
ti
m
e
p
r
ed
ictio
n
m
o
d
el.
T
h
e
y
in
tr
o
d
u
ce
d
th
e
d
ata
lo
ca
lit
y
alg
o
r
ith
m
w
h
ic
h
h
as
m
i
n
i
m
u
m
co
s
t
an
d
also
co
n
s
id
er
s
a
w
ei
g
h
t.
R
ea
l
-
ti
m
e
p
r
ed
ictio
n
m
o
d
el
i
s
u
s
ed
to
b
etter
s
er
v
e
d
i
f
f
er
en
t
s
ize
j
o
b
s
.
T
h
e
y
al
s
o
s
tated
th
at
r
es
o
u
r
ce
u
ti
lizatio
n
o
f
u
n
e
x
ec
u
ted
tas
k
s
ca
n
b
e
p
r
ed
icted
b
y
ca
lc
u
lati
n
g
t
h
e
r
u
n
n
in
g
tas
k
s
.
B
o
Z
h
an
g
et
al.
,
[1
7
]
p
r
o
p
o
s
e
d
a
f
ee
d
b
ac
k
co
n
tr
o
l
lo
o
p
b
ased
ap
p
r
o
ac
h
.
B
ased
o
n
th
e
cu
r
r
en
t
s
tate
o
f
th
e
clu
s
ter
t
h
e
y
d
y
n
a
m
ical
l
y
ad
j
u
s
ted
t
h
e
Had
o
o
p
r
eso
u
r
ce
m
an
a
g
er
co
n
f
i
g
u
r
atio
n
.
T
h
e
y
i
m
p
r
o
v
ed
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
s
y
s
te
m
b
y
3
0
% a
s
co
m
p
ar
ed
to
d
ef
au
lt H
ad
o
o
p
s
etu
p
.
Mu
h
a
m
m
ad
I
d
r
is
e
t
al.
,
[
18
]
p
r
o
v
id
ed
g
o
o
d
s
u
r
v
e
y
o
n
Had
o
o
p
Ma
p
R
ed
u
ce
s
c
h
ed
u
li
n
g
an
d
en
h
a
n
ce
m
en
ts
d
o
n
e
s
o
f
ar
.
T
h
e
y
a
ls
o
d
is
c
u
s
s
ed
o
p
en
i
s
s
u
es,
ch
alle
n
g
es
r
elate
d
to
th
e
s
ch
ed
u
lin
g
d
o
n
e
i
n
Ma
p
R
ed
u
ce
.
4
.
4
.
Cha
lleng
e
I
V:
B
ef
o
re
Reduce
F
ini
s
he
s
if
M
a
p Wo
rk
er
F
a
ils
,
T
a
s
k
M
us
t
be
Co
m
plet
e
ly
Rer
un
I
n
o
r
d
er
to
s
o
lv
e
th
is
p
r
o
b
le
m
,
th
e
s
a
m
e
tas
k
ca
n
b
e
ex
ec
u
t
ed
o
n
d
if
f
er
e
n
t
n
o
d
es.
T
h
e
n
o
d
e
w
h
ich
f
i
n
is
h
es e
x
ec
u
tio
n
f
ir
s
t
g
iv
e
s
o
u
tp
u
t.
T
h
en
s
i
m
p
l
y
w
e
ca
n
ab
o
r
t a
ll o
th
er
ex
ec
u
tio
n
s
[
12
]
.
4
.
5
.
Cha
lleng
e
V:
I
nte
r
m
ed
ia
t
e
D
a
t
a
Yax
io
n
g
Z
h
ao
et
al.
,
[1
9
]
p
r
o
p
o
s
ed
a
n
o
v
el
Dac
h
e
(
Data
Aw
ar
e
C
ac
h
e)
tech
n
iq
u
e.
C
ac
h
e
m
a
n
a
g
er
g
ets
i
n
ter
m
ed
iate
r
es
u
lts
f
r
o
m
d
if
f
er
e
n
t
tas
k
s
.
B
ef
o
r
e
ex
ec
u
ti
n
g
a
n
y
tas
k
,
a
task
q
u
er
ies
t
h
e
ca
ch
e
m
a
n
ag
er
.
I
f
it
is
av
ailab
le
i
n
ca
ch
e
t
h
en
s
a
m
e
is
u
s
ed
,
if
n
o
t
th
e
n
o
n
l
y
n
e
w
co
m
p
u
ti
n
g
is
d
o
n
e.
T
h
ey
d
e
s
ig
n
ed
a
n
e
w
ca
ch
e
r
eq
u
est
an
d
r
ep
l
y
p
r
o
to
co
l,
ca
ch
e
d
escr
ip
tio
n
s
ch
em
e.
T
h
r
o
u
g
h
t
h
eir
r
esu
l
ts
th
e
y
h
a
v
e
s
h
o
w
n
th
e
s
ig
n
i
f
ica
n
t i
m
p
r
o
v
e
m
e
n
t i
n
th
e
co
m
p
letio
n
ti
m
e
o
f
Ma
p
R
ed
u
ce
j
o
b
s
.
R
.
Ud
en
d
r
an
et
al.
,
[
20
]
d
o
n
e
r
ev
ie
w
o
n
th
e
d
ata
-
a
w
ar
e
ca
ch
e
(
Dac
h
e)
f
o
r
b
ig
d
ata
a
p
p
licatio
n
s
.
T
h
ey
also
s
tated
t
h
at
b
etter
l
i
f
e
-
ti
m
e
m
an
a
g
e
m
e
n
t
o
f
ca
c
h
e
is
r
eq
u
ir
ed
if
it
r
eq
u
ir
e
s
h
u
g
e
a
m
o
u
n
t
o
f
ca
c
h
e.
Dian
a
Mo
is
e
et
al.
,
[
5
]
in
th
ei
r
p
ap
e
r
f
o
cu
s
ed
o
n
in
ter
m
ed
i
ate
d
ata
g
en
er
ated
in
m
ap
r
ed
u
ce
p
r
o
ce
s
s
.
T
h
ey
p
r
o
p
o
s
ed
a
n
e
w
s
to
r
ag
e
m
ec
h
an
is
m
f
o
r
in
ter
m
ed
iate
d
ata
o
n
th
e
B
lo
b
Seer
d
ata
m
a
n
ag
e
m
en
t
s
er
v
ice.
Fail
u
r
e
h
an
d
li
n
g
,
m
i
n
i
m
u
m
e
x
ec
u
t
io
n
ti
m
e,
co
n
cu
r
r
en
c
y
co
n
tr
o
l
etc
.
ar
e
m
an
a
g
ed
p
r
o
p
er
ly
.
T
h
u
s
r
ed
u
ce
d
th
e
lo
ca
l
s
to
r
ag
e
d
ep
en
d
en
c
y
.
T
h
u
s
r
eu
s
e
o
f
in
ter
m
ed
iate
d
ata
is
p
o
s
s
ib
le.
Mr
u
d
u
la
Var
ad
e
et
al.
,
[2
1
]
g
iv
en
a
g
o
o
d
co
m
p
ar
ativ
e
s
tu
d
y
o
f
a
m
etad
ata
m
a
n
ag
e
m
e
n
t
s
ch
e
m
es.
T
o
m
ai
n
tai
n
r
eliab
ilit
y
,
m
etad
ata
is
r
ep
licated
in
d
if
f
er
en
t
Na
m
eNo
d
es.
L
o
g
r
ep
licatio
n
tech
n
o
lo
g
y
i
s
u
s
ed
f
o
r
r
ep
licatio
n
.
T
o
m
ai
n
tai
n
r
e
p
licatio
n
co
n
s
i
s
te
n
c
y
P
ax
o
s
al
g
o
r
ith
m
is
u
s
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
6
,
No
.
6
,
Dec
em
b
er
201
6
:
2
9
1
1
–
2
919
2918
4
.
6
.
Cha
lleng
e
VI:
H
et
er
o
g
eneo
us
Da
t
a
J
u
n
Q
u
et
a
l.
,
[2
2
]
p
r
o
p
o
s
ed
a
n
e
w
f
r
a
m
e
w
o
r
k
ca
lled
a
s
Ma
p
-
R
ed
u
ce
-
Me
r
g
e.
W
eb
h
et
er
o
g
en
eo
u
s
d
ata
p
r
o
ce
s
s
in
g
is
e
f
f
icie
n
tl
y
d
o
n
e.
T
h
ey
d
o
n
e
t
h
eir
ex
p
er
i
m
e
n
ts
o
n
f
ea
t
u
r
es o
f
w
eb
d
ata.
Nen
a
v
ath
Sri
n
i
v
as
Nai
k
et
al.
,
[
2
3
]
p
r
o
p
o
s
ed
Ma
p
R
ed
u
ce
R
ein
f
o
r
ce
m
e
n
t
L
ea
r
n
i
n
g
s
c
h
e
d
u
ler
.
T
h
is
s
ch
ed
u
ler
s
u
g
g
est
s
th
e
r
e
-
ex
ec
u
tio
n
o
f
s
lo
w
er
tas
k
s
to
o
th
er
av
ailab
le
n
o
d
es b
y
o
b
s
er
v
i
n
g
t
h
e
s
y
s
te
m
s
ta
te
an
d
task
ex
ec
u
tio
n
.
T
h
u
s
f
aster
e
x
ec
u
tio
n
o
f
th
e
tas
k
ca
n
b
e
d
o
n
e.
No
p
r
io
r
k
n
o
w
led
g
e
o
f
th
e
s
y
s
te
m
is
r
eq
u
ir
ed
.
T
h
u
s
o
v
er
all
j
o
b
co
m
p
letio
n
ti
m
e
is
s
ig
n
i
f
ica
n
tl
y
m
in
i
m
ized
.
5.
CO
NCLU
SI
O
N
B
ig
d
ata
is
in
cr
ea
s
in
g
tr
e
m
e
n
d
o
u
s
l
y
d
a
y
b
y
d
a
y
w
h
ich
g
a
v
e
r
is
e
to
n
e
w
d
if
f
ic
u
ltie
s
an
d
ch
a
llen
g
es
a
s
w
e
h
a
v
e
to
s
to
r
e,
p
r
o
ce
s
s
,
an
aly
ze
,
m
o
d
if
y
s
u
c
h
a
h
u
g
e
a
m
o
u
n
t
o
f
d
ata.
E
x
is
ti
n
g
d
atab
ases
,
to
o
ls
ar
e
n
o
t
g
o
o
d
en
o
u
g
h
to
h
a
n
d
le
t
h
is
is
s
u
e.
I
n
o
u
r
p
ap
er
w
e
h
a
v
e
p
r
o
v
id
ed
o
v
er
v
ie
w
o
f
t
h
e
b
ig
d
ata,
its
c
h
alle
n
g
e
s
w
it
h
r
esp
ec
t to
Ma
p
r
ed
u
ce
.
Ma
n
y
ef
f
o
r
t
s
tak
e
n
to
r
ed
u
ce
th
o
s
e
c
h
alle
n
g
e
s
ar
e
also
d
is
cu
s
s
ed
.
T
h
u
s
b
etter
p
lan
n
i
n
g
o
f
B
ig
Data
p
r
o
j
ec
ts
ca
n
b
e
d
o
n
e.
Fo
r
r
esear
ch
er
s
‟
o
p
p
o
r
tu
n
i
ties
f
o
r
f
u
t
u
r
e
r
esear
ch
ca
n
b
e
id
en
ti
f
ied
.
RE
F
E
R
E
NC
E
S
[1
]
Ha
sh
e
m
I
.
A
.
T
.
,
e
t
a
l.
,
“
T
h
e
rise
o
f
b
ig
d
a
ta
o
n
c
lo
u
d
c
o
m
p
u
ti
n
g
:
Re
v
ie
w
a
n
d
o
p
e
n
re
se
a
rc
h
issu
e
s
,
”
El
se
v
ier
In
fo
rm
a
t
io
n
S
y
ste
ms
,
v
o
l.
4
7
,
p
p
.
9
8
–
1
1
5
,
2
0
1
5
.
[2
]
W
a
n
g
L
.
a
n
d
A
lex
a
n
d
e
r
C
.
A
.
,
“
Big
Da
ta:
In
f
ra
stru
c
tu
re
,
tec
h
n
o
l
o
g
y
p
ro
g
re
ss
a
n
d
c
h
a
ll
e
n
g
e
s
,
”
J
o
u
rn
a
l
o
f
D
a
t
a
M
a
n
a
g
e
me
n
t
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l/
issu
e
:
2
(
1
),
p
p
.
0
0
1
-
0
0
6
,
2
0
1
5
.
[3
]
W
.
F
a
n
a
n
d
A
.
Bi
f
e
t,
“
M
in
in
g
Big
Da
ta:
Cu
rre
n
t
S
tatu
s,
a
n
d
F
o
re
c
a
st
to
th
e
F
u
t
u
re
,
”
S
IGKD
D
Exp
lo
ra
t
io
n
s
,
v
o
l/
issu
e
:
1
4
(2
),
2
0
1
2
.
[4
]
P
.
Bh
a
ti
a
a
n
d
S
.
G
u
p
ta,
“
Co
rre
l
a
ted
A
p
p
ra
isa
l
o
f
Big
Da
ta,
Ha
d
o
o
p
a
n
d
M
a
p
Re
d
u
c
e
,
”
Ad
v
a
n
c
e
s
in
C
o
mp
u
ter
S
c
ien
c
e
:
An
I
n
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
,
v
o
l/
issu
e
:
4
(4
),
p
p
.
1
6
,
2
0
1
5
.
[5
]
D
.
M
o
ise
,
e
t
a
l.
,
“
Op
ti
m
izin
g
In
term
e
d
iate
Da
ta
M
a
n
a
g
e
m
e
n
t
in
M
a
p
Re
d
u
c
e
Co
m
p
u
tatio
n
s
,
”
1
st
In
ter
n
a
ti
o
n
a
l
W
o
rk
sh
o
p
o
n
Cl
o
u
d
C
o
mp
u
ti
n
g
P
la
tf
o
rm
s
,
2
0
1
1
.
[6
]
S
.
A
g
a
r
wa
l
a
n
d
Z
.
K
h
a
n
a
m
,
“
M
a
p
Re
d
u
c
e
:
A
S
u
rv
e
y
P
a
p
e
r
o
n
Re
c
e
n
t
Ex
p
a
n
sio
n
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Ad
v
a
n
c
e
d
Co
mp
u
ter
S
c
ien
c
e
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
/i
ss
u
e
:
6
(8
),
2
0
1
5
.
[7
]
N
.
M
a
ll
e
sw
a
ri
T
.
Y
.
J
.
a
n
d
V
a
d
iv
u
G
.
,
“
M
a
p
Re
d
u
c
e
:
A
Tec
h
n
ica
l
Re
v
ie
w
,
”
In
d
ia
n
J
o
u
r
n
a
l
o
f
sc
ien
c
e
a
n
d
te
c
h
n
o
l
o
g
y
,
v
o
l/
issu
e
:
9
(1
),
p
p
.
1
-
6
,
2
0
1
6
.
[8
]
K
.
A
.
A
l
m
o
h
se
n
a
n
d
H
.
Al
-
Jo
b
o
ri,
“
Re
c
o
m
m
e
n
d
e
r
S
y
ste
m
s
in
L
ig
h
t
o
f
Big
D
a
ta
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
),
v
ol
/i
ss
u
e
:
5
(
6
)
,
p
p
.
1
5
5
3
~
1
5
6
3
,
2
0
1
5
.
[9
]
D
.
Zh
a
n
g
,
“
In
c
o
n
siste
n
c
ies
i
n
Big
Da
t
a
,
”
1
2
th
IEE
E
In
t.
Co
n
f.
o
n
Co
g
n
it
ive
In
f
o
rm
a
ti
c
s
&
Co
g
n
it
i
v
e
Co
mp
u
ti
n
g
,
2
0
1
3
.
[1
0
]
H
.
Ba
g
h
e
ri
a
n
d
A
.
A
.
S
h
a
lt
o
o
k
i,
“
Big
Da
ta:
Ch
a
ll
e
n
g
e
s,
Op
p
o
rt
u
n
it
ies
a
n
d
Cl
o
u
d
Ba
se
d
S
o
lu
ti
o
n
s
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
t
e
r E
n
g
i
n
e
e
rin
g
(
IJ
ECE
),
v
ol
/i
ss
u
e
:
5
(
2
)
,
p
p
.
3
4
0
~
3
4
3
,
2
0
1
5
.
[1
1
]
A
.
R.
El
sa
y
e
d
,
e
t
a
l.
,
“
M
a
p
Re
d
u
c
e
:
S
tate
-
of
-
th
e
-
A
rt
a
n
d
Re
se
a
rc
h
Dire
c
ti
o
n
s
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Co
mp
u
ter
a
n
d
El
e
c
trica
l
E
n
g
in
e
e
rin
g
,
v
o
l/
is
su
e
:
6
(1
)
,
2
0
1
4
.
[1
2
]
K.
G
ro
li
n
g
e
r,
e
t
a
l.
,
“
Ch
a
ll
e
n
g
e
s
f
o
r
M
a
p
Re
d
u
c
e
in
Big
Da
ta
,
”
IEE
E
1
0
th
2
0
1
4
W
o
rld
C
o
n
g
re
ss
o
n
S
e
rv
ice
s
(
S
ER
VICE
S
2
0
1
4
)
A
la
sk
a
U
S
A
,
J
u
ly
2
0
1
4
.
[1
3
]
V
.
A
.
Ay
m
a
,
e
t
a
l.
,
“
Clas
sif
ica
ti
o
n
A
lg
o
rit
h
m
s
f
o
r
b
ig
d
a
ta an
a
ly
sis
,
A
m
a
p
re
d
u
c
e
a
p
p
ro
a
c
h
,
”
T
h
e
In
ter
n
a
t
io
n
a
l
Arc
h
ive
s
o
f
t
h
e
Ph
o
t
o
g
ra
mm
e
tr
y
,
Rem
o
te
S
e
n
sin
g
a
n
d
S
p
a
t
ia
l
In
f
o
rm
a
ti
o
n
S
c
ien
c
e
s,
XL
-
3
/W
2
,
J
o
in
t
IS
PR
S
c
o
n
fer
e
n
c
e
Ge
rm
a
n
y
2
0
1
5
.
[1
4
]
D
.
P
o
o
n
a
m
B
.
a
n
d
G
.
Ba
isa
L
.
,
“
S
u
rv
e
y
P
a
p
e
r
o
n
T
ra
d
it
i
o
n
a
l
Ha
d
o
o
p
a
n
d
P
ip
e
li
n
e
d
M
a
p
Re
d
u
c
e
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Co
m
p
u
t
a
ti
o
n
a
l
E
n
g
i
n
e
e
rin
g
Res
e
a
rc
h
,
v
o
l/
issu
e
:
0
3
(
1
2
),
2
0
1
3
.
[1
5
]
N
.
Ka
d
a
le
a
n
d
U.
A
.
M
a
n
d
e
,
“
S
u
rv
e
y
o
f
T
a
s
k
S
c
h
e
d
u
li
n
g
M
e
t
h
o
d
f
o
r
M
a
p
Re
d
u
c
e
F
ra
m
e
w
o
r
k
in
Ha
d
o
o
p
,
”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
A
p
p
li
e
d
I
n
fo
rm
a
t
io
n
S
y
ste
ms
(
IJ
AIS
)
NCIP
ET
,
2
0
1
3
.
[1
6
]
J
.
L
iu
,
e
t
a
l.
,
“
A
n
Eff
icie
n
t
Jo
b
S
c
h
e
d
u
li
n
g
f
o
r
M
a
p
Re
d
u
c
e
Cl
u
ste
rs
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
F
u
tu
re
Ge
n
e
ra
ti
o
n
Co
mm
u
n
ica
ti
o
n
a
n
d
Ne
two
rk
in
g
,
v
o
l/
issu
e
:
8
(
2
),
p
p
.
3
9
1
-
3
9
8
,
2
0
1
5
.
[1
7
]
B.
Zh
a
n
g
,
e
t
a
l
.
,
“
S
e
lf
-
c
o
n
f
ig
u
ra
ti
o
n
o
f
th
e
Nu
m
b
e
r
o
f
c
o
n
c
u
rre
n
tl
y
Ru
n
n
in
g
M
a
p
Re
d
u
c
e
Jo
b
s
in
a
Ha
d
o
o
p
Clu
ste
r
,
”
ICAC
2
0
1
5
,
p
p
.
1
4
9
-
1
5
0
,
2
0
1
5
.
[1
8
]
M
.
Id
ris
,
e
t
a
l
.
,
“
Co
n
tex
t
-
a
w
a
r
e
s
c
h
e
d
u
li
n
g
in
M
a
p
Re
d
u
c
e
:
a
c
o
m
p
a
c
t
re
v
ie
w
,
”
Co
n
c
u
rr
e
n
c
y
a
n
d
Co
mp
u
t
a
ti
o
n
:
Pra
c
ti
c
e
a
n
d
Exp
e
rie
n
c
e
,
v
o
l/
issu
e
:
2
7
(
1
7
),
p
p
.
5
3
3
2
–
5
3
4
9
.
[1
9
]
Y
.
Zh
a
o
,
e
t
a
l.
,
“
Da
c
h
e
:
A
Da
ta
Aw
a
re
Ca
c
h
in
g
f
o
r
Big
-
Da
ta
A
p
p
li
c
a
ti
o
n
s
Us
in
g
th
e
M
a
p
Re
d
u
c
e
F
ra
m
e
w
o
rk
,
”
T
S
INGH
UA
S
CIENC
E
AND
T
EC
HNO
L
OG
Y
,
v
o
l/
issu
e
:
1
9
(
1
),
p
p
.
3
9
-
5
0
,
2
0
1
4
.
[2
0
]
R.
Ud
e
n
d
ra
n
,
e
t
a
l.
,
“
Re
v
ie
w
P
a
p
e
r
o
n
Da
ta
-
a
w
a
re
Ca
c
h
in
g
f
o
r
Big
Da
ta
A
p
p
li
c
a
ti
o
n
s
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Ad
v
a
n
c
e
d
Res
e
a
rc
h
in
Co
mp
u
ter
S
c
ien
c
e
a
n
d
S
o
ft
w
a
re
En
g
in
e
e
rin
g
,
v
o
l/
iss
u
e
:
5
(
3
),
2
0
1
5
.
[2
1
]
M
.
V
a
ra
d
e
a
n
d
V
.
Je
th
a
n
i
,
“
Distrib
u
te
d
m
e
ta
d
a
ta
m
a
n
a
g
e
m
e
n
t
sc
h
e
m
e
in
HD
F
S
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
Ad
v
a
n
c
e
d
Co
mp
u
ter
S
c
ien
c
e
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
/i
ss
u
e
:
6
(8
),
2
0
1
5
.
[2
2
]
J
.
Qu
,
e
t
a
l.
,
“
T
h
e
Op
ti
m
iza
ti
o
n
a
n
d
Im
p
ro
v
e
m
e
n
t
o
f
M
a
p
Re
d
u
c
e
in
W
e
b
Da
ta
M
in
in
g
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Fu
tu
re
Ge
n
e
ra
ti
o
n
Co
mm
u
n
ica
t
i
o
n
a
n
d
Ne
two
rk
in
g
,
v
o
l/
issu
e
:
8
(2
),
p
p
.
3
9
1
-
3
9
8
,
2
0
1
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
B
ig
Da
ta
a
n
d
Ma
p
R
e
d
u
ce
C
h
a
llen
g
es,
Op
p
o
r
tu
n
ities
a
n
d
T
r
en
d
s
(
S
a
ch
in
A
r
u
n
Th
a
n
ek
a
r
)
2919
[2
3
]
N
.
S
.
Na
ik
,
e
t
a
l.
,
“
P
e
rf
o
r
m
a
n
c
e
I
m
p
r
o
v
e
m
e
n
t
o
f
M
a
p
Re
d
u
c
e
F
ra
m
e
w
o
rk
in
H
e
tero
g
e
n
e
o
u
s
Co
n
tex
t
u
sin
g
Re
in
f
o
rc
e
m
e
n
t
Lea
rn
in
g
,
”
El
se
v
ie
r IS
BCC’1
5
,
2
0
1
5
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
S
a
c
h
i
n
Ar
u
n
T
h
a
n
e
k
a
r
,
re
c
e
iv
e
d
h
is
B.
E
(Co
m
p
u
ter)
a
n
d
M
.
E.
(Co
m
p
u
ter).
d
e
g
re
e
s
f
ro
m
P
u
n
e
Un
iv
e
rsity
,
In
d
ia,
in
2
0
0
5
a
n
d
2
0
1
3
re
sp
e
c
ti
v
e
ly
.
Cu
rre
n
tl
y
h
e
is
a
P
h
.
D.
sc
h
o
lar
in
KL
Un
iv
e
rsit
y
,
A
n
d
h
ra
P
ra
d
e
sh
,
In
d
i
a
.
His
c
u
rre
n
t
in
tere
sts
in
c
lu
d
e
b
ig
d
a
ta,
in
f
o
r
m
a
ti
o
n
se
c
u
rit
y
,
d
a
tab
a
se
s,
so
f
t
w
a
re
tes
ti
n
g
.
Dr
.
K
.
S
u
b
r
a
h
m
a
n
y
a
m
is
a
p
ro
f
e
ss
o
r
in
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
E
n
g
in
e
e
rin
g
d
e
p
a
rtm
e
n
t
o
f
KL
Un
iv
e
rsit
y
,
A
n
d
h
ra
P
ra
d
e
sh
.
His
c
u
rre
n
t
i
n
tere
sts in
c
lu
d
e
so
f
tw
a
r
e
e
n
g
in
e
e
rin
g
,
b
ig
d
a
ta.
Dr
.
A.
B
.
B
a
g
w
a
n
is
w
o
rk
in
g
a
s
a
P
ro
f
e
ss
o
r
in
Co
m
p
u
ter
En
g
in
e
e
rin
g
d
e
p
a
rtm
e
n
t
o
f
Ra
jar
sh
i
S
h
a
h
u
C
o
ll
e
g
e
o
f
En
g
in
e
e
rin
g
,
P
u
n
e
.
His
c
u
rre
n
t
i
n
tere
sts
in
c
lu
d
e
d
a
ta
W
a
re
h
o
u
se
,
d
a
ta
M
in
i
n
g
,
A
lg
o
rit
h
m
s an
d
b
ig
d
a
ta.
Evaluation Warning : The document was created with Spire.PDF for Python.