Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
15
,
No.
1
,
Febr
uary
20
25
, pp.
1
200
~
1
208
IS
S
N:
20
88
-
8708
, DO
I:
10
.11
591/ij
ece.v
15
i
1
.
pp
1
200
-
1
208
1200
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om
Intrusio
n detecti
on and
prev
ention
using
Bayesia
n decisi
on
with
fuzzy l
ogic syste
m
Sa
t
heesh
kum
ar
Se
ka
r
1
, P
ala
nira
j R
aj
idu
rai P
arvathy
1
,
Gopal
Kum
ar
Gupt
a
2
,
Thi
ruvenkad
ac
ha
ri
Raj
agopalan
3
, Chet
han C
handr
a
S
u
bha
s
h
Chand
ra
B
as
app
a
B
as
av
ar
ad
di
4
,
Kuppan
Pad
manab
an
5
,
Subbi
ah M
uru
gan
6
1
Mph
asis
Co
rpo
ration
,
Ch
an
d
ler,
Un
ited
States of
Am
er
ica
2
Sy
m
b
io
sis
I
n
stitu
t
e of Tech
n
o
lo
g
y
N
ag
p
u
r
Campu
s, Symbio
sis
I
n
ternatio
n
al (
Dee
m
ed
Univ
ersity
),
Pun
e,
Ind
ia
3
Dep
artm
en
t of
M
ath
em
atics,
Univ
er
sity
Co
lleg
e of E
n
g
in
eering
,
An
n
a U
n
iv
ersity
,
Ariyalu
r
,
I
n
d
ia
4
Dep
artm
en
t of
Ar
tificial
Intellig
en
ce
and
M
achi
n
e L
ear
n
in
g
,
Do
n
Bo
sco
I
n
stitu
te of T
echn
o
lo
g
y
,
Ban
g
alo
re,
I
n
d
ia
5
Dep
artm
en
t of
Co
m
p
u
ter
Sci
en
ce a
n
d
E
n
g
in
eering
,
Ko
n
eru
Laks
h
m
aiah
Edu
catio
n
Fou
n
d
at
io
n
,
An
d
h
ra
Prade
sh
,
Ind
ia
6
Dep
artm
en
t of
Bi
o
m
ed
ical E
n
g
in
eer
in
g
,
Sav
eeth
a Scho
o
l of E
n
g
in
eering
,
Sav
eeth
a I
n
stitu
te
o
f
Medical
and
Te
ch
n
ical Sciences,
Sav
eeth
a Univ
ersit
y
,
Ch
en
n
ai,
Ind
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
M
a
r 20,
2024
Re
vised
Sep 6
,
2024
Accepte
d
Oct
1,
2024
Now
ada
ys,
int
ru
sion
de
te
c
ti
on
a
n
d
pre
v
ent
ion
met
hod
has
co
mprehende
d
th
e
noti
c
e
to
d
ec
r
ea
s
e
the
eff
ec
t
of
in
trude
rs.
Deni
al
o
f
servi
ce
(DoS
)
i
s
an
atta
ck
tha
t
for
mul
a
te
s
ma
licious
tr
aff
i
c
is
distri
bu
te
d
int
o
an
ex
ac
t
i
ng
net
work
devi
c
e.
Th
ese
a
t
ta
ck
ers
abso
rb
w
it
h
a
va
li
d
n
et
wo
rk
dev
ice
,
th
e
va
li
d
device
will
b
e
com
pro
mi
sed
to
insert
ma
licious
tra
ff
ic.
To
solve
th
ese
proble
ms,
th
e
Baye
sian
d
ec
isi
on
mod
el
with
a
fu
zz
y
logi
c
sys
te
m
base
d
o
n
int
rusio
n
det
e
ct
ion
and
pr
eve
nt
ion
(BDF
L)
is
int
roduc
ed.
Thi
s
mecha
nis
m
sepa
rates
the
DoS
pa
ckets
base
d
on
the
t
ype
of
v
al
id
atio
n,
su
ch
as
pac
k
et
and
flow
val
id
at
ion
.
Th
e
BDF
L
me
ch
an
ism
uses
a
fu
z
zy
logic
sys
te
m
(FLS)
for
val
id
at
ing
the
d
at
a
pac
k
et
s.
Al
so,
the
key
feature
s
of
the
al
g
orit
hm
are
exc
erp
te
d
fro
m
dat
a
p
ac
k
et
s
a
nd
ca
t
egor
ized
int
o
norm
al
,
do
ubtful
,
and
ma
licious.
Furth
erm
ore
,
the
Bay
esia
n
d
ecision
(
BD)
dec
id
e
two
queue
s
as
ma
licious
and
n
orma
l
.
The
BDF
L
m
ec
h
ani
sm
is
exp
eri
m
ent
a
l
in
a
ne
twork
simul
at
or
envi
r
onme
nt
,
and
th
e
oper
a
ti
ons
ar
e
measures
reg
ard
ing
DoS
at
t
ac
ker
de
tecti
o
n
ratio,
delay
,
tr
a
ffic
loa
d
,
and
thr
oughput.
Ke
yw
or
d
s
:
Ba
yesian deci
s
ion
al
gorithm
Den
ia
l
of ser
vi
ce
Fu
zz
y
lo
gic s
yst
em
In
tr
us
i
on d
et
ec
ti
on
a
nd
pr
e
ve
ntion
Qu
e
ue mana
ge
ment
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
BY
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Gopal
Kumar
Gupta
Sy
m
biosi
s Inst
it
ute o
f
Tec
hnol
ogy Nagp
ur Camp
us
,
S
ymbiosis I
nter
natio
nal (Deeme
d
U
niv
e
rsity)
Lavale,
Mulshi
, Pun
e
, Ma
har
a
sh
tra
4121
15, I
nd
ia
Emai
l:
gopalg
upta
.ii
tbhu90
@
gm
ai
l.co
m
1.
INTROD
U
CTION
The
mode
rn
cr
eat
ion
of
intr
usi
on
detect
io
n
sy
ste
ms
(
IDS)
gro
wingly
de
man
ds
machi
ne
-
co
ntr
olled
and
i
ntell
igent
ID
S
t
o
deal
with
th
reats,
c
ausin
g
a
rise
i
n
the
numb
e
r
of
e
nc
ourag
e
d
at
ta
cker
s
in
t
he
cy
be
r
env
i
ronme
nt
[
1]
.
Sp
eci
fical
ly,
the
re
ha
ve
be
en
great
nece
ssit
at
es
fo
r
a
uto
nom
ous
age
nt
-
base
d
I
DS
s
ol
ution
s
that
ne
cessi
ta
te
as
li
tt
le
hum
an
i
nter
fer
e
nce
as
feasi
ble
w
hen
bein
g
cap
able
of
dev
el
opin
g
a
nd
en
ha
ncin
g
it
sel
f
an
d
de
ve
lop
in
g
into
more
rob
us
t
t
o
pro
bab
le
th
reats.
E
nhan
ci
ng
c
yb
e
rsec
uri
ty
by
detect
ing
a
nd
pr
e
ve
nting
i
ntr
us
io
ns
at
en
dpoin
ts
usi
ng
ma
chine
le
ar
ning
al
gorithms
is
a
su
ccess
fu
l
stra
te
gy
[2]
.
Tr
adi
ti
on
al
intru
si
on
detec
ti
on
a
nd
preve
ntion
meth
ods
of
te
n
dep
e
nd
on
sig
natu
re
-
base
d
ap
proac
hes.
H
ow
e
ve
r,
these
methods
ma
y
no
t
be
en
ough
to
ide
ntify
f
res
h
a
nd
c
omplex
assault
s.
O
n
t
he
oth
e
r
hand,
al
gorithms
that
le
arn
via
machi
ne
le
arn
i
ng
ca
n
eva
luate
patte
rn
s
,
beh
a
viors,
a
nd
anomal
ie
s
within
the
data
to
detect
po
s
sibl
e
risks
in r
eal
-
ti
me
[
3]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
omp E
ng
IS
S
N:
20
88
-
8708
In
tr
us
io
n detec
ti
on
and preve
ntion u
sin
g
B
ay
esi
an d
eci
si
on wi
th fuzzy
lo
gic
syste
m
(
Sa
t
he
eshk
umar Sek
ar
)
1201
Den
ia
l
of
ser
vi
ce
(
D
oS
)
a
nd
distrib
uted
de
nial
of
se
rv
ic
e
(DD
oS)
a
re
both
well
-
know
n
ty
pes
of
assault
s
that
ha
ve
the
same
go
al
,
w
hich
i
s
to
exh
a
us
t
t
he
c
omp
uting
ca
pabi
li
t
ie
s
of
a
ho
st
or
ta
r
get.
Net
work,
making
the
m
inacce
ssible
to
le
giti
mate
us
ers
w
ho
are
a
ll
ow
ed
t
o
us
e
them
or
ad
ve
rsely
harmi
ng
the
performa
nce
of their comp
uter
sy
ste
m in so
m
e w
ay
[
4]
. D
oS at
ta
cks
may b
e b
r
oken d
own
into
a fe
w diffe
ren
t
kinds,
th
e
mos
t
comm
on
of
wh
ic
h
are
s
of
t
war
e
vu
l
ner
a
bi
li
ti
es
and
flo
odin
g
at
ta
cks
[
5]
.
W
he
n
an
a
tt
acker
util
iz
es
softwa
re
e
xploit
s,
t
he
y
a
re
ta
king
ad
va
ntage
of
vulne
ra
bili
ti
e
s
in
their
ta
r
get
se
rv
e
r
to
ei
ther
com
plete
ly
ce
ase
the
ser
vic
es
bei
ng
pro
vid
ed
by
the
se
rv
e
r
or
dr
ast
ic
al
ly
de
grade
t
he
perf
or
m
anc
e
it
is
capab
le
of
.
In
the
e
ve
nt
of
fl
ooding
at
ta
cks
,
the
at
ta
cke
r
us
es
up
al
l
of
the
s
ys
te
m'
s
re
so
urces
by
se
ndin
g
man
y
inc
orrect
requests,
le
a
din
g
t
o
the
c
halle
ng
e
s
disc
us
se
d
in
the
previ
ous
sect
io
n.
T
he
mo
st
rece
nt
ve
rsion
of
t
he
DDoS
ta
kes
a
dvanta
ge
of
m
ulti
ple
netw
ork
c
omp
on
e
nts'
powe
r
to
inc
rease
the
threat
ca
us
e
d
by
the
at
ta
ck
by
distr
ibu
ti
ng
it
acr
oss
sla
ve
machi
nes
[6]
.
Im
plementin
g
mac
hi
ne
le
ar
ning
(
M
L
)
ba
sed
i
ntru
si
on
de
te
ct
ion
a
nd
pr
e
ve
ntion
at
endp
oin
ts
re
quires
ca
refull
y
choosi
ng
a
ppr
opriat
e
al
go
rithms
,
qual
it
y
da
ta
,
an
d
ongoin
g
monit
or
i
ng
an
d
ada
ptati
on
.
B
y
c
onti
nuously
re
fi
ning
t
he
mod
el
s
an
d
i
nteg
r
at
ing
t
hem
int
o
t
he
secur
it
y
w
orkf
l
ow,
orga
nizat
ion
s
ca
n
si
gn
ifi
cantl
y
en
ha
nce
their
abili
ty
to
detect
and
pr
e
ven
t
int
ru
si
ons
at
the
endp
oin
t l
e
vel
[7]
.
Re
searc
h
ga
p
:
a
n
intel
li
gen
t
I
DS
is
intr
oduced
f
or
the
s
ecur
it
y
inter
ne
t
protoc
ol
(
IP
)
mu
lt
ime
dia
su
bsyste
m
(
IMS)
by
a
pplyin
g
M
L
t
o
raise
cl
assifi
er
acc
ur
a
cy
,
this
mec
ha
nism
c
hoos
es
vital
featur
es
to
buil
d
an
I
DS
.
This
mecha
nism
presents
the
decisi
on
tree
(DT),
s
upport
vecto
r
mac
hin
e
(
SVM),
as
well
as
Naiv
e
Ba
yesian
cl
ass
ifie
r
to
meas
ure
intr
us
io
n
detect
ion
e
ff
ic
ie
nt
ly.
H
oweve
r,
these
cl
assifi
er
s
can
no
t
detec
t
th
e
DoS
at
ta
ck
e
ffi
ci
ently
[8]
.
T
o
so
l
ve
t
hese
issues
i
ntr
us
i
on
detect
io
n
an
d
pr
e
ve
ntion
us
i
ng
Ba
yesian
de
ci
sion
with
f
uzzy
lo
gi
c
syst
em
is
pro
po
s
ed
.
T
he
re
mainde
r
of
t
he
pa
pe
r
is
str
uctur
e
d
as
f
ollows.
Se
ct
ion
2
dis
cusses
the
Ba
yesian
decisi
on
mode
l
with
a
fu
zz
y
lo
gic
s
ys
te
m
base
d
on
int
rusion
detect
io
n
and
pr
e
ve
ntion.
T
he
netw
ork
sim
ul
at
ion
a
nalysis
is sp
eci
fied
i
n
s
ect
ion
3.
Finall
y,
we
c
oncl
ude
the
pap
e
r
i
n
se
ct
ion
4.
In
tr
us
i
on
pr
e
ve
ntion
sy
ste
m
(I
P
S)
detect
s
c
yb
e
r
-
at
ta
c
ks
by
ap
ply
i
ng
ma
chine
le
ar
ning
al
gorithms
li
ke
SVM
an
d
forest
al
gorith
m
[9]
.
Hac
king
ide
ntific
at
ion
thr
ough
the
fu
zz
y
lo
gic
syst
em
us
in
g
the
fu
zz
y
al
gorithms
to
no
ti
ce
t
he
inje
ct
ion
at
ta
cks
[
10]
.
H
oweve
r,
it
create
s
im
porta
nt
er
rors.
Gr
a
ph
-
base
d
i
ntr
us
io
n
detect
ion
s
ys
te
m
int
r
oduces
t
he
seq
ue
nce
of
interc
hange
d
message
s.
T
houg
h,
it
can
not
no
ti
ce
at
ta
ck
s
by
analyzin
g
se
pa
rated
frames
a
nd
al
so
has
a
n
importa
nt
er
ror
[
11]
.
I
DS
is
hi
gh
l
y
pro
fici
ent,
bu
t
it
is
al
so
a
ble
to
detect
thre
at
s.
The
a
nomaly
de
te
ct
ion
s
ys
te
m
is
ca
pa
ble
of
disco
ver
i
ng
a
nomali
es
with
a
le
sser
false
posit
ive
and
false
ne
ga
ti
ve.
T
his
mec
han
is
m
rec
ogni
zes
dee
p
le
ar
ni
ng
t
o
e
nh
a
nce
scal
abili
ty
an
d
ef
fici
enc
y.
F
eat
ur
e
sta
nd
a
rd
iz
at
io
n
can
util
iz
e
to
raise
acc
ur
a
cy.
Diff
e
re
nt
featur
e
sel
ect
ion
mecha
nism
s
to
ch
oose
c
ertai
n
featur
e
s that ca
n
ma
nipulat
e r
esult
s b
et
te
r
. H
ow
e
ve
r,
t
his m
echan
is
m r
ai
se
s the
netw
ork
t
raffic
[
12]
.
A
deep
Q
-
le
ar
ning
met
hod
of
fer
s
an
aut
o
-
le
arn
i
ng
capa
bili
ty
f
or
an
en
vir
onment
t
hat
ca
n
disti
nguish
sever
al
netw
ork
i
ntr
us
io
ns
.
T
his
mecha
nism
ca
ught
a
nd
e
xamine
d
to
noti
ce
mali
ci
ou
s
pay
l
oads
i
n
a
sel
f
-
l
earn
in
g
fas
hi
on.
Howe
ver
,
t
hi
s
mechan
is
m
do
e
s
not
us
e
th
e
cl
oud
en
vir
onme
nt
[13
]
.
A
neural
netw
ork
with
deep
le
ar
ning
(D
L
)
is
t
h
e
m
os
t
a
pprop
riat
e
f
or
detect
in
g
DoS
at
ta
cks
th
at
hav
e
reac
he
d
high
-
perf
ormance
accurac
y.
H
oweve
r,
t
his
m
echan
is
m
does
no
t
guara
nte
e
they
opti
mall
y
cat
egorize
un
i
de
ntifie
d
pack
et
s
incomi
ng
t
hro
ugh
a
w
eb
se
r
ver
beca
us
e
th
e
identific
at
ion
po
s
sibil
it
y
is
high
but
has
a
degree
of
ins
ecur
it
y
[14]
.
T
he
intr
usi
on
detect
io
n
and
pr
e
ven
ti
on
sy
ste
m
a
pp
l
y
m
od
el
-
base
d
intru
si
on
detect
ion
a
nd
M
L
-
ba
se
d
intru
si
on
pr
e
ve
ntion
to
def
e
nd
the
net
work.
The
detect
io
n
ph
a
se
reduces
netw
ork
featu
r
es
an
d
exa
min
es
them
to
dete
rmin
e
wh
et
her
the
ne
twork
is
in
a
nor
mal
sta
te
.
T
his
mec
ha
nism
util
iz
es
Q
-
le
ar
ning
a
nd,
t
hro
ughout
interact
ions,
di
sco
ver
s
the
bes
t scheme
ag
ai
nst
an
att
ack
[15]
.
Cyb
e
r
def
e
ns
e
dema
nds
f
unct
ion
s
co
nducte
d
in
the
cy
b
e
rse
cur
it
y
fiel
d,
de
fendin
g
missi
on
ta
r
gets
to
recog
nize
an
d
av
oid
c
yberat
ta
cks,
i
nclu
ding
ID
S
,
as
wel
l
as
intr
us
io
n
pr
e
ve
ntion.
E
xpla
inable
A
rtif
ic
ia
l
In
te
ll
igence
al
gorithm
for
a
nomaly
-
base
d
ID
S
i
n
i
nter
ne
t
of
t
hings
(
IoT
)
net
wor
ks.
In
it
ia
ll
y,
the
I
DS
s
c
on
ce
ntrate
on
an
om
al
y
-
ba
sed
detect
io
n
te
chn
iq
ues
t
o
offer
tr
us
t
an
d
co
nf
i
den
ce
.
Nex
t,
util
iz
e
DL
to
eff
ic
ie
ntly
dete
ct
an
an
om
al
y,
pro
vid
in
g
bette
r
pe
rfo
rma
nce
s
[1
6]
.
M
L
al
gorith
ms
act
a
vi
ta
l
ta
sk
in
buil
ding
an
I
DS
.
An
e
nse
mb
le
meth
od
usi
ng
ra
ndom
su
bspace
i
n
t
ha
t
an
e
xtreme
l
earn
i
ng
mac
hi
ne
is
pr
e
ferre
d
as
the
base
cl
assifi
er
.
An
e
ns
e
mb
le
p
r
unin
g
meth
od
wa
s
est
ablish
ed
on
the b
at
al
gorithm f
it
ness
functi
on
to
e
nhance
the
cl
assifi
er
s
ub
s
et
[17
]
.
A
deep
rein
force
ment
le
ar
ning
al
gorithm
that
util
iz
es
the
M
a
rko
v
decisi
on
process
to
e
nh
a
nce
t
he
ID
S
d
eci
sio
n
operati
on.
T
his mech
a
nism pr
ov
i
des
bette
r
de
te
ct
ion
p
er
for
mance
an
d
min
imi
zes
the
false
al
ar
m
co
unt.
Howe
ve
r,
t
his
mecha
nism
does
not
detect
the
do
ubtf
ul
at
ta
cker
[
18]
.
A
fe
de
rated
dee
p
reinfo
rceme
nt
le
arn
in
g
-
base
d
IDS
in
t
hat
sever
al
a
gen
ts
are
distri
bu
te
d
on
the
netw
ork
,
an
d
the
se
agen
ts
exten
d
a
d
ee
p
Q
-
netw
ork
l
ogic
.
It
c
on
cei
ve
d
e
ver
y
a
ge
nt's
data
pr
i
vacy
oc
cup
ie
d
wh
il
e
desig
ning
t
he
s
ys
te
m
,
and eve
r
y
a
gent
d
oe
s
no
t
distr
ibu
te
the
dat
a to
oth
e
r nodes
[
19]
.
M
ulti
-
a
gen
t
fe
at
ur
e
s
el
ect
ion
-
base
d
I
DS
c
onsti
tutes
a
feat
ur
e
sel
f
-
sel
ect
ion
a
nd
a
deep
reinfo
rceme
nt
le
arn
in
g
at
ta
ck
d
et
ect
ion
. T
he
f
eat
ure sel
f
-
se
le
ct
ion
meth
od
p
urc
hases
m
ul
ti
-
agen
t
rein
for
cement le
ar
ning that
sp
eci
fies
the
is
su
es
of
featu
re
sel
ect
ion
.
F
ur
therm
or
e
,
it
m
inimi
zes
the
c
omplexit
y
a
nd
impro
ves
the
sear
ch
strat
egy
to
c
ho
os
e
the
feat
ur
e
.
Mor
eo
ver,
th
e
gr
a
ph
c
onvo
luti
on
al
netw
ork
meth
od
ev
okes
deep
e
r
fea
tures
from
the
data.
This
mec
ha
nism
e
nhances
t
he
acc
uracy
[
20]
.
T
he
t
hr
ea
t
of
DDoS
ha
s
de
vel
op
e
d
w
it
h
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
omp E
ng,
V
ol.
15
, No
.
1
,
Febr
uary
20
25
:
1
200
-
1
208
1202
increase
of
i
nt
el
li
gen
t
in
form
at
ion
s
ys
te
m
s.
This
mec
ha
nism
util
iz
es
a
de
ci
sion
tree
-
ba
sed
mode
l
t
o
detect
DDoS
at
ta
cks
[
21]
.
C
yb
e
r
secur
it
y
a
ppr
oa
ches
to
en
ha
nce
t
he
secu
ri
ty
e
valuates
a
gainst
cy
be
r
-
a
tt
ac
ks
.
Conve
ntion
al
secur
it
y
so
l
ution
s
de
scribi
ng
an
d
dev
el
op
i
ng
secu
rity
t
hreat
s
[
22]
.
Th
e
wireless
fid
el
it
y
(
Wi
-
Fi
)
-
e
na
ble
d
e
nerg
y
obse
r
v
in
g
f
or
s
olar
-
powe
red
buil
di
ng
s
that
will
pe
rmit
f
or
daily
an
d
week
l
y
le
arn
of
energ
y
util
iz
e
[2
3]
.
I
DS
pur
po
se
is
to
exte
nd
the
sec
ur
it
y
in
an
Io
T
e
nv
ir
on
ment
[24]
.
Distrib
uted
mu
lt
i
-
a
gen
t
ID
S
util
iz
ed
le
arn
i
ng
a
ge
nts
separ
at
e
th
e
nor
mal
or
at
ta
c
k
ba
sed
on
ne
twork
be
hav
i
or
[
25]
.
T
he
Ba
yesian
decisi
on
(
BD
)
model
est
ablis
hed
a
reli
able
r
ou
te
formati
on
and
it
separa
te
s
an
unreli
abl
e
sens
or
a
nd
f
orward
the
data
ef
fici
ently
[26]
.
T
he
ID
S
usi
ng
a
n
e
ns
em
ble
m
odel
to
desig
n
a
s
mart
homes
t
o
recog
nize
the
a
tt
acks
[27]
.
2.
PROP
OSE
D MET
HO
D
The
int
ru
si
on
pr
e
ve
ntion
sy
s
te
m
(IPS)
is
a
n
e
xtensiv
e
I
DS
that
a
ssist
s
in
obse
rv
i
ng
al
l
normal
models
of
tra
ffi
c
and
t
ransmi
ts
al
erts
in
case
of
an
y
dif
fer
e
nce
f
r
om
t
he
nor
mal
m
od
el
.
Since
the
pu
blic
c
a
n
gr
a
nted
acce
ss
to
the
netw
ork,
the
data
pac
ke
ts
f
orward
fro
m
intr
uders
a
re
co
mb
i
ned
int
o
the
net
work
tr
aff
ic
as
input.
It
is
e
ssentia
l
to
obse
rv
e
al
l
in
ward an
d
de
par
ti
ng
t
raffic
.
Th
e
dat
a
pack
et
s f
r
om
the
valid u
se
rs
come
into
s
witc
hes,
w
her
eas
the
data
pa
ckets
a
re
c
orrob
or
at
e
d
util
iz
ing
fe
a
tures.
Alth
ough
the
at
ta
ckers
ar
e
intercepte
d,
th
e
com
promise
d
us
e
r's
in
volv
ement
is
prese
nt
in
the
netw
ork
.
The
c
omp
r
om
ise
d
us
e
rs
di
stribu
te
mali
ci
ou
s
pac
kets
to
al
l
nodes,
drai
ni
ng
the
netw
ork
r
eso
ur
ces
.
T
his
mec
han
is
m
de
te
ct
s
an
d
pr
e
ven
ts
intru
si
on
li
ke
DoS
at
ta
c
ks
ba
sed
on
pa
cke
t
an
d
fl
ow
cl
assifi
e
rs.
Fi
gure
1
ex
plains
the
st
ru
ct
ur
e
of
the
Ba
yesian deci
s
ion
model
with
a fuzz
y
lo
gic s
ys
te
m
(
BD
FL
)
appr
oach.
Figure
1. Str
uc
ture of
BD
FL a
ppr
oach
Fr
om
F
ig
ur
e
1,
the
up
c
omi
ng
data
pac
ket
is
validat
e
d
by
f
uzz
y
l
og
ic
s
ys
te
m
(
F
LS
)
membe
rs
hip
functi
ons
a
nd
the
q
ue
ue
man
ageme
nt
va
li
da
te
s
the
data
fl
ow.
T
he
i
nw
a
r
ds
pac
kets
are
cl
assifi
ed
mali
ci
ou
s
,
normal
a
nd
do
ub
t
fu
l;
th
us
,
se
par
at
es
a
n
a
bn
ormal
be
ha
vio
r
.
The
n,
the
Ba
yesian
d
eci
sio
n
al
gorit
hm
c
he
cks
th
e
doubtf
ul
pac
ke
t
qu
e
ue.
I
f
c
he
ck
ed
the
pac
ke
t
is
an
a
bnorm
al
beh
a
vi
or
t
ype,
it
is
disti
ng
uish
e
d
as
the
a
tt
ack
and
in
vokes
th
e
al
arm
as
a
s
ign
al
t
o
t
he
a
ppreciat
ed
dev
i
ce
holde
r.
T
hus,
the
BD
FL
mecha
nism
uti
li
zi
ng
In
te
ll
igent
I
DS f
or
deter
minin
g wh
et
her the
performa
nce
of traf
fic is n
orm
a
l or n
ot.
2.1
.
FLS
-
b
as
ed
Do
S attack
de
tec
tion
The
packet
-
ba
s
ed
validat
ion mo
du
le
t
hat
de
te
ct
s
the
D
oS
by
the FLS
.
He
r
e,
the
i
nput
pac
ket
feat
ur
es
are
e
xcerpted
from
t
he
pac
ke
t
hea
der
,
li
ke
sen
der
a
nd
re
cei
ver
a
ddres
s,
t
yp
e
of
p
ro
t
oc
ol,
a
nd
se
nd
e
r
a
nd
rec
ei
ver
port
[
28]
.
T
he
ob
ta
i
ned
i
nputs
are
confirme
d
with
the
pr
i
nciple
s,
an
d
it
makes
an
outp
ut
li
ke
normal,
doubtf
ul,
a
nd
mali
ci
ou
s
pa
c
kets.
T
he
rig
ht
ness
of
t
he
pa
cket
featu
re
def
i
nes
the
F
LS.
Fig
ure
2
exp
la
in
s
FLS
-
base
d D
oS at
ta
ck dete
ct
ion an
d
T
able
1 sh
ows FL
S ta
ble.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
omp E
ng
IS
S
N:
20
88
-
8708
In
tr
us
io
n detec
ti
on
and preve
ntion u
sin
g
B
ay
esi
an d
eci
si
on wi
th fuzzy
lo
gic
syste
m
(
Sa
t
he
eshk
umar Sek
ar
)
1203
Tab
le
1.
FLS
t
able
Typ
e of Pr
o
to
co
l
Sen
d
er
an
d
r
eceive
r
ad
d
ress
Sen
d
er
an
d
r
eceive
r
p
o
rt
Ou
tp
u
t
Hig
h
(H)
Hig
h
Low (
L)
No
rm
al
pack
et
Midd
le (
M
)
Midd
le
Midd
le
Do
u
b
tful p
acket
Hig
h
Low
Hig
h
Maliciou
s p
acket
Figure
2. FLS
-
base
d Do
S
att
ack
detect
ion
Inwards
pack
e
ts
fr
om
nor
mal
us
ers
a
re
puri
fied,
a
nd
t
her
e
fore,
the
us
a
ge
of
ov
e
rloa
d
ba
ndwidt
h
is
minimi
zed
.
Ba
ndwidt
h
is
a
ne
cessar
y
res
our
ce
to
op
e
rate
the
ar
rive
d
pac
kets.
T
hus,
t
he
band
width
de
fi
ci
ency
create
s
hea
vy
pack
et
l
os
s.
T
he
in
ward
pac
kets
co
rr
e
spo
nd
wit
h
the
fiel
ds
;
the
featu
re
s
are
e
xpresse
d
f
or
purified
act
io
n.
T
he
purificat
ion
is
an
im
por
ta
nt
f
un
ct
io
n
of
ke
y
pac
kets
f
rom
in
div
i
du
al
us
e
rs.
The
f
unct
ion
of
filt
rati
on
is
t
o
e
va
de
distrib
uting
mali
ci
ou
s
traf
fic
at
a
prel
imi
nar
y
sta
ge
.
T
he
FLS
out
pu
t
dis
plays
no
rmal,
Dou
btf
ul,
an
d
mali
ci
ou
s p
ac
ke
ts.
The
n,
t
he
normal p
ac
kets
are
tra
ns
fe
rred
,
the
mali
ci
ous p
ackets
a
re r
ej
ect
ed
,
and the
pac
ket
band
width ve
ri
fies the
pac
ket.
2.2
.
Queue
ma
nag
ement
system
-
ba
sed
DoS
attack
d
etec
tion
In
this
mec
hani
sm,
the
doubtf
ul
pack
et
is
ve
rified
by
the
B
D
al
gorithm
.
T
he
pac
ket
ov
e
r
flo
w
is
al
s
o
encou
ntere
d
i
n
the
que
ue
bec
ause
of
t
he
i
nt
ru
si
on
in
volve
ment.
Be
f
or
e
c
onfirmi
ng
th
e
flo
w,
the
ba
nd
width
exp
e
ndit
ur
e
is
eval
uated.
B
ecause
t
he
ba
ndwidt
h
is
th
e
f
undame
ntal
res
ource
extr
emel
y
e
ngage
d
by
intru
si
ons
an
d
is
no
t
ap
pro
pr
i
at
ed
f
or
norma
l
pack
et
s,
it
pr
esents
tw
o
que
ues,
s
uc
h
as:
nor
mal
queue
(
NQ)
mali
ci
ou
s
queu
e
(
M
Q
).
The
data
packet
s
that
are
pa
rtly
c
oupled
w
it
h
the
fu
zz
y
r
ule
act
ivit
y
a
re
treat
ed.
Alte
r
nativel
y,
the
pack
et
s
in
doubtf
ul
a
re
e
xpli
ci
t
within
the
port
as
well
as
out
port
numbe
r.
All
i
nw
a
r
ds
do
ub
t
fu
l
pac
ke
ts
are
treat
ed
one
at
a
ti
me
us
i
ng
B
D
al
gorithm
.
T
he
sp
eci
fied
tr
aces
a
re
a
ppli
ed
i
n
de
velo
ping
a
rati
onal
de
ci
sion
from
tr
ue
e
vide
nc
e.
He
re,
the
BD
al
gorit
hm
is
a
ppli
ed
t
o
de
ci
de
t
he
possi
bili
ti
es
of
s
pec
ific
eve
nt
t
ype
s.
T
he
trace
is
em
ployed
to
cal
c
ulate
the
c
onditi
onal
pro
ba
bili
ty
f
or
var
ia
bles
with
s
pecifie
d
i
nfo
rmati
on
re
ga
rd
i
ng
pack
et
que
ue.
The
B
D
al
gorithm
in
Ba
yes'
r
ule
is
uti
li
zed
t
o
i
nform
the
possibil
it
y
of
e
va
luati
ng
a
trac
e
as
a
n
add
it
io
nal
co
nfi
rmati
on.
T
he
BD
al
gorithm
app
li
es
to
t
he
NQ,
an
d
the
M
Q
is
s
pecifie
d
in
(1)
an
d
(2).
He
re,
de
no
te
s
the t
otal qu
e
ue,
repr
esents the
mali
ci
ou
s
que
ue,
a
nd
exp
la
in
s a
no
rmal
queue
.
(
|
)
=
(
)
∗
P
(
|
)
P
(
)
(1)
(
|
)
=
(
)
∗
P
(
|
)
P
(
)
(2)
The
qu
e
ue
possibil
it
y
is
gr
e
at
er
than
t
he
t
hr
es
hold
wh
e
n
the
us
e
r
qu
e
ue
is
ide
ntifie
d
in
act
io
n
detect
ion
;
t
he
BD
determi
ne
s
w
hethe
r
nor
mal
or
mali
ci
ou
s
qu
e
ue.
T
he
valu
e
of
t
he
thres
hold
is
pr
ese
nt
amo
ng
0
t
o
1.
T
he
t
hr
es
hold
val
ue
is
set
to
0.5,
a
nd
the
p
ossi
bili
ty
is
be
tt
er
tha
n
a
th
r
esh
old
that
qu
eue
is
normal
or
mali
ci
ou
s
in t
he netwo
rk.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
omp E
ng,
V
ol.
15
, No
.
1
,
Febr
uary
20
25
:
1
200
-
1
208
1204
3.
SIMULATI
O
N
ANALY
SIS
This
mec
han
is
m
util
iz
es
a
net
work
sim
ulato
r
-
3
to
e
valuate
the
net
work
performa
nce
[
29]
.
It
buil
ds
a
flo
w
ta
ble
i
n
t
he
entit
y
O
pe
nF
lo
w
s
witc
h
base
d
on
t
hat
inwa
rd
s
pac
ke
t
is
ei
the
r
lost
or
tra
ns
mit
te
d.
T
his
mecha
nism
a
ppli
es
the
sim
ul
at
ion
par
a
mete
rs
to
pla
n
the
pro
po
se
d
s
ys
te
m.
T
he
obje
ct
ive
of
this
mec
han
is
m
is
to
disti
ng
uish
a
nd
al
le
via
te
the
D
oS
at
ta
cker
s
in
t
he
netw
ork;
th
us
,
the
par
a
mete
r
s
prefe
rab
le
i
n
this
mecha
nism
are
dela
y,
detect
io
n
rati
o,
th
rou
ghput,
an
d
traf
fi
c
loa
d
i
n
t
he
ne
twork
[3
0]
.
T
he
detect
ion
ra
te
is
a
vital
par
a
mete
r
that
re
present
s
the
e
ff
ic
ie
nt
forecast
in
g
of
DoS
at
ta
cks
.
F
igure
3
e
xp
la
in
s
a
detect
ion
r
at
io
of
SVM
[31]
, DT
, and
B
DF
L
appr
oac
hes base
d o
n DoS
at
ta
ck
er
.
Figure
3
.
Dete
ct
ion
r
at
io
of S
VM,
DT, an
d B
DF
L a
ppr
oac
hes base
d on D
oS
a
tt
acke
r
F
r
o
m
t
h
i
s
r
e
s
u
l
t
,
t
h
e
B
D
F
L
m
e
c
h
a
n
i
s
m
h
a
s
a
g
r
e
a
t
e
r
d
e
t
e
c
t
i
o
n
r
a
t
i
o
a
t
r
a
i
s
i
n
g
a
n
a
t
t
a
c
k
e
r
c
o
u
n
t
.
T
h
i
s
r
a
i
s
e
i
s
b
e
c
a
u
s
e
o
f
t
h
e
s
u
i
t
a
b
l
e
r
e
c
o
g
n
i
t
i
o
n
o
f
a
t
t
a
c
k
s
u
t
i
l
i
z
i
n
g
F
L
S
a
n
d
B
D
a
l
g
o
r
i
t
h
m
s
.
T
h
e
p
r
e
l
i
m
i
n
a
r
y
l
e
v
e
l
i
.
e
.
,
w
i
t
h
o
u
t
D
o
S
a
t
t
a
c
k
e
r
f
o
r
B
D
F
L
,
S
V
M
a
n
d
D
T
m
e
c
h
a
n
i
s
m
s
d
e
t
e
c
t
i
o
n
r
a
t
i
o
i
s
1
,
0
.
9
7
,
a
n
d
0
.
9
5
.
T
h
e
e
x
i
s
t
i
n
g
S
V
M
a
n
d
D
T
-
b
a
s
e
d
I
D
S
m
e
c
h
a
n
i
s
m
s
t
h
a
t
t
he
p
a
c
k
e
t
s
a
r
e
r
e
c
o
g
n
i
z
e
d
a
s
n
o
r
m
a
l
o
r
m
a
l
i
c
i
o
u
s
;
h
o
w
e
v
e
r
,
i
t
r
a
i
s
e
s
t
h
e
f
a
l
s
e
a
l
a
r
m
.
T
h
e
p
r
o
p
o
s
e
d
B
D
F
L
m
e
c
h
a
n
i
s
m
d
e
t
e
c
t
s
t
h
e
c
o
m
p
r
o
m
i
s
e
d
p
a
c
k
e
t
s
e
f
f
i
c
i
e
nt
l
y
,
ho
w
e
v
e
r
,
t
h
e
S
V
M
a
n
d
D
T
-
b
a
s
e
d
I
D
S
m
e
c
h
a
n
i
s
m
s
c
a
n
n
o
t
d
e
t
e
c
t
t
h
e
c
o
m
p
r
o
m
i
s
e
d
p
a
c
k
e
t
s
e
f
f
i
c
i
e
n
t
l
y
.
Traffic
loa
d
is
a
si
gn
i
ficant
par
a
mete
r
i
n
i
ntr
us
io
n
detect
ion
an
d
pre
vent
ion
w
her
e
the
co
ntri
bu
ti
o
n
of
the
DoS
at
ta
cker
s
is
noti
c
ed.
M
os
t
re
gula
rly,
the
DoS
at
ta
cker
's
obje
ct
ive
is
to
ex
ha
us
t
al
l
the
res
ources.
The
enlar
ge
me
nt
in
ir
regular
traff
ic
loa
d
de
ci
des
w
hich
at
ta
ck
pac
kets
m
ay
be
in
vo
l
ved.
T
he
traf
fic
lo
ad
f
or
DT, SV
M, a
nd BDFL
is
disp
l
ayed in
Fig
ur
e
4,
res
pecti
ng th
e rise in
num
be
r of
D
oS
att
ack
ers.
Figure
4
.
Tra
ffi
c
l
oad
of S
V
M
, DT
, a
nd B
DF
L
ap
proac
he
s b
ase
d o
n DoS
a
tt
acker
0
0.2
0.4
0.6
0.8
1
1.2
0
5
10
15
20
25
SV
M
DT
B
D
F
L
DoS
Atta
cker
Detection
Ra
tio
0
1
2
3
4
5
6
7
5
10
15
20
25
SV
M
DT
B
D
F
L
DoS
Atta
ckers
Traffic
Lo
ad
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
omp E
ng
IS
S
N:
20
88
-
8708
In
tr
us
io
n detec
ti
on
and preve
ntion u
sin
g
B
ay
esi
an d
eci
si
on wi
th fuzzy
lo
gic
syste
m
(
Sa
t
he
eshk
umar Sek
ar
)
1205
T
h
e
r
i
s
e
i
n
t
r
a
f
f
i
c
l
o
a
d
d
e
p
l
e
t
e
s
a
h
i
g
h
e
r
a
m
o
u
n
t
o
f
b
a
n
d
w
i
d
t
h
.
T
h
i
s
i
s
m
i
n
i
m
i
z
e
d
b
y
a
p
p
r
o
p
r
i
a
t
i
n
g
n
o
r
m
a
l
u
s
e
r
s
f
o
r
r
e
q
u
e
s
t
i
n
g
t
h
e
i
r
s
e
r
v
i
c
e
.
I
n
a
c
c
o
r
d
a
n
c
e
w
i
t
h
t
h
e
D
o
S
m
a
l
i
c
i
o
u
s
p
a
c
k
e
t
s
,
t
h
e
B
D
F
L
m
e
c
h
a
n
i
s
m
i
s
c
o
r
r
e
c
t
l
y
r
e
c
o
g
n
i
z
e
d
b
a
s
e
d
o
n
c
a
t
e
g
o
r
i
z
i
n
g
q
u
e
u
e
;
a
s
a
r
e
s
u
l
t
,
i
t
m
i
ni
m
i
z
e
s
t
h
e
t
r
a
f
f
i
c
l
o
a
d
.
H
o
w
e
v
e
r
,
t
h
e
D
T
a
n
d
S
V
M
a
l
g
o
r
i
t
h
m
d
o
e
s
n
o
t
a
c
c
u
r
a
t
e
l
y
d
e
t
e
c
t
t
h
e
m
a
l
i
c
i
o
u
s
n
o
d
e
;
t
h
u
s
,
i
t
c
r
e
a
t
e
s
m
o
r
e
t
r
a
f
f
i
c
l
o
a
d
.
Delay
is
a
vital
par
amet
e
r,
and
t
he
small
e
r
dela
y
will
su
rel
y
en
ha
nce
the
netw
ork
performa
nce.
Gen
e
rall
y,
a
rise
in
at
ta
c
k
pa
ckets
will
ulti
mate
ly
e
nh
a
nc
e
dela
y
a
nd
m
inimi
ze
the
t
hroug
hput.
The
BDFL
mecha
nism
de
te
ct
s
the
D
oS
at
ta
ck
base
d
on
the
FL
S
sy
ste
m
an
d
queue
met
hod.
T
he
de
velo
pme
nt
of
mali
ci
ou
s
pac
ke
ts
util
iz
es
gr
e
at
er
re
source
s,
an
d
it
ca
us
es
disgrace
t
o
t
he
pe
rfo
rma
nces
of
nor
mal
pac
kets.
Figure
5
c
omp
ares
delay pa
ra
mete
rs for S
V
M
, DT
, a
nd B
DF
L
mec
han
is
ms.
Figure
5
.
Dela
y of SV
M,
DT, an
d
B
DF
L a
ppr
oac
hes
base
d o
n DoS
a
tt
ack
er
The
al
le
viati
on
of
dis
honest
use
rs
with
the
a
ct
ion
of
mali
ci
ou
s
pac
kets
m
od
i
fies
to
e
nha
nce
delay.
The
dela
y
inc
r
eases
with
the
rise
in
D
oS
at
ta
cker
s
beca
us
e
of
the
par
ti
ci
pa
ti
on
of
se
ver
a
l
Do
S
at
ta
cke
r
s
who
can
handle
t
he
pack
et
s
.
I
DS
ut
il
iz
ing
SVM
a
nd
D
T
al
gorit
hms
detect
the
DoS
at
ta
cke
rs
bu
t
ca
nnot
det
ect
the
mali
ci
ou
s
pac
kets
e
ff
ic
ie
ntl
y.
The
pro
po
s
ed
meth
od
use
s
the
FL
S
a
nd
BD
to
detect
mali
ci
ou
s
pack
et
s
eff
ic
ie
ntly
, hen
ce re
du
ci
ng the
n
et
w
ork
d
el
a
y.
The
netw
ork
th
rou
ghput
ris
es
wh
il
e
acce
ssi
bili
ty
of
ba
ndwidt
h
res
ource
is
a
dequate
f
or
the
pac
ket
t
o
proce
dure.
I
n
t
he
netw
ork,
th
e
ba
ndwidt
h
util
iz
at
ion
is
a
m
ai
n
rest
raint
th
at
will
be
e
nga
ged
w
hile
the
at
ta
ck
pack
et
s
raise
.
Figure
6
e
xp
la
i
ns
t
he
net
work
th
rou
ghput
of
SVM,
D
T,
an
d
BDF
L
mecha
nisms
ba
sed
on
D
oS
at
ta
cker
s.
Figure
6
.
Th
r
ough
pu
t
of S
V
M
, DT
, a
nd B
DF
L
ap
proac
he
s b
ase
d o
n DoS
a
tt
acker
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
5
10
15
20
25
SV
M
DT
B
D
F
L
DoS
Atta
ckers
Avera
ge Dela
y (ms
)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
omp E
ng,
V
ol.
15
, No
.
1
,
Febr
uary
20
25
:
1
200
-
1
208
1206
In
t
he
pro
pose
d
BD
FL
mech
anism,
the
rapi
d
reject
io
n
of
mali
ci
ou
s
pac
kets
raises
the
ba
ndwidt
h
avail
abili
ty
by
app
l
ying
the
BD
al
gorith
m.
T
he
FLS
al
gorithm
al
s
o
separ
at
es
t
he
mali
ci
ou
s
pa
ckets
eff
ic
ie
ntly
.
Th
us
,
the
t
hro
ughput
is
raise
d
wh
e
n
c
ompare
d
with
DT
a
nd
SVM
al
gorith
ms.
But,
t
he
e
xisti
ng
DT
a
nd
SVM
al
gorithms
can
no
t
detect
t
he
mali
ci
ou
s
data
pac
ket,
an
d
da
ta
flo
w
well
s
inc
e
the
par
ti
ci
pation
of DoS att
ac
ke
r
is t
he
mai
n
re
aso
n
f
or r
ai
si
ng att
ack
pac
ket
s.
4.
CONCL
US
I
O
N
In
this
sect
io
n,
a
Ba
yesian
de
ci
sion
m
odel
with
a
f
uzzy
l
og
ic
s
ys
te
m
ba
sed
on
int
ru
si
on
detect
ion
and
pre
ven
ti
on
m
echa
nism
is
pla
nn
e
d
pa
rtic
ularly
to
gu
a
ra
ntee
sec
ur
it
y.
T
his
m
echan
is
m
detect
s
an
d
pr
e
ve
nts
i
ntrusion
li
ke
D
oS
at
ta
cks
bas
ed
on
pac
ket
an
d
fl
ow
cl
assifi
ers.
T
he
in
wards
pa
c
kets
a
re
authe
ntica
te
d
by
a
pplyin
g
F
LS
that
pr
ov
i
de
s
the
outp
ut
of
t
he
pac
ket
as
normal,
ma
li
ci
ou
s,
a
nd
doubtf
ul
pack
et
s
.
The
n, the
doubtf
ul
pa
ckets
are v
al
id
at
ed
base
d
on
q
ue
ue
band
widt
h.
T
his
mec
ha
nism
a
pp
li
ed
t
he
BD
al
gorithm
to
s
epar
at
e
the
no
r
mal
an
d
mali
ci
ou
s
pac
kets
in
the
net
wor
k.
The
simulat
io
n
res
ults
il
lustra
te
the
BDFL
mecha
ni
sm
en
ha
nces
t
he
detect
ion
ra
ti
o
an
d
mi
nim
iz
es
the
net
work
dela
y.
F
urt
hermo
re,
t
he
BDFL
appr
oach
redu
ces
the
tra
ff
ic
load
since
t
he
que
ue
ma
nagement
proce
ss
di
d
only
doubtf
ul
pack
et
s;
thu
s
,
it
reduces
the
net
work
tra
ff
ic
l
oa
d.
I
n
t
he
fu
t
ure,
we
will
util
iz
e
an
en
sem
ble
le
arn
i
ng
al
go
rithm
t
o
i
m
pro
ve
t
he
IP
S i
n
a
larg
e
-
s
cal
e sett
ing
.
REFERE
NCE
S
[1]
A.
Kh
,
I
.
G
,
P.
,
J.
K
,
“
:
h
q
,
h
,”
Cyb
ers
ecur
ity
,
v
o
l.
2
,
n
o
.
1
,
p
p
.
1
–
2
2
,
Dec.
2
0
1
9
,
d
o
i: 1
0
.11
8
6
/s4
2
4
0
0
-
019
-
0
0
3
8
-
7.
[2]
.
.
K
.
H
,
“C
h
:
,”
2
0
1
8
Inter
n
a
tio
n
a
l
Yo
u
n
g
E
n
g
in
eers
For
u
m (
YE
F
-
ECE)
,
May 2
0
1
8
,
p
p
.
1
9
–
2
4
.
d
o
i: 1
0
.1
1
0
9
/
YEF
-
ECE.
2
0
1
8
.83
6
8
9
3
3
.
[3]
A.
A
j,
“E h
h
(I
)
h
I
(I
P
AI),
”
IE
E
E
A
ccess
,
p
.
1,
2
0
2
1
,
d
o
i: 1
0
.11
0
9
/ACC
ESS.
2
0
1
9
.28
9
3
4
4
5
.
[4]
T.
Mahjab
in
,
Y.
Xiao
,
G.
Su
n
,
an
d
W
.
J
,
“A
-
of
-
serv
i
ce
attack,
p
reven
tio
n
,
an
d
m
itig
atio
n
h
q
,”
Inter
n
a
tio
n
a
l
Jo
u
r
n
a
l
o
f
Distr
ib
u
ted
S
en
so
r
Netw
o
r
ks
,
v
o
l.
1
3
,
n
o
.
1
2
,
p
p
.
1
–
1
2
,
Dec.
2
0
1
7
,
d
o
i:
1
0
.11
7
7
/1
5
5
0
1
4
7
7
1
7
7
4
1
4
6
3
.
[5]
C.
Birk
in
sh
aw,
E.
Ro
u
k
a,
an
d
.
G.
,
“I
-
d
efined
n
etwo
rkin
g
:
Defe
n
d
in
g
ag
ain
st
p
o
rt
-
scan
n
in
g
an
d
d
en
ial
-
of
-
,
”
Jo
u
rn
a
l
o
f
Netw
o
rk
a
n
d
Co
mp
u
ter
App
lica
tio
n
s
,
v
o
l.
1
3
6
,
p
p
.
7
1
–
8
5
,
Jun. 20
1
9
,
d
o
i: 10
.1
0
1
6
/j.jnca.2
0
1
9
.0
3
.00
5
.
[6]
.
E
.
A
új
,
.
.
,
J.
A.
,
“
I
’
-
of
-
,”
2
0
1
7
CHILEA
N
Co
n
feren
ce
o
n
Electrica
l,
Ele
ct
ro
n
ics
Eng
in
eering,
Info
rma
tio
n
a
n
d
Co
mmu
n
ica
ti
o
n
Tech
n
o
lo
g
ies (CHILE
CON
)
,
Oct
.
2
0
1
7
,
p
p
.
1
–
6
.
d
o
i: 10
.
1
1
0
9
/CHILE
CON.
2
0
1
7
.8
2
2
9
5
1
9
.
[7]
.
h
,
R.
h
,
.
h
,
.
Ch
,
A.
,
.
,
“
h
-
learnin
g
-
ass
isted
secu
rity
an
d
p
rivac
y
:
,”
IE
EE
I
n
tern
et
o
f
Th
in
g
s
Jo
u
rn
a
l
,
v
o
l.
9
,
n
o
.
1
,
p
p
.
2
3
6
–
2
6
0
,
Jan
.
2
0
2
2
,
d
o
i:
1
0
.11
0
9
/JI
OT.
2
0
2
1
.30
9
8
0
5
1
.
[8]
C.
-
Y.
H
,
.
W
,
Y.
,
“I
h
,”
Mu
ltimed
ia
To
o
ls
a
n
d
App
lica
tio
n
s
,
v
o
l.
8
0
,
n
o
.
1
9
,
p
p
.
2
9
6
4
3
–
2
9
6
5
6
,
Au
g
.
2
0
2
1
,
d
o
i: 10
.1
0
0
7
/
s1
1
0
4
2
-
021
-
1
1
1
0
0
-
x.
[9]
P.
F
reitas
De
Ar
a
u
jo
-
h
,
A.
J.
P
h
,
G.
K
,
.
R.
C
,
.
.
,
“A
sy
stem
for
CAN:
h
in
d
ering
cy
b
er
-
attacks
with
a
lo
w
-
,
”
IE
EE
Ac
cess
,
v
o
l.
9
,
p
p
.
1
6
6
8
5
5
–
1
6
6
8
6
9
,
2
0
2
1
,
d
o
i:
1
0
.11
0
9
/ACC
ESS.
2
0
2
1
.3
1
3
6
1
4
7
.
[10
]
.
,
.
,
.
,
A.
,
“C
h
h
h
h ,”
2
0
1
7
IE
E
E
Inter
n
a
tio
n
a
l Co
n
feren
ce on
F
u
zz
y Sys
tems (F
UZ
Z
-
IE
EE
)
,
Ju
l.
20
1
7
,
p
p
.
1
–
7
.
d
o
i: 10
.1
1
0
9
/FUZZ
-
I
EE
E
.20
1
7
.8015
4
6
4
.
[11
]
R
.
I
,
R
.
.
.
R
,
.
.
Y
,
H
.
,
“
G
h
-
,
”
I
E
E
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
I
n
t
e
l
l
i
g
e
n
t
T
r
a
n
s
p
o
r
t
a
t
i
o
n
S
y
s
t
e
m
s
,
v
o
l
.
2
3
,
n
o
.
3
,
p
p
.
1
7
2
7
–
1
7
3
6
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
T
I
T
S
.
2
0
2
0
.
3
0
2
5
6
8
5
.
[12
]
A.
h
,
“A
h
h
,”
2
0
2
0
5
th
Inter
n
a
tio
n
a
l
C
o
n
feren
ce
o
n
Commun
ica
tio
n
a
n
d
Electro
n
ics
S
ystems
(I
CC
ES)
,
Ju
n
.
2
0
2
0
,
p
p
.
4
7
6
–
4
7
9
.
d
o
i:
1
0
.11
0
9
/ICCES4
8
7
6
6
.2020
.9137
9
8
7
.
[13
]
H.
Alav
izad
eh
,
H
.
Alav
izad
eh
,
an
d
J.
Jan
g
-
J
,
“
-
l
earnin
g
b
ased
reinforce
m
en
t
learnin
g
ap
p
roach
for
n
etwo
rk
,”
Co
mp
u
ters
,
v
o
l.
1
1
,
n
o
.
3
,
p
p
.
1
–
1
9
,
Ma
r.
20
2
2
,
d
o
i: 10
.33
9
0
/co
m
p
u
ters
1
1
0
3
0
0
4
1
.
[14
]
J.
F.
Can
o
la
Garc
i
a
an
d
G
.
E.
.
,
“A
-
b
ased
in
trus
i
o
n
d
etectio
n
an
d
p
reven
tatio
n
sy
stem
for
d
etectin
g
an
d
p
reven
tin
g
den
ial
-
of
-
,
”
IE
EE
A
cc
ess
,
v
o
l.
1
0
,
p
p
.
8
3
0
4
3
–
8
3
0
6
0
,
2
0
2
2
,
d
o
i: 10
.
1
1
0
9
/ACC
ESS.
2
0
2
2
.31
9
6
6
4
2
.
[15
]
P.
J
.
C
.
.
,
“I
n
tru
sio
n
d
etectio
n
an
d
p
reven
tio
n
for
Zi
g
Bee
-
h
,”
I
EE
E
Tra
n
sa
ctio
n
s o
n
S
ma
rt Grid
,
v
o
l.
9
,
n
o
.
3
,
p
p
.
1
8
0
0
–
1
8
1
1
,
2
0
1
8
,
d
o
i: 1
0
.1
1
0
9
/TSG.
2
0
1
6
.260
0
5
8
5
.
[16
]
N.
Mou
stafa,
N
.
Ko
ron
io
tis,
M.
K
e
sh
k
,
A.
Y.
Zo
m
ay
a,
an
d
Z.
T
ari,
“
E
x
h
I
h
:
,”
IE
E
E
Co
mmu
n
ica
tio
n
s
S
u
rveys
a
n
d
Tu
t
o
ria
ls
,
v
o
l.
2
5
,
n
o
.
3
,
p
p
.
1
7
7
5
–
1
8
0
7
,
2
0
2
3
,
d
o
i: 1
0
.11
0
9
/COMST
.20
2
3
.3280
4
6
5
.
[17
]
Y.
Sh
en
,
K.
Zhen
g
,
C.
W
u
,
M.
Zh
,
X.
,
Y
.
Y
,
“A
h
h
,”
Th
e Co
mp
u
ter J
o
u
rn
a
l
,
v
o
l.
6
1
,
n
o
.
4
,
p
p
.
5
2
6
–
5
3
8
,
Ap
r.
20
1
8
,
d
o
i: 10
.10
9
3
/co
m
jn
l/b
x
x
1
0
1
.
[18
]
H.
Ben
ad
d
i,
K.
Ib
rahimi,
A.
Ben
sli
m
an
e,
an
d
J.
,
“A
( R
-
I )
I
h
,”
1
2
th
EA
I
Inter
n
a
tio
n
a
l
Co
n
ference
,
WiC
ON
2
0
1
9
,
Ta
iCh
u
n
g
,
Ta
iwa
n
,
No
vember
,
2
0
2
0
,
p
p
.
7
3
–
8
7
.
d
o
i: 10
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
5
2
9
8
8
-
8
_
7
.
[19
]
.
,
K.
h
,
.
h
,
.
P
.
,
P.
,
“
sing
h
,”
Jo
u
rn
a
l
o
f
Info
rmation
S
ecur
ity
a
n
d
App
lica
tio
n
s
,
v
o
l.
7
8
,
p
p
.
1
–
1
2
,
No
v
.
2
0
2
3
,
d
o
i:
1
0
.10
1
6
/j.jisa.20
2
3
.10
3
6
0
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
omp E
ng
IS
S
N:
20
88
-
8708
In
tr
us
io
n detec
ti
on
and preve
ntion u
sin
g
B
ay
esi
an d
eci
si
on wi
th fuzzy
lo
gic
syste
m
(
Sa
t
he
eshk
umar Sek
ar
)
1207
[20
]
K.
R
,
Y
.
Z
,
Y.
Zh
,
.
h
,
Y.
Zh
,
“
A
I :
-
b
ased
in
trus
io
n
d
et
ectio
n
m
o
d
el
for
m
u
lti
-
,”
Jo
u
rn
a
l
o
f
Bi
g
Da
ta
,
v
o
l.
1
0
,
n
o
.
1
,
p
p
.
1
–
3
0
,
Se
p
.
2
0
2
3
,
d
o
i:
1
0
.11
8
6
/s
4
0
5
3
7
-
0
2
3
-
0
0
8
1
4
-
4.
[21
]
W
.
A h
,
A
.
G
,
.
A
,
I.
.
Z
,
R.
A.
A
,
“ h
-
b
ased
d
istrib
u
te
d
d
en
ial
o
f
serv
ices
( )
,”
Inter
n
a
tio
n
a
l
Jo
u
rn
a
l
o
n
S
ema
n
tic
Web
a
n
d
Info
rma
tio
n
S
ystems
,
v
o
l.
1
9
,
n
o
.
1
,
p
p
.
1
–
1
7
,
Au
g
.
2
0
2
3
,
d
o
i: 10
.4018
/IJSW
IS.
3
2
7
2
8
0
.
[22
]
C.
.
R
h
,
R.
R
,
K.
K.
,
R
.
A
,
.
,
“
h
h
q
,”
2
0
2
3
7
th
Int
ern
a
tio
n
a
l
Co
n
fere
n
ce
o
n
Electro
n
ics
,
Co
mmu
n
ica
tio
n
a
n
d
Aero
space
Tech
n
o
lo
g
y
(I
CECA)
,
No
v
.
2
0
2
3
,
p
p
.
1
7
5
5
–
1
7
5
9
.
d
o
i: 1
0
.1
1
0
9
/ICECA5
8
5
2
9
.20
2
3
.1
0
3
9
4
9
6
2
.
[23
]
R.
Krish
n
a
Van
ak
am
a
m
id
i,
L.
Ra
m
alin
g
am
,
N
.
Ab
ira
m
i,
S
.
P
riyan
k
a,
C.
S.
Ku
m
ar
,
an
d
S.
Muru
g
an
,
“I
h
,
”
2
0
2
3
S
econ
d
Inter
n
a
tio
n
a
l
Co
n
f
eren
ce
On
S
ma
rt
Tech
n
o
lo
g
ies
For
S
ma
rt
Na
tio
n
(Sma
rtTech
Co
n
)
,
Au
g
.
2
0
2
3
,
p
p
.
6
8
3
–
6
8
7
.
d
o
i: 10
.11
0
9
/Sma
rtT
echC
o
n
5
7
5
2
6
.2
0
2
3
.1039
1
7
2
7
.
[24
]
A.
Ku
m
ar
,
K.
Ab
h
ish
ek
,
M.
R
.
Gh
alib
,
A.
Sh
a
,
X.
Ch
,
“
I
I
,”
Dig
ita
l Co
mmu
n
ica
tio
n
s a
n
d
Netwo
rks
,
v
o
l.
8
,
n
o
.
4
,
p
p
.
5
4
0
–
5
5
1
,
Au
g
.
2
0
2
2
,
do
i: 10
.10
1
6
/j.dcan
.20
2
2
.0
5
.02
7
.
[25
]
A.
J
,
P.
P
,
.
J ’
,
W
.
Zh
,
“
A
IDPS:
a
d
istrib
u
ted
m
u
lti
-
ag
en
t
in
tru
sio
n
d
etectio
n
an
d
p
reven
tio
n
sy
stem
I
,”
Clu
ster
Co
mp
u
tin
g
,
v
o
l.
2
6
,
n
o
.
1
,
p
p
.
3
6
7
–
3
8
4
,
Feb
.
2
0
2
3
,
d
o
i:
10
.10
0
7
/s
1
0
5
8
6
-
0
2
2
-
0
3
6
2
1
-
3.
[26
]
M.
J.
Ku
m
a
r,
S.
Mish
ra,
E.
G
.
Red
d
y
,
M.
Raj
m
o
h
an
,
S.
Mur
,
.
A.
h
,
“
h
,”
Ind
o
n
esia
n
J
o
u
rn
a
l
o
f
Electrica
l
Eng
in
eerin
g
a
n
d
Co
mp
u
ter
S
cien
ce
,
v
o
l.
3
4
,
n
o
.
3
,
p
p
.
1
6
6
5
–
1
6
7
3
,
Jun. 20
2
4
,
d
o
i:
1
0
.1
1
5
9
1
/ijeecs.v
3
4
.i3.p
p
1
6
6
5
-
1
6
7
3
.
[27
]
M.
A
m
ru
et
a
l.
,
“
h
,”
Inter
n
a
tio
n
a
l
Jo
u
rn
a
l
o
f
Electrica
l an
d
Co
mp
u
ter Eng
in
eering
,
v
o
l.
1
4
,
n
o
.
3
,
p
p
.
3
4
8
5
–
3
4
9
4
,
Ju
n
.
2
0
2
4
,
d
o
i: 10
.1
1
5
9
1
/i
jece.v1
4
i3
.pp3
4
8
5
-
3
4
9
4
.
[28
]
D.
Jav
ah
eri
,
S.
G
o
rgin
,
J.
-
A
.
,
.
,
“
-
b
ase
d
DDo
S
attacks
a
n
d
n
etwo
rk
tra
ff
ic
an
o
m
aly
d
et
ectio
n
h
:
,
,
,”
Info
rma
tio
n
S
cien
ces
,
v
o
l.
6
2
6
,
p
p
.
3
1
5
–
3
3
8
,
May
2
0
2
3
,
d
o
i
:
1
0
.10
1
6
/j.ins
.20
2
3
.01
.06
7
.
[29
]
S.
K.
Sek
ar
et
a
l
.
,
“R
h
h
h
h
,”
Ind
o
n
esia
n
Jo
u
rnal
o
f
Electrica
l
Eng
in
e
erin
g
a
n
d
Co
mp
u
ter
S
cien
ce
,
v
o
l.
3
7
,
n
o
.
1
,
p
.
1
3
4
,
Jan
.
2
0
2
5
,
d
o
i
:
1
0
.11
5
9
1
/ijeecs.v
3
7
.i1.p
p
1
3
4
-
142.
[30
]
P.
Rad
h
ak
rish
n
an
et
a
l.
,
“
h
,”
Ind
o
n
esia
n
Jo
u
rn
a
l
o
f
Electrica
l E
n
g
in
ee
rin
g
an
d
Co
m
p
u
ter
Scien
ce
,
v
o
l.
3
6
,
n
o
.
2
,
p
.
8
8
2
,
No
v
.
2
0
2
4
,
d
o
i: 1
0
.11
5
9
1
/ijeecs.v
3
6
.i2.p
p
8
8
2
-
891.
[31
]
A.
P
.
h
,
“
A
h
h
h
h
,”
App
lied
S
o
ft
Compu
tin
g
,
v
o
l.
1
1
6
,
p
.
1
0
8
2
9
5
,
Feb
.
2
0
2
2
,
d
o
i:
1
0
.10
1
6
/j.aso
c.20
2
1
.10
8
2
9
5
.
BIOGR
AP
HI
ES OF
A
UTH
ORS
Sath
ee
sh
k
umar
Sek
a
r
with
over
15
ye
ars
of
sea
soned
exp
ertise
in
infor
mati
on
te
chno
logy,
h
e
brings
a
wea
l
th
of
expe
ri
ence
spanning
projec
t
and
portfo
li
o
ma
nag
em
en
t,
te
chn
ic
a
l
de
li
ve
r
y,
and ma
n
age
d
servic
es.
His e
xt
ensive
b
ac
kgrou
nd
inc
lud
es
a
str
ong
foc
us on
dat
a
and
cl
oud
pro
je
c
ts,
wher
e
he
has
excel
l
ed
in
sys
tem
an
alys
is,
req
uir
emen
t
ga
the
ring
,
design,
develop
me
nt
,
te
sting
,
qu
al
it
y
assuranc
e
,
im
plementat
ion,
and
support
ac
r
oss
banki
ng,
insuranc
e
,
he
al
t
hca
re
,
and
ma
n
ufa
ct
ur
ing
dom
ai
ns.
Notab
le
s
kil
ls
include
pr
ofic
i
enc
y
in
sno
wflak
e,
Azur
e
Data
b
ric
ks
,
an
d
Azure
servi
ce
s
,
with
a
spec
ia
l
em
phasis
on
HV
R
rea
l
-
ti
m
e
rep
lication.
H
e
has
succ
essful
ly
ma
n
age
d
en
d
-
to
-
end
pro
je
c
t
pla
nn
ing,
execut
ion
,
and
ma
nag
em
en
t,
a
ligning
a
ct
iv
it
i
es
with
cor
e
busin
ess
obje
c
ti
v
es.
His
com
pe
te
n
cies
exte
nd
to
dat
a
ana
lysis
,
g
over
nance,
int
e
gra
ti
on
,
quality
,
appl
i
ca
t
ion
tu
ning,
and
s
ec
ur
it
y.
He
h
as
dem
onstra
te
d
m
aste
ry
in
dev
el
o
ping
custom
Pyt
hon
uti
l
it
i
es
for
sea
mless
dat
a
m
igra
ti
on
and
exhi
bit
s
h
ands
-
o
n
expe
r
ie
nc
e
in
Spark,
Sca
la,
Py
thon,
and
UN
IX
shell
s
cri
pt
ing.
A
standout
ac
hi
eve
m
ent
in
c
lude
s
d
esigni
ng
and
bu
il
ding
HV
R
EL
T
p
ipeline
s
for
var
iou
s
platforms,
highl
ighting
h
is
expe
r
ti
se
in
d
at
a
movemen
t.
Furtherm
ore
,
h
i
s
bac
kground
e
ncom
passes
ree
ngin
ee
ring
l
e
gac
y
appl
i
cation
s
int
o
mi
c
roservi
ce
s
on
the
Dat
a
b
ric
ks
pl
at
form
a
nd
execut
ing
succ
essful
Te
rad
at
a
to
Snow
fla
k
e
m
igrations
and
Teradata
to
GCP
BigQuery.
W
e
ll
-
ver
sed
in
Azure
DevOps
a
nd
Data
br
ic
ks
M
LOps,
he
brings
a
co
mpre
h
ensive
under
stand
ing
o
f
tool
s
and
te
chno
logi
es
in
IBM
ma
in
fra
m
e,
vision
plus,
and
I
DM
S.
He
c
an
be
con
tact
ed
at
e
ma
i
l:
sathe
eshkum
ar.
s
eka
r24@gm
ai
l
.
c
om.
Pal
aniraj
Rajid
urai
Par
vathy
is
a
proj
ec
t
m
ana
ger
at
Mphasis
Corpora
ti
on
in
Chandl
er
,
Arizo
na,
US
A.
He
h
as
over
16
y
ea
r
s
of
IT
exp
eri
e
nce
in
the
BI
a
nd
ana
ly
ti
cs
doma
in
,
with
a
foc
us
on
d
at
a
mode
li
ng
,
in
te
gr
at
ion
,
and
visua
li
z
at
ion
(Snow
fl
ake
,
Azu
re,
AWS,
GCP
,
Azure
Da
ta
Fa
ctory,
Dat
a
bri
cks,
Ta
bl
ea
u
,
Pow
e
r
BI,
Python
,
R,
SA
P
BO,
Alte
ryx,
Xc
ept
or
(RPA
)).
He
has
bee
n
re
cogni
z
ed
by
custo
me
rs
for
provid
ing
“
cust
ome
r
val
u
e
addi
ti
on
”
throug
h
per
forma
n
ce
t
uning
on
sche
du
le
.
He
r
ecei
ved
t
he
“
star
p
erf
orm
er
”
awa
rd
of
the
quar
te
r
fo
r
a
support
p
roject
from
Hexa
war
e
le
ad
ership.
Addi
ti
onally,
h
e
was
awa
rde
d
the
“
Star
Perfor
me
r
”
awa
rd
of
th
e
quar
te
r
for
a
m
igra
ti
on
proj
ec
t
from
Hexa
war
e
leade
rship.
Moreove
r,
he
wa
s
rec
ogn
iz
ed
as
t
he
“
Mos
t
va
lua
b
le
pla
y
er
”
for
a
s
upport
pro
je
c
t
f
r
om
Wi
pro
-
best
buy
a
cc
oun
t
leade
rship
.
He
al
so
re
ce
iv
ed
a
“
Feat
h
er
in
my
ca
p
”
awa
rd
for
outsta
nding
c
ontri
but
ion
to
t
he
proj
ect
busin
ess
group
hi
era
r
chy
i
te
r
at
ion
.
H
e
c
an
be
contac
t
ed
a
t
email:
pal
an
ira
jrps@g
ma
il.c
o
m.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
omp E
ng,
V
ol.
15
, No
.
1
,
Febr
uary
20
25
:
1
200
-
1
208
1208
G
o
p
a
l
K
u
m
a
r
G
u
p
t
a
h
’
m
a
t
h
e
m
a
t
i
c
s
f
r
o
m
B
a
n
a
r
a
s
H
i
nd
u
U
n
i
v
e
r
s
i
t
y
,
V
a
r
a
n
a
s
i
,
U
t
t
a
r
P
r
a
d
e
s
h
,
I
n
d
i
a
,
i
n
2
0
1
1
.
H
e
e
a
r
n
e
d
h
i
s
P
h
.
D
.
d
e
g
r
e
e
i
n
2
0
1
9
f
r
o
m
t
h
e
D
e
p
a
r
t
m
e
n
t
o
f
M
a
t
h
e
m
a
t
i
c
a
l
S
c
i
e
n
c
e
s
,
I
n
d
i
a
n
I
n
s
t
i
t
u
t
e
o
f
T
e
c
h
n
o
l
o
g
y
,
B
a
n
a
r
a
s
H
i
n
d
u
U
n
i
v
e
r
s
i
t
y
,
V
a
r
a
n
a
s
i
,
U
t
t
a
r
P
r
a
d
e
s
h
,
I
n
d
i
a
.
H
e
i
s
c
u
r
r
e
n
t
l
y
w
o
r
k
i
n
g
a
s
a
n
a
s
s
i
s
t
a
n
t
p
r
o
f
e
s
s
o
r
a
t
t
h
e
S
y
m
b
i
o
s
i
s
I
n
t
e
r
n
a
t
i
o
n
a
l
U
n
i
v
e
r
s
i
t
y
N
a
g
p
u
r
c
a
m
p
u
s
,
I
n
d
i
a
.
H
e
h
a
s
p
u
b
l
i
s
h
e
d
r
e
s
e
a
r
c
h
p
a
p
e
r
s
i
n
v
a
r
i
o
u
s
j
o
u
r
n
a
l
s
,
s
u
c
h
a
s
t
h
e
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Q
u
a
l
i
t
y
T
e
c
h
n
o
l
o
g
y
a
n
d
Q
u
a
n
t
i
t
a
t
i
v
e
M
a
n
a
g
e
m
e
n
t
,
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
O
p
e
r
a
t
i
o
n
a
l
R
e
s
e
a
r
c
h
,
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
p
p
l
i
e
d
a
n
d
C
o
m
p
u
t
a
t
i
o
n
a
l
M
a
t
h
e
m
a
t
i
c
s
,
a
n
d
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
M
a
t
h
e
m
a
t
i
c
s
i
n
O
p
e
r
a
t
i
o
n
a
l
R
e
s
e
a
r
c
h
,
a
m
o
n
g
o
t
h
e
r
s
.
H
e
c
a
n
b
e
c
o
n
t
a
c
t
e
d
a
t
e
m
a
i
l
:
g
o
p
a
l
g
u
p
t
a
.
i
i
t
b
h
u
9
0
@
g
m
a
i
l
.
c
o
m
.
Th
ir
uvenkadachari
Rajagop
al
an
is
cur
ren
tl
y
working
wit
h
Anna
Unive
rsity
,
Univer
sity
Col
lege
of
Eng
ineeri
ng
Ariya
lur
,
T
a
mi
l
Nadu
,
Ind
ia
.
He
has
two
d
ecade
s
of
ri
ch
expe
ri
ence
in
teac
hing
and
rese
arc
h.
His
ar
ea
s
of
res
ea
rch
inc
l
ude
li
ne
ar
a
lge
b
ra,
app
li
ed
al
gebr
a,
and
theoret
i
ca
l
com
put
e
r
sc
ie
n
ce
.
He
was
awa
rd
ed
a
doct
ora
l
d
egr
e
e
in
2012
by
Bhara
th
ida
san
Univer
sity,
Ti
ru
chi
rap
al
l
i,
Tamil
Nadu,
India.
H
e
is
p
assionat
e
and
has
b
ee
n
foc
used
on
co
nduct
ing
interd
isci
pli
n
ary
r
ese
arc
h.
He
ca
n
be
contac
t
ed
at
emai
l:
raj
gopa
la
nt@g
m
ai
l
.
com.
Ch
ethan
Ch
an
dra
Su
bh
as
h
Ch
and
ra
Basap
p
a
Basavaradd
i
working
as
a
n
associa
t
e
profe
ss
or
in
th
e
Depa
r
t
me
nt
of
Artifici
al
Int
el
l
ige
n
ce
a
nd
Mac
hine
L
earning
at
Don
Bosco
Instit
ute
of
Te
chno
logy,
Kumbal
agodu
,
Banga
lor
e
-
5600
74.
He
has
over
12
yea
rs
of
ric
h
exp
eri
en
ce in
t
eachi
ng
a
t
rep
ute
d
insti
tut
ions.
He
is
a
lso
an
a
c
com
pli
shed
rese
arc
her
in
the
fie
lds
of
art
i
fici
al
intelligence,
dee
p
learni
ng
,
a
nd
image
pro
ces
sing.
To
his
cr
edi
t
,
h
e
has
publi
shed
many
rese
arc
h
articles
in
well
-
r
eput
ed
j
ourna
ls
and
has
fil
ed
p
at
en
ts
as
well
.
He
ca
n
be
con
tacte
d
a
t
e
ma
il: r
addi
04@
y
ahoo.
co
m.
Kup
pan
Padma
nab
an
joi
n
ed
K
L
Univ
ersit
y
i
n
2019
upon
co
mpl
eting
h
is
Ph.
D
.
in
com
put
er
sci
enc
e
and
engi
n
e
eri
ng
.
Curr
ent
ly
,
he
serv
es
as
a
n
associ
ate
prof
essor
in
th
e
Depa
rtment
o
f
Comput
er
Scie
n
ce
and
Engi
ne
e
ring
(Honors
)
a
t
th
e
School
of
Comput
ing
,
Koneru
L
akshma
i
ah
Educ
a
ti
o
n
Foundati
on
,
K
L
Univer
sit
y,
Andhra
Pra
desh,
Ind
ia.
Speci
a
li
z
ing
in
full
-
sta
ck
d
ev
el
opm
en
t,
he
p
oss
esses
expe
rtis
e
in
va
rious
t
ec
hnolog
ie
s,
inc
ludi
ng
MER
N,
Python
,
sprin
g
m
ic
roserv
ic
es
,
and
the
.
NET
fr
a
me
work.
He
hol
ds
bac
h
el
or's
and
ma
st
er's
de
gre
es
in
co
mputer
scie
n
ce
and
engi
ne
eri
ng
fro
m
Anna
Univer
sity
in
T
am
i
l
Nadu,
Indi
a.
Hi
s
rese
arc
h
fo
cus
es
on
wire
le
ss
sensor
net
works
,
IoT,
m
ac
hin
e
l
ea
rning
,
and
dat
a
ana
ly
ti
cs
.
He
has
publ
ished
nu
me
rous
r
ese
arc
h
art
i
cles
in
r
eput
ab
le
j
ourna
ls
and
conf
ere
n
ce
s.
He can
b
e cont
a
cted
at email:
p
adma
naba
n.
k@y
ahoo.com
.
Su
b
biah
Murugan
is
an
a
djunc
t
profe
ss
or
,
Sav
ee
th
a
Scho
ol
of
Engi
n
ee
r
i
n
g
,
Savee
th
a
Instit
u
te
of
Medi
ca
l
and
Techni
c
al
Scie
nc
es,
Chen
nai
,
Ta
m
il
Nad
u,
I
ndia.
He
publi
shed
his
re
sea
rch
ar
ticle
s
i
n
ma
ny
interna
t
iona
l
and
nation
al
conf
ere
n
ce
s
a
nd
journa
ls.
His
rese
arc
h
areas
i
ncl
ud
e
ne
twork
sec
urit
y
and
ma
ch
ine
learni
ng.
He
ca
n
b
e
cont
a
ct
ed
at
smuresjur@gm
a
il
.
com
.
Evaluation Warning : The document was created with Spire.PDF for Python.