Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
6, N
o
. 1
,
Febr
u
a
r
y
201
6,
pp
. 23
5
~
24
1
I
S
SN
: 208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i
1.9
013
2
35
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
QoS Des
i
gn Cons
ideration for En
terp
ris
e
and Provider’s
Network at Ingress and Egress
Router for VoIP
P
r
ot
ocols
Ma
nju
r
K
o
l
h
ar*
,
Mo
sl
eh M Abu
a
l
h
aj
*
*
, Fai
z
a Ri
z
w
an
*
* Dept. Computer Science
and I
n
forma
tion, Prince Sattam B
i
n A
bdulaziz Univ
ers
i
ty
, Wadi Ad
Dwaser, Kingdom
of
Saudi Arabia
** Dept. of N
e
twork and Inform
ation
Secu
rity
, Faculty
of
Information
Technolog
y
Al
-Ahl
iy
y
a
Amma
n Uni
v
e
r
si
ty
,
Am
ma
n,
Jorda
n
Article Info
A
B
STRAC
T
Article histo
r
y:
Received Sep 13, 2015
Rev
i
sed
No
v
15
, 20
15
Accepte
d Dec 3, 2015
Com
p
liance wi
t
h
the Serv
ice
Leve
l Ag
reem
e
n
t (SLA) m
e
tri
c
is a m
a
jo
r
chal
lenge in a
Multiprotocol
Labe
l Switchin
g
Virtual Priva
t
e Network
(M
P
L
S
VP
N) becaus
e
m
a
ndator
y
m
ode
ls
m
u
s
t
be m
a
int
a
ined
o
n
both s
i
d
e
s
of the MPLS VPN in order to achiev
e
end-to
-en
d
service lev
e
ls. The
end-
to-
end service of an MPLS VPN can be de
gr
aded
owing to variou
s issues such
as distribu
ted d
e
nial of service (
DDo
S), and Ran
dom Early
D
e
tection
(RED)
that pr
even
ts
congestion
and
diffe
r
e
ntiates betwee
n legitimate and
illeg
itim
at
e user
traffi
c. In th
is stud
y
,
we propos
e a centralized solution th
at
uses a SLA Violation Det
ector (SLAVD) and intrusi
on detection to prevent
SLA violation.
Keyword:
MPLS
Qo
S
Vo
IP
VP
N
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Mosleh M
Abualhaj,
Dept
.
o
f
Net
w
or
k a
n
d I
n
fo
rm
at
i
on Sec
u
ri
t
y
, Facul
t
y
o
f
I
n
fo
rm
ati
on Tec
h
n
o
l
o
gy
Al-Ah
liyya Amman
Un
iv
ersity,
A
mman
,
Jo
rd
an
Em
a
il: m
.
ab
u
a
lh
aj@amm
an
u
.
ed
u.jo
1.
INTRODUCTION
Mu
ltip
ro
t
o
co
l
Lab
e
l Switch
i
ng
(MPLS)
Virt
u
a
l Pri
v
ate Net
w
ork
(VPN) tech
no
log
y
h
a
s en
ab
led
th
e
servi
ce
pr
ovi
d
e
r net
w
o
r
k a
n
d
ent
e
rp
ri
se net
w
o
r
k t
o
ag
ree
on c
o
m
m
on t
e
rm
s for t
h
e p
r
ovi
si
on
of e
n
d
-
t
o
-e
n
d
ser
v
ice levels. Th
is agr
eem
e
n
t is th
e k
e
y facto
r
in
t
h
e incr
easing
pr
ef
eren
ce fo
r
M
P
LS V
P
N
technolo
g
y
.
MPLS VPN ele
m
ents such as
custom
er
edge
(C
E) ro
ut
er a
n
d pr
o
v
i
d
e
r
ed
g
e
(PE) r
o
ut
er p
l
ay
a pi
vot
al
ro
l
e
i
n
managing the service level agreem
ents
(SLA
s) bet
w
ee
n t
h
e
ent
e
r
p
ri
se an
d pr
o
v
i
d
er
net
w
o
r
ks
. Fi
g
u
re 1 s
h
o
w
s
t
h
e M
P
LS
VP
N a
r
chi
t
ect
u
r
e
t
o
ac
hi
eve
e
n
d-t
o
-e
nd
Q
o
S
.
En
d-t
o
-e
n
d
Q
o
S i
s
achi
e
ve
d
by
a
p
pl
y
i
ng
v
a
ri
o
u
s
p
o
licies to
th
e PE and
CE [1
,
2
]
. In
th
is stu
d
y
,
we inv
e
stig
ate th
e p
o
l
i
c
ies th
at can
ach
i
ev
e th
e
b
e
st Qo
S.
Nu
m
e
rou
s
issues prev
en
t th
e
p
r
ov
id
er
from
achieving t
h
e a
g
ree
d
se
rvice l
e
vels. T
h
ese is
sues ca
n
be cla
ssified
as technical iss
u
es a
nd t
h
reats
.
The
standard
technical issue
s
are TCP st
arv
a
tio
n [3
], Rand
o
m
Early Detectio
n
(R
ED
) [
4
]
t
o
avoi
d co
n
g
est
i
o
n
,
an
d t
h
e m
i
xi
ng
o
f
TC
P
and
U
D
P
pr
ot
ocol
s
.
The
s
e t
echni
cal
i
ssu
es
cause
u
n
d
e
rp
erfo
rm
an
ce in
t
h
e
p
r
ov
id
er n
e
t
w
ork
.
Furth
e
r, th
reats affect th
e ab
ility o
f
th
e
prov
id
er
n
e
two
r
k
to
achieve
agree
m
ents. Num
e
rous
threats
ha
m
p
er the pe
r
f
o
r
m
a
n
ce of
t
h
e pr
ov
id
er
n
e
t
w
or
k an
d en
t
e
r
p
r
i
se
net
w
or
k.
Am
ong t
h
ese t
h
re
at
s, Di
st
ri
b
u
t
e
d De
ni
al
o
f
Ser
v
i
ce (D
D
o
S) i
s
co
nsi
d
er
ed t
o
be e
x
t
r
em
ely
dan
g
e
r
o
u
s
bec
a
use i
t
fl
o
o
d
s t
h
e C
E
a
nd
PE
net
w
or
ks
[5]
.
I
n
s
u
ch si
t
u
at
i
o
ns, t
h
e com
p
l
i
a
nce wi
t
h
ag
re
em
ent
s
on a
VoIP net
w
ork
becom
e
s extrem
ely
critical and challe
ngi
ng.
Hence
,
th
is stud
y prop
o
s
es ar
ch
itectu
r
e to
p
r
ov
id
e seam
less conn
ectiv
ity b
e
tween
t
h
e en
dpo
in
ts
v
i
a the serv
ice pro
v
i
d
e
r n
e
t
w
ork.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 1, Feb
r
uar
y
20
1
6
:
23
5 – 24
1
23
6
Fig
u
r
e
1
.
MPLS VPN Ar
ch
itectu
r
e
The standard
SLA agreem
e
n
t specifies the
m
a
xim
u
m
one-
w
ay
l
a
t
e
ncy
from
t
h
e
mout
h t
o
ear
,
max
i
m
u
m
j
itte
r, and
m
a
x
i
m
u
m
p
ack
et lo
ss
as 150
m
s
, 30
m
s
, an
d
1
%
, resp
ectiv
ely [6
].
In order t
o
m
eet th
is
chal
l
e
ng
e of
pr
ovi
di
n
g
ser
v
i
ce t
o
t
h
e end
poi
nt
s, we s
h
o
u
l
d
agree
on c
o
m
m
on t
e
r
m
s such as
m
a
ppi
n
g
m
odel
s
bet
w
ee
n t
h
e
en
t
e
rpri
se a
n
d
pr
ovi
der
net
w
o
r
k
s
. T
h
e m
o
st
im
po
rt
ant
t
a
s
k
i
s
t
o
assi
g
n
pri
o
ri
t
y
cl
ass of se
rv
i
ce t
o
the carriers
of real-tim
e data. He
nce,
t
h
e
V
o
IP
net
w
or
k s
h
o
u
l
d
ha
ve
t
h
e
hi
g
h
est
p
r
i
o
ri
t
y
fo
r si
gnal
i
n
g
an
d
v
i
d
e
o
traffic.
In
add
ition
,
si
gn
alin
g p
l
ays an i
m
p
o
r
tan
t
ro
le
in
a
VoIP n
e
twork
alth
oug
h
it
d
o
e
s n
o
t
u
s
e a
real-
t
i
m
e
prot
oc
ol
a
s
i
t
s
car
ri
er.
H
o
we
ve
r, i
n
or
d
e
r t
o
ac
hi
eve t
h
e SLA for the
sign
alin
g, t
h
e
PE m
u
st p
r
ov
id
e t
h
e
best service e
v
en for non-real time
traffic; furthe
r, the
traffic (ba
ndwid
th
requ
iremen
t) fo
r signalin
g
is
negl
i
g
i
b
l
e
w
h
e
n
c
o
m
p
ared
wi
t
h
t
h
e R
T
P
dat
a
fo
r a
u
di
o a
n
d
vi
de
o.
Di
f
f
er
ent
i
a
t
e
d ser
v
i
c
e (Di
ffSe
r
v),
a
not
her
ele
m
ent of an
MPLS
VPN, i
n
volve
s
the
m
a
rki
n
g and
re
m
a
r
k
i
n
g of
traff
i
c (
c
lassif
i
cation and
m
a
n
a
g
e
men
t
of
traffic)
to
p
r
ov
id
e Qo
S. In th
e
case
of
Di
ffSe
r
v,
bot
h t
h
e si
de
s o
f
a
n
M
P
LS
VP
N
are res
p
on
si
bl
e. Th
e
ent
e
r
p
ri
se
net
w
o
r
k m
u
st
ad
opt
t
h
e m
e
t
h
o
d
s an
d
pol
i
c
i
e
s appl
i
e
d by
t
h
e p
r
ovi
der
ne
t
w
o
r
k
.
Fo
r e
x
a
m
pl
e, if
expe
di
t
e
d f
o
rw
ardi
ng
(EF
)
i
s
use
d
t
o
sha
p
e t
h
e t
r
af
fi
c of
vi
deo
,
au
di
o
,
an
d si
g
n
al
i
n
g
,
i
t
sho
u
l
d
be a co
m
m
on
feature
betwee
n CE and PE route
r
s.
Traffic mixing of TCP and
UDP should not
be use
d
because it will
starve
TC
P. The
r
ef
o
r
e, si
gnal
i
n
g a
n
d real
-t
i
m
e dat
a
m
u
st
use a si
ngl
e p
r
ot
oc
ol
. Th
us,
we can co
ncl
u
de t
h
at
, t
o
achi
e
ve Q
o
S f
o
r
an
M
P
LS
V
P
N wi
t
h
V
o
IP
pr
ot
oc
ol
,
t
h
e fo
l
l
o
wi
n
g
poi
nt
s are
i
m
port
a
nt
.
Service Level Agreem
ent.
Si
gnal
i
ng shoul
d be m
a
rked as real
t
i
m
e.
Not
m
i
xi
ng of UDP and TC
P.
M
a
rki
ng an
d re
m
a
rki
ng of
Tra
ffi
c.
2.
LITERATU
R
E
SU
RVE
Y
I
n
[7
], th
e au
t
h
or
s pr
opo
se
d
e
lay
m
a
r
g
in
-b
ased
traffic fo
r th
e MPLS n
e
two
r
k
,
an
d
su
gg
est three
alg
o
rith
m
s
to
ach
iev
e
end
-
t
o
-end
QoS in
MPLS n
e
t
w
ork
s
.
I
n
[8
], th
e au
tho
r
s pr
opo
se th
at MPLS and
DiffServ a
r
e t
h
e
only elements that
provide QoS fo
r
m
u
l
t
i
m
ed
ia tr
af
f
i
c
b
y
using n
e
twor
k r
e
sou
r
ces
effectiv
ely. Furth
e
r, th
ey u
tili
ze lab
e
l switch
e
d
p
a
th
s to
m
e
asu
r
e th
e
n
e
two
r
k
state in
o
r
d
e
r to
ad
ap
t n
e
twork
co
nfigu
r
ation
s
to
ch
ang
i
ng
t
r
affic con
d
itions. In [9
],
t
h
e au
tho
r
s
p
r
o
v
e
t
h
at m
e
rg
in
g
IP
an
d
WDM
will
n
o
t
au
to
m
a
tical
ly
h
a
nd
le traffic th
at is ad
ap
tive in
n
a
tu
re. Fu
rt
h
e
r, th
ey sh
ow th
at th
eir p
r
opo
sed
n
e
t
w
ork
efficien
tly ad
ap
ts to
th
e ch
an
g
e
s in
traffic p
a
ttern
s t
h
at
di
sr
upt
i
t
s
o
p
e
r
at
i
o
n
.
They
a
ppl
y
Q
o
S e
n
d-
t
o
-e
nd
fu
nct
i
o
ns i
n
t
h
ei
r p
r
o
p
o
sed
net
w
or
k. I
n
[
1
0]
, t
h
e aut
h
or
s use pre
-
co
n
g
est
i
o
n and
n
o
t
i
f
i
cat
i
on t
o
pr
o
v
i
d
e
feedb
a
ck
about lo
ad
co
nd
ition
s
o
n
t
h
e p
a
t
h
to
th
e
b
oun
d
a
ry no
d
e
s. Fu
rt
h
e
r, th
ey u
tilize th
is in
fo
rm
atio
n
to
pr
o
pose l
i
g
ht
w
e
i
ght
adm
i
ssi
on co
nt
r
o
l
an
d f
l
ow t
e
rm
i
n
at
i
on, a
nd
do
n
o
t
use t
h
e
kn
o
w
l
e
dge
of
per
-fl
o
w
st
at
e
o
n
in
ter
i
o
r
nodes; h
e
n
ce, th
ey
avo
i
d u
s
i
n
g
Dif
f
S
er
v to
ach
i
ev
e t
h
e
Q
o
S
f
o
r
th
e pro
p
o
s
ed
n
e
two
r
k
.
In
[
1
1
]
, th
e
aut
h
ors
pr
o
p
o
s
e pr
ovi
si
oni
ng
t
echni
q
u
es f
o
r
a
m
e
sh net
w
o
r
k
,
an
d use l
i
n
k vect
or t
ech
ni
que
s t
o
ex
pl
o
r
e t
h
e
shari
ng
pote
n
tial am
ong bac
k
up
paths a
nd a
c
hieve
bandwid
th-assignm
e
nt flexibility.
In [12], the m
a
nuscript
descri
bes l
a
y
e
r
2 t
u
n
n
el
i
n
g t
e
chni
que
s s
u
ch
as M
P
LS
-ba
s
e
d
t
u
n
n
el
s t
o
es
t
a
bl
i
s
h a
n
e
n
d
-
t
o
-e
nd
V
P
N
se
rvi
ce
by
m
e
rgi
ng t
h
e servi
ces o
ffe
red by
va
ri
o
u
s
net
w
o
r
k
dom
ai
ns al
on
g t
h
e p
a
t
h
bet
w
ee
n en
d use
r
s. I
n
[1
3
]
, t
h
e
au
tho
r
s u
s
e d
e
l
a
y-b
a
sed
co
ngestio
n
d
e
tection
and
ad
m
i
ssi
o
n
con
t
ro
l for vo
ice qu
ality in
en
terp
rise
n
e
tworks.
In [1
4
]
, th
e
SLA is
u
s
ed
t
o
p
r
o
v
i
d
e
QoS
b
y
u
tilizin
g
th
e ele
m
en
ts o
f
SLA su
ch
as co
nn
ectio
n
h
o
l
d
i
ng
t
i
m
e
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Qo
S
Desi
g
n
C
onsi
d
er
at
i
o
n
f
o
r Ent
e
r
p
ri
se
a
n
d
Pr
ovi
d
e
r
’s
N
e
t
w
ork at
I
ngr
ess a
n
d
…
(
M
o
s
l
e
h M
Ab
u
a
l
h
aj
)
23
7
im
prove the routing e
fficienc
y
. In
[1
5
]
, th
e
au
tho
r
s
d
e
scribe th
e rou
ting
mechanism
of Ethernet-s
pecific load
bal
a
nce
r
s,
wh
i
c
h i
s
act
i
v
e fo
r dy
nam
i
c t
r
affi
c
de
m
a
nds. T
h
i
s
m
echani
s
m
si
gni
fi
ca
nt
l
y
red
u
ces
ove
r
p
r
ovi
si
oni
ng
, an
d re
qui
re
s onl
y
t
h
e ba
n
d
wi
dt
h p
r
ofi
l
e
s
associ
at
ed wi
t
h
SL
As at
t
h
e PE an
d C
E
. I
n
[1
6]
,
t
h
e aut
h
o
r
s
des
c
ri
be a m
e
t
hod
t
o
re
d
u
ce
net
w
or
k c
o
st
wi
t
h
i
m
prove
d ba
n
d
w
i
d
t
h
ef
fi
ci
enc
y
, an
d o
ffe
r
pr
act
i
cal
opt
i
o
ns
f
o
r
i
n
t
e
rnet
w
o
r
k
i
n
g
d
u
ri
ng
t
h
e
m
i
grat
i
on.
I
n
[1
7]
, t
h
e a
u
t
h
ors
p
r
op
ose
Vi
rt
ual
Net
w
o
r
k
-
ba
sed
Di
ff
Ser
v
/
M
PL
S II
I t
r
ans
p
ort
net
w
or
k arc
h
i
t
ect
ure f
o
r scal
abl
e
IP ser
v
i
c
e
depl
oy
m
e
nt
and e
ffi
ci
ent
ne
t
w
o
r
k
reso
u
r
ce m
a
nagem
e
nt
. In [
1
8
]
, t
h
e aut
h
o
r
s p
r
o
p
o
se a Qo
S m
a
nagem
e
nt
and c
ont
rol
sy
st
em
t
h
at
co
m
b
ines
t
w
o
services
, suc
h
as service admission control and the ra
te feedbac
k
control, and uses
the co
m
b
ination t
o
main
tain
th
e preset Qo
S p
a
rameters in
th
e
back
bon
e
n
e
tw
or
k
of
th
e
pr
ov
i
d
er. In
[
1
9
]
, the au
thor
s
d
e
scrib
e
a
n
e
w Rate Con
t
ro
l Sch
e
m
e
, R
C
S, fo
r
real-time ap
p
licatio
ns in
n
e
twork
s
with
h
i
g
h
b
it erro
r
rates. RC
S u
s
es
d
u
m
m
y
p
ack
ets to
p
r
ob
e th
e av
ailab
ility o
f
network resou
r
ces.
In
sum
m
ary
,
every
st
udy
foc
u
ses
o
n
o
n
e
o
f
t
h
e a
b
ove
m
e
nt
i
one
d a
s
pec
t
s t
o
achi
e
ve
Qo
S
on
a
n
M
P
LS V
P
N
.
Ho
we
ver
,
p
r
ev
i
ous st
udi
es
ha
ve n
o
t
u
s
ed
t
h
ese aspects to
create an
a
r
chi
t
ecture for
real
-tim
e
traffic.
Our MPLS VPN arc
h
itecture uses al
l these asp
ects in
th
e CE an
d PE n
e
two
r
k
s
to
ob
tain
end
-
t
o
-end
Qo
S.
3.
PROP
OSE
D
AR
CHITE
C
T
URE
Fi
gu
re
2 s
h
o
w
s t
h
e
pr
op
ose
d
archi
t
ect
u
r
e
f
o
r ac
hi
e
v
i
n
g Q
o
S
fo
r M
P
L
S
VP
N at
C
E
a
n
d PE
r
out
e
r
s.
We
p
e
rform
e
d
an
in
itial stu
d
y
o
n
Vo
IP pro
t
oco
l
s, an
d
d
e
termin
ed
th
at th
e
Session
In
itiati
o
n
Pro
t
o
c
o
l
(SIP) i
s
t
h
e m
o
st
sui
t
a
bl
e pr
ot
oc
ol
f
o
r
ou
r
pr
o
pose
d
a
r
chi
t
ect
u
r
e. S
I
P can
be
use
d
t
o
creat
e,
m
odi
f
y
, an
d di
sc
o
n
n
ect
an
SIP s
e
ssi
o
n
bet
w
een
t
w
o o
r
m
o
re
rem
o
t
e
par
t
i
e
s [2
0]
. F
u
rt
h
e
r, S
I
P
has
bee
n
used
i
n
vari
o
u
s c
o
m
p
ari
s
on
t
e
st
s
wi
t
h
t
h
e
I
n
t
e
r
Ast
e
ri
s
k
p
r
ot
o
c
ol
an
d
ITU
-
b
a
sed
pr
ot
oc
ol
,
H.
32
3
[2
1
-
2
2
]
. SIP
per
f
o
rm
ed b
e
t
t
e
r t
h
a
n
t
h
ese
ot
he
r p
r
ot
ocol
s
un
der
va
ri
o
u
s
net
w
or
k i
m
pairm
e
nt
s t
h
at
were i
n
duce
d
by
usi
n
g t
h
e
Net
E
m
t
ool
. SIP i
s
onl
y
a
si
gnal
i
n
g
pr
ot
o
c
ol
, an
d t
h
e
re
al
-t
im
e dat
a
t
r
ansm
i
ssi
on of
audi
o an
d
vi
de
o t
r
af
fi
c i
s
pe
r
f
o
r
m
e
d by
R
T
P o
v
er
UD
P [
2
3-
2
4
]
.
SIP m
e
di
a ne
g
o
t
i
a
t
i
on ca
n
be
per
f
o
r
m
e
d wi
t
h
t
h
e
Sessi
on
Descri
pt
i
o
n
P
r
ot
oc
ol
.
Fi
gu
re
2.
Pr
o
p
o
se
d A
r
c
h
i
t
ect
ure
The propose
d architecture
m
u
st
av
oi
d D
D
o
S at
t
acks by
us
i
ng t
h
e I
n
t
r
usi
on
Det
ect
i
on s
y
st
em
(IDS),
an
d real-tim
e med
i
a tran
sm
i
ssio
n
shou
ld co
m
p
ly with
the SLA ag
reemen
t and
ro
u
t
ers th
at h
a
v
e
capab
ilities
to
m
a
rk
a
nd re
m
a
rk traffic.
O
u
r SI
P ser
v
er is equip
p
e
d
to
prov
id
e an
au
then
ticatio
n
-
b
a
sed
serv
ice. SIP
clien
t
s
are s
o
ft
ware
-based a
n
d
hardware
-ba
s
ed IP
phones, an
d
Pu
b
lic
Serv
ice Telep
hon
e N
e
tw
or
k
(PSTN
)
p
hon
es.
Th
e
SLA
can
b
e
co
m
p
r
o
m
i
s
e
d
b
y
D
D
o
S
attack
s, an
d m
a
r
k
i
n
g and r
e
mar
k
i
n
g of
t
r
aff
i
c; ho
w
e
v
e
r,
one-
w
ay
del
a
y
sho
u
l
d
n
o
t
be co
nsi
d
e
r
ed o
w
i
n
g t
o
t
h
e sy
nchr
o
n
i
zat
i
on p
r
obl
em
and asy
m
m
e
t
r
i
c
l
i
nks [2
5]
. A
n
SLA
violation ca
nnot be veri
fied a
nd re
ctif
ied at core routers be
cause they se
rve the provi
d
er
network according to
t
h
e DS fi
el
d
o
f
Di
f
f
Ser
v
.
He
nce, i
t
i
s
m
a
ndat
o
ry
fo
r PEs
and C
E
s t
o
b
e
equi
ppe
d f
o
r
t
h
e det
ect
i
o
n
and
cor
r
ect
i
on
o
f
S
L
A vi
ol
at
i
ons
.
Ou
r p
r
op
ose
d
archi
t
ect
u
r
e u
s
es an SL
A sy
ste
m
to detect attacks that cause a
n
MPLS
VPN to
violate the
SL
A a
g
reem
ent.
Figure
3
s
h
ows
the a
r
chitecture for
SL
A vi
ol
at
i
on det
ect
or
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 1, Feb
r
uar
y
20
1
6
:
23
5 – 24
1
23
8
Fi
gu
re
3.
SL
A
Vi
ol
at
i
o
n
Det
ect
or
Our propo
sed
SLA
Vio
l
atio
n Detecto
r
(SLAVD)
w
ill
m
o
n
ito
r traffic acro
ss t
h
e MPLS VPN edg
e
r
o
u
t
er
s of
t
h
e
p
r
ov
id
er
n
e
twor
k and
en
terpr
i
se n
e
t
w
or
k.
T
h
e SL
A
violations t
h
at are
det
ected in the
SLAVD
syste
m
are:
Hig
h
e
r th
an
the SLA b
it rate.
Illeg
i
t
i
m
ate an
d
Leg
iti
m
a
te u
s
ers.
DDoS attack.
3
.
1
.
SLAVD
SLAVD assu
mes th
at real-ti
m
e traffic will u
s
e th
e
assig
n
e
d
IP and
po
rts for
VoIP. Th
e b
a
n
d
wid
t
h
is
calcu
lated
to
en
sure th
at each
u
s
er ob
tain
s
th
e ag
reed
ban
d
wi
dt
h.
SLA
V
D
i
s
resp
o
n
si
b
l
e for m
easuri
ng t
h
e
traffic and taki
ng
necessa
ry actions if a use
r
excee
ds
the bandwidth. If
the Ingre
ss router receive
s a VoIP
packet, the router copies th
e
heade
r
in orde
r to calculate the delay.
If the d
e
lay b
e
tween
th
e PE and CE is
determ
ined to be large
,
it is reporte
d
to the SL
AVD for furthe
r action.
Wh
e
n
the SLAVD recei
ves
the
in
v
e
stig
ation
rep
o
rt o
f
th
e delay
b
e
tween
th
e
ing
r
ess
and
eg
ress rou
t
ers, it
relo
ad
s t
h
e
IPtab
l
es for th
at
p
a
rticu
l
ar ro
u
t
e. In
add
itio
n, SLAVD
d
e
tect
s o
v
e
rfl
o
o
d
i
n
g
u
s
ers. Howev
e
r, it is ex
trem
el
y d
i
fficu
lt to
iden
tify
th
e v
i
cti
m
o
r
in
tru
d
e
r. SLAVD co
m
e
s to
th
e rescu
e
o
f
th
e
user
who m
a
y
be a victim
or intruder. This
case is
considere
d
t
o
be a s
p
ecial one, and this
re
port
is pr
ov
id
ed
to
th
e ID
S fo
r fu
rth
e
r
in
v
e
stig
atio
n.
The ingres
s router m
a
y receive less
ba
ndwi
dth of data
owing to a gl
obal
attack or RE
D as pa
rt of
congestion c
ontrol [26]. T
h
e
ingress
router
buffe
rs t
h
e pa
ck
et to
t
h
e greatest ex
ten
t
po
ssib
le, and
drops th
e
packet
i
f
t
h
e t
h
res
h
ol
d i
s
rea
c
hed
.
The
s
e d
r
op
pe
d pa
c
k
ets
may affect the QoS
of t
h
e PE and CE
.
When the
SLAVD learn
s
abou
t th
e RED an
d
g
l
ob
al attack
, it redu
ce
s the
traffic
width bet
w
een the PE
and CE
.
Whe
n
th
is abno
rm
alit
y ceases to
ex
i
s
t, th
e
SLAVD
resets t
h
e traffic b
e
tween
t
h
e
PE an
d CE.
3.2.
SLA P
a
r
a
meter
Me
as
urements
Th
e b
a
sic SLA
p
a
r
a
m
e
ter
s
ar
e d
e
lay, lo
ss, an
d
th
rou
ghpu
t.
W
e
classif
y
th
e tr
af
f
i
c
m
e
asu
r
em
en
t as
i
n
t
r
usi
v
e an
d
n
o
n
-
i
n
t
r
usi
v
e. A
n
i
n
t
r
usi
v
e t
r
a
f
fi
c
m
e
t
hod
req
u
i
r
es t
h
e al
t
e
ra
t
i
on o
f
t
h
e t
r
a
f
fi
c fl
o
w
bet
w
e
e
n t
h
e
PE an
d C
E
ro
ut
ers,
whe
r
eas
a no
n-i
n
t
r
usi
v
e
m
e
t
hod w
o
ul
d
req
u
i
r
e i
n
ject
i
on
of
dum
m
y
packet
s
bet
w
e
e
n t
h
e
PE an
d C
E
.
Th
ese m
e
t
hod
s i
n
t
r
o
duce a
d
di
t
i
onal
del
a
y
i
n
t
h
e sy
st
em
[27,
28]
, a
n
d he
nce
,
we
d
o
n
o
t
co
nsi
d
e
r
t
h
ese m
e
t
hods
of
del
a
y
cal
cul
a
t
i
on.
We
use
t
h
e e
quat
i
o
n
1
f
o
r
cal
cul
a
t
i
n
g
t
h
e
del
a
y
bet
w
e
e
n t
h
e
PE
an
d
C
E
.
To
tal Delay =
(Tp
CE
-
Tp
PE
)
(
1
)
Accord
ing
to
Equ
a
tio
n
1
,
Tp CE is th
e ti
me tak
e
n
b
y
th
e
p
ack
et at th
e CE rou
t
er, an
d Tp
PE is th
e
ti
m
e
tak
e
n
b
y
th
e p
a
ck
et at t
h
e PE rou
t
er. If th
e to
ta
l d
e
lay is foun
d to
b
e
m
o
re th
an
t
h
e SLA b
it
rate, the
SLAVD
will perform
an appropriate ac
tion
according to the report.
Packe
t
loss is the num
ber of pac
k
et
s sent
t
o
t
h
e i
n
g
r
ess
r
out
e
r
by
t
h
e e
g
ress
ro
ut
er
. It
c
a
n
be cal
c
u
l
a
t
e
d
fr
om
equat
i
o
n
2:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Qo
S
Desi
g
n
C
onsi
d
er
at
i
o
n
f
o
r Ent
e
r
p
ri
se
a
n
d
Pr
ovi
d
e
r
’s
N
e
t
w
ork at
I
ngr
ess a
n
d
…
(
M
o
s
l
e
h M
Ab
u
a
l
h
aj
)
23
9
Packet L
o
ss
=
Packet Se
nt
–
Packet Receive
d
(2)
Sim
ilarly, average
packet
loss
i
s
com
put
ed
f
r
o
m
equat
i
o
n
3:
A
v
er
ag
e
p
ack
et lo
ss =
(A
vg
pk
et Sen
t
– Avg Pk
et Recd
)
/
Av
g Pk
et Sen
t
(
3
)
3.3.
SLA P
a
r
a
meter
Me
as
urements
The RED and
Weighted RE
D (W
RE
D) queue m
a
nage
m
e
n
t
p
o
licy cau
se th
e
eg
r
e
ss rou
t
er
to
dr
op
p
ack
ets. A DDoS attack
in
th
e n
e
twork
at th
is ti
me
is
considere
d
to be the worst
-
case scena
r
io. The
perform
a
nce is already degra
d
ed
ow
i
ng t
o
t
h
e q
u
eue
pol
i
c
y
,
and i
t
deg
r
a
d
es f
u
rt
her be
cause o
f
t
h
e D
D
o
S
attack. H
o
we
v
e
r, o
u
r
SL
AV
D is eq
ui
pp
ed
w
ith
th
e o
p
e
n-so
ur
ce S
n
ort
I
D
S t
o
pr
ot
ect
th
e CE fro
m
th
e DDoS
attacker.
If a
n
enterprise
network nod
e is
misbehavi
ng
or is a victim
o
f
a DDoS attack, its ba
ndwi
dth i
s
m
easured;
i
f
t
h
e ba
n
d
wi
dt
h c
o
nsum
ed e
x
cee
ds t
h
e
val
u
e
i
n
th
e SLA agreemen
t, it is con
s
id
ered
to
b
e
a
v
i
cti
m
or i
n
t
r
ude
r.
Fi
gu
re
4 s
h
o
w
s t
h
e e
xpa
n
d
ed
v
e
rsi
o
n
of t
h
e p
r
o
p
o
sed
sy
st
em
. It
consi
s
t
s
of t
h
e S
n
ort
I
D
S,
a
n
d
traffic is
di
vert
ed
from
the ingress
router t
o
the SL
AV
D syste
m
. The
SL
AVD system
is use
d
t
o
a
voi
d
threats
an
d to
m
a
in
tain
th
e Qo
S m
e
t
r
ics m
e
n
tio
n
e
d in
th
e SLA ag
reem
en
t b
e
tween
th
e CE and
PE n
e
t
w
o
r
k
s
. Ou
r
pr
o
pose
d
net
w
or
k has
been i
m
pl
em
ent
e
d at
our
uni
versi
t
y
.
In o
r
de
r t
o
m
e
asure a
nd i
n
d
u
ce l
o
ss, pac
k
et
del
a
y
,
an
d thr
oug
hpu
t w
e
u
s
ed
t
h
e
Netw
or
k Em
u
l
at
o
r
, NetEm
.
Fi
gu
re 4.
Ex
pa
nde
d ve
rsi
o
n p
r
o
p
o
sed
net
w
o
r
k
arc
h
i
t
ect
ure
4.
CO
NCL
USI
O
N
Th
e arch
itecture propo
sed
in th
is
m
a
n
u
s
crip
t con
s
is
ts of
n
e
two
r
k
elem
e
n
ts su
ch
as SLVD, NetEm
,
and
IPtables
. We ha
ve strived to create a sim
p
le arch
itecture s
o
that it can be im
ple
m
ented easily, and the
Qo
S m
e
t
r
i
c
can be m
easur
ed wi
t
h
o
u
t
ad
di
t
i
onal
ef
fo
rt
s or
net
w
or
k
el
em
ent
s
. The
del
a
y
cause
d
by
t
h
e
redirecti
on
of t
r
affic from
the ingress
rou
t
er to
th
e SLAVD i
n
stead
o
f
d
i
rectly to
th
e eg
ress rou
t
er is m
i
n
i
mal.
If t
h
e i
n
g
r
ess
ro
ut
er i
s
eq
ui
ppe
d
wi
t
h
t
h
e
IPt
a
bl
es c
o
nfig
uratio
n, th
en, th
e IPtab
l
es
co
nfigu
r
ation
at th
e
SLA
VD
ca
n b
e
rem
oved. We
m
easured
t
h
e l
o
ss, del
a
y
,
and throughput
between the
PE
and CE; these
values
are found to
be adequate to ensure
n
o
n
-
v
i
o
l
a
t
i
on o
f
SLA
m
e
t
r
i
c
s. When
packet
re
or
der
i
ng an
d si
m
u
l
t
aneo
us
floo
d
i
n
g
of
p
a
ck
ets occur in
a n
e
two
r
k, th
ey will cau
se
loss,
d
e
lay, and
ch
ang
e
i
n
throug
hpu
t; ho
wev
e
r, th
ese
v
a
lu
es satisfy th
e
QoS m
e
tric
s.
In
th
e
fu
tu
re
work, we
will furth
e
r classify traffi
c acco
r
d
i
n
g
t
h
e th
reats su
ch
as v
i
ru
s, b
o
t
n
e
ts
and
DD
oS
. M
o
re-e
ver
we
t
r
y
t
o
p
o
rt
a
n
t
i
-
t
h
reat
s
m
odul
e at
SL
A
V
D
so
t
h
at
i
t
c
a
n t
a
ke
i
n
st
a
n
t
act
i
on.
REFERE
NC
ES
[1]
CISCO, “
Enterprise QoS Solutio
n Referen
ce N
e
twork Design Gu
ide
”, San
Jose:
CISCO, 2005, p
p
. 5-1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJEC
E V
o
l
.
6, No
. 1, Feb
r
uar
y
20
1
6
:
23
5 – 24
1
24
0
[2]
Haer
y
ong Lee;
Jeong
y
e
on Hwang; B
y
ungr
y
ong
Kang; K
y
oung
p
y
o Jun, "
End-to-end QoS architecture for VPNs:
MPLS VPN deployment in a
backbone n
e
two
r
k
,"
Para
lle
l Processing, 2000.
Proceedings. 20
00 Internat
iona
l
Workshops on ,
vol., no., pp
. 479
, 483
, 2000
[3]
Widmer, J.; Denda, R.; Mauve,
M., "A surv
ey
o
n
TCP-friendly
congestion con
t
r
o
l",
Network, IEEE
, vol.15, no.3
,
pp. 28
, 37
, May
2001
[4]
Baklizi, Mahmoud, Hussein Abdel-Jabe
r
,
Ahmad Adel Ab
u-Shareha, Mosl
eh M. Abualh
aj, and Sureswaran
Ramadass, "Fuzzy
logi
c
controller of gentle random early
d
e
tect
ion based on
averag
e queue length and d
e
lay
rate"
,
Internatio
nal Journal of Fuzzy Systems
, V
o
l. 16
, No
. 1
,
pp
. 9, 19 Mar
c
h 20
14.
[5]
Zargar
, S.T.; Joshi, J.;
Tipper
,
D
.
, "A Survey
of
D
e
fense
M
e
chan
is
m
s
Agains
t Dis
t
ributed
Denial of
Service (DDoS)
Flooding Attack
s".
Communicati
ons
Sur
veys
&
Tutor
i
als
,
I
EEE
, n
o
. 99
, pp
. 1
,
24
,
March 2013
[6]
Janssen, Jan, Dann
y
De Vleesch
auwer, and Guid
o H. Petit. "
Delay and distortion bounds fo
r pack
e
tized voice ca
lls
of tradit
i
onal
PS
TN quality
"
,
Pro
ceed
ings of
the
1
s
t IP Te
lephon
y
workshop (IPTEL 2000). 2000.
[7]
Ashour, Mohamed; Tho Le-Ngoc, "Delay
-
m
argin based traffic eng
i
neeri
ng for MPLS-
D
iffServ networks",
Communications and Ne
tworks, Journal
of
, vol.10
,
no
.3, pp.351
,36
1
, Sept. 2008
[8]
Anjali
, T.
; Sco
g
lio, C
.
; de Ol
iveir
a
, J.C
., "New MPLS
network management tec
hniqu
es based on adap
tive
learn
i
ng",
N
e
ural Networks, I
E
EE Transactions
on
, vol.16, no.5
,
pp.1242,1255
, S
e
pt. 2005
[9]
Androulidakis, S.;
Doukoglou, T.;
Patikis,
G.;
Kagklis, D., "Service Diff
erentia
tion and Tr
affic Engineer
ing in
IP
over WDM Networks",
Commu
nications
Magazine, IEEE
, vol. 4
6
, no
. 5
,
pp
. 52
,
59, May
2008
[10]
Menth, M.; Br
iscoe, B.; Tsou, T., "P
recon
g
es
tion notif
ication
:
new QoS sup
port for differ
e
ntiated services
IP
networks",
Com
m
unications Ma
gazine, I
E
EE
, vo
l. 50
, no
. 3
,
pp
. 9
4
, 103
, Mar
c
h 2
012
[11]
Lei Song; Jing
Zhang; Mukherjee, B., "D
y
n
amic provision
ing w
ith
availability
g
u
arantee for diff
erentiated serv
ices
in survivable m
e
sh networks",
Se
lec
t
ed Areas in Communications, IEEE Journal
on
, vol. 25, no.
3, pp. 35, 43, April
2007
[12]
Saad,
T.; Alawieh, B.; Mouf
tah,
H.T.; Gulder
, S., "Tunne
ling tech
niques for
end-to-end
VPNs: generic deplo
y
men
t
in an
optical testbed env
i
ronment",
Communicatio
n
s Magazine, IEEE
, vol. 44
, no
.
5, pp
. 124
, 132
,
May
2006
[13]
Burst,
Ke
n; Joine
r
,
L.; Grime
s
,
Ga
ry
,
"De
l
ay
Base
d C
onge
stion
De
te
ction
a
nd A
d
mission Control for Voic
e qua
lity
in en
terprise or
carrier
controlled IP Networks",
Network and S
e
rvice Management, I
EEE Transactions on
, vol
. 2
,
no. 1
,
pp
. 1
,
8
,
N
ov. 2005
[14]
Tornator
e, M.; Baruffald
i, A.; Hong
y
u
e Zhu
;
Mukherjee, B.;
Pattavin
a
, A.,
"
H
olding-Time-A
ware D
y
n
a
mic Traffic
Grooming",
Sele
cted
Ar
eas
in Co
mmuni
cations, I
EEE Journal on
, vol. 26, no. 3, p
p
. 28
, 35
, April
2008
[15]
Van Haalen
,
R.; Malhotr
a
,
R.; de-He
e
r,
A., "Optim
i
zed
routing for p
r
oviding eth
e
rn
et LAN services",
Communications Magazine, IEEE
, vo
l. 43, no. 1
1
, pp
. 158
, 164
,
Nov. 2005
[16]
Hache,
L
.
; L
i
L
i
, "Unified
contr
o
l infra
stru
ctur
e for carr
i
er n
e
tw
ork evolution",
Communications Magazine, I
E
EE
,
vol. 38
, no
. 11
, p
p
. 74
, 77
, Nov 2
000
[17]
Yu Cheng; Farh
a, R
.
;
Tizgh
a
dam, A.; M
y
ung-S
up Kim;
Hashemi, M.; Leon-G
a
rcia, A
.
; Hong, J.W.-K., "Virtu
al
network appro
a
ch to
scalable IP service d
e
plo
y
m
e
nt
and
effic
i
ent
resour
ce man
a
gement",
Communicatio
n
s
Magazine, I
E
EE
, vol. 43
, no
. 10
,
pp. 76
, 84
, Oct.
2005
[18]
Dongli Zhang;
Ionescu, D.,
"Measurem
ent and C
ontrol
of Packet L
o
ss Probabilit
y for MPLS
VPN
S
e
rvices
",
Instrumentation
and
Measurem
ent, I
EEE Transactio
ns on
, vol. 55
, n
o
. 5
,
pp
. 1587
,15
98, Oct. 2006
[19]
Ak
y
ild
iz
, I
.
F.; A
k
an, O
.
B.;
Mora
bito,
G.,
"A rat
e
contro
l sc
h
e
m
e
for adap
tive
re
al
-tim
e ap
p
lic
at
io
ns in IP ne
tworks
with loss
y
links
and long round trip times",
N
e
tworking, IEEE/ACM Transactions on
, vol. 13, n
o
. 3, pp. 554,56
7,
June 2005
[20]
Session Initiatio
n Protocol RFC
3261.
[21]
Abu-Alhaj, Mosleh M., S.K. Manjur,
R.
Sure
swa
r
a
n
,
Tat-Chee
Wa
n,
Ima
d
J.
Moh
a
ma
d,
a
nd Ahme
d M.
Ma
na
sra
h
.
"ITTP: A New Transport Protocol for VoIP Applications",
International Journal of
Innovative Computing,
Information and
Control (
I
JICIC)
, vol. 8
.
, no. 3, p
p
1–10-12026.,
March 2013
.
[22]
A
buA
lhaj, M
o
s
l
eh, e
t
a
l
. "
Multiplexing SIP app
lications voi
ce p
a
ckets between
SWVG gateways
", Proceedings
o
f
International Co
nference on
Co
mputer Engi
n
eer
ing and
Applications (ICCEA 20
09). 2009
.
[23]
Abu-Alhaj, Mosleh M., et a
l
. "
D
elta-Mul
tipl
e
xi
ng: A
Novel Technique to Im
prove
VoIP Bandwidth Utiliz
atio
n
between VoIP Gateway
s
",
Computer and Information Technolog
y (
C
IT)
,
2010 IEEE 10th Int
e
rnat
ional Confere
n
c
e
on. IEEE
, 2010
.
[24]
Abu-Alhaj, Mo
sleh,
et
al
. "M
uxCom
p-A New Archite
ctur
e
to Im
prove VoI
P
Bandwidth U
tili
zat
ion",
Futu
re
Networks, 2009
International Co
nference on
. I
E
EE
, 2009
.
[25]
Ahsan Habib, Sonia Fahm
y
,
Srinivas R.
Avasar
ala, Venkatesh
Prabhakar, Bh
ar
at Bharg
a
va
, On dete
cting s
e
rv
ic
e
violations and b
a
ndwidth th
eft in QoS network domains,
Computer Communica
tions
, Volume 2
6
, Issue 8, 20
M
a
y
2003, Pages
861
-871
[26]
Flo
y
d, Sally
;
Jacobson, V.,
"Random early
d
e
tection gatewa
y
s
f
o
r congestion
av
oidance",
N
e
two
r
king, IE
EE/
AC
M
Transactions on
, vol. 1, no. 4, pp.
397, 413, Aug 1
993.
[27]
Wei-Zhou Lu
,
Wei-Xuan Gu,
Shun-Zheng Yu, One-way
queuing delay
measurement
and its application o
n
dete
cting DDoS
att
ack
,
Journal of Network and
Computer Applications
, Volume 32, Issue 2
,
March 2009, Pages
367-376,
[28]
Hong-hua Zhao
; Ming Chen,
"Network Topolog
y
Inf
e
ren
ce
Based on Delay
Variation"
,
A
d
vanced Compu
t
er
Control, 2009
. I
C
ACC '09. Inter
national Con
f
erence on
, vo
l., no
., pp. 772,
776, 22
-24 Jan. 2009
Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE
ISS
N
:
2088-8708
Qo
S
Desi
g
n
C
onsi
d
er
at
i
o
n
f
o
r Ent
e
r
p
ri
se
a
n
d
Pr
ovi
d
e
r
’s
N
e
t
w
ork at
I
ngr
ess a
n
d
…
(
M
o
s
l
e
h M
Ab
u
a
l
h
aj
)
24
1
BIOGRAP
HI
ES OF
AUTH
ORS
Dr.
Manjur Kolhar
(m.
kolhar@ps
au.
e
du.
s
a) rece
ived his Bachelor
of Science from KUD,
INDIA
in 1999 and Master in Computer Applications
sy
stem from KUD, India in 2001
, receiv
ed PhD
degree from National Advanced IP
v6 Centre (NAV6) in Universiti Sains Malaysia (USM). I
n
2010. He has published more than 25 research pape
rs in International Journals an
d Conference
s
of high repu
te.
His research
interest in
clud
es
advanced
Computer n
e
tworks an
d security
and
cloud
computing
resource manag
e
ment.
Dr. M
o
s
l
eh M
.
Abu-Alhaj is
a senior le
cturer
in
Al-Ahli
yya Am
m
a
n Univers
i
t
y
. He receiv
e
d his
first degr
ee
in
Computer Scien
ce from Philadel
phia Univ
ersity, Jordan, in
July
2004
, master
degree in Com
puter Inform
ation
S
y
s
t
em
from
t
h
e Arab Acade
m
y
for Bank
ing
and F
i
nancia
l
Sciences, Jordan in July
2007, and doctor
a
te
degree
in Multimedia Networks Protocols from
Universiti
Sains
Malay
s
i
a
in
2
011. His r
e
sear
ch ar
ea of in
ter
e
st in
cludes Vo
IP, Multim
edi
a
Networking,
and
Congestion
Control. Apart fro
m
research
, Dr.
Mosleh M. Abu-
Alhaj
also does
cons
ultan
c
y
s
e
r
v
ices
in the abo
v
e res
earch ar
ea
s
and directs
the Cis
c
o acad
em
y te
am
at Al-
Ahliy
y
a
Amman University
.
F
a
iza R
i
zwan
c
o
m
p
leted Ba
che
l
or of S
c
i
e
nc
e
in Computer Application f
r
om Patna Women’s
College, Ind
i
a
i
n
2002. Com
p
leted Mast
er of C
o
m
puter Application from
L
.
N.
Mishra Institut
e
of Economic Development and
Social Change, Pa
tna, India in
2005. Working as Lecturer in
Computer Scien
ce
& Informatio
n depar
t
ment in
Co
llege of Arts
and Scie
nce, W
o
men in Prince
Sattam
Bin Ab
dulaz
iz Univ
ersit
y
, Wad
i
Addawasir, since Febr
uar
y
12 to
til
l Date
. I h
a
ve
inter
e
sted
in d
a
t
a
base,
software
development an
d user in
terf
ace.
Evaluation Warning : The document was created with Spire.PDF for Python.