T
E
L
KO
M
NI
K
A
,
V
ol
.
14,
N
o.
3,
S
ept
em
ber
20
16,
pp.
89
4
~
9
03
I
S
S
N
:
1
693
-
6
930
,
ac
c
r
edi
t
ed
A
b
y
D
IK
T
I,
D
e
c
r
e
e
N
o
:
58/
D
I
K
T
I
/
K
ep/
2013
D
O
I
:
10.
12928/
T
E
LK
O
M
N
I
K
A
.
v
1
4
i
3
.
3155
89
4
R
ec
ei
v
ed
N
ov
e
mber
2
8
,
20
1
5
;
R
ev
i
s
ed
May
1
9
,
2
01
6
;
A
c
c
ept
ed
J
une
4
,
201
6
Erro
r
R
esilien
t
M
ul
ti
p
a
th
V
i
de
o
D
eli
v
er
y
o
n
W
ir
eless
O
ve
r
l
a
y
N
et
w
o
r
ks
B
U
m
a M
ah
esw
ar
i
*
1
,
T
SB
S
u
d
ar
sh
an
2
D
epar
t
m
ent
of
C
o
m
put
er
S
c
i
e
nc
e
and E
ngi
n
eer
i
ng,
A
m
r
i
t
a S
c
hoo
l
of
E
ngi
neer
i
ng,
B
engal
ur
u,
A
m
ri
t
a
Vi
s
h
w
a
Vi
d
y
a
p
eet
ha
m
,
A
m
r
i
t
a
U
n
i
ve
r
si
t
y,
I
ndi
a
*
C
or
r
es
po
ndi
ng a
ut
hor
,
e
-
ma
i
l
:
um
ar
ahul
05@
g
m
ai
l
.
c
om
1
,
s
u
dar
s
h
an.
t
s
b@
g
m
ai
l
.
c
om
2
A
b
st
r
act
R
eal
t
i
m
e app
l
i
c
at
i
ons
del
i
v
er
i
ng m
ul
t
i
m
edi
a d
at
a o
v
er
w
i
r
el
e
s
s
ne
t
w
or
k
s
s
t
i
l
l
p
os
e
m
any
c
hal
l
en
ges
du
e t
o
hi
gh t
hr
ou
ghput
and
s
t
r
i
ng
ent
del
ay
r
e
qui
r
em
ent
s
.
O
v
er
l
ay
net
w
or
k
s
w
i
t
h
m
ul
t
i
pat
h
t
r
ans
m
i
s
s
i
on
i
s
t
he
pr
om
i
s
i
n
g s
o
l
ut
i
on
t
o
addr
e
s
s
t
h
e ab
ov
e
pr
ob
l
em
s
.
B
ut
i
n w
i
r
el
e
s
s
ne
t
w
or
k
s
t
h
e
m
ai
nt
ena
nc
e of
ov
er
l
ay
n
et
w
or
k
s
i
ndu
c
e add
i
t
i
onal
o
v
er
h
eads
af
f
ec
t
i
ng t
he
bul
k
y
a
nd
del
ay
s
en
s
i
t
i
v
e
del
i
v
er
y
of
m
u
l
t
i
m
edi
a d
at
a.
T
o m
i
ni
m
i
z
e t
he o
v
er
he
ad
s
,
t
h
i
s
w
or
k
i
nt
r
od
uc
es
t
he E
r
r
o
r
C
o
m
pens
at
ed D
at
a
D
i
s
t
r
i
b
ut
i
on M
odel
(
E
C
D
D
)
t
hat
ai
d
s
i
n r
ed
uc
i
ng e
nd t
o en
d del
a
y
s
a
nd o
v
er
he
ads
ar
i
s
i
ng f
r
om
pac
k
e
t
r
et
r
an
s
m
i
s
s
i
on
s
.
T
he E
C
D
D
adopt
s
m
T
r
ee
bon
e al
gor
i
t
hm
t
o i
dent
i
f
y
t
he
un
s
t
a
bl
e
w
i
r
e
l
es
s
n
ode
s
a
n
d
c
on
s
t
r
u
c
t
ov
er
l
a
y
t
r
ee.
T
he
ov
er
l
ay
t
r
e
e i
s
f
ur
t
her
s
p
l
i
t
t
o s
upp
or
t
m
ul
t
i
p
at
h t
r
an
s
m
i
s
s
i
ons
.
A
s
ub
pac
k
et
i
z
at
i
on
m
ec
ha
ni
s
m
i
s
adopt
ed
f
or
m
ul
t
i
pat
h v
i
de
o dat
a del
i
v
er
y
i
n
t
he
E
C
D
D
.
A
f
o
rw
a
rd
e
rro
r
c
or
r
e
c
t
i
o
n m
ec
ha
ni
s
m
and s
u
b
-
pac
k
et
r
et
r
an
s
m
i
s
s
i
o
n t
ec
hn
i
que
s
ado
pt
ed
i
n E
C
D
D
enab
l
es
t
o r
e
duc
e t
h
e
ov
er
h
ead
and
end
t
o
en
d
de
l
a
y
.
T
he
s
i
m
u
l
at
i
on
r
e
s
ul
t
s
pr
es
e
nt
ed
i
n
t
h
i
s
p
aper
pr
o
v
e
t
h
at
t
h
e
E
C
D
D
m
odel
pr
opo
s
ed a
c
h
i
ev
es
l
ow
er
en
d
t
o end
del
a
y
an
d out
per
f
or
m
s
t
he ex
i
s
t
i
ng m
od
el
s
i
n p
l
a
c
e.
R
et
r
ans
m
i
s
s
i
on
r
eque
s
t
s
ar
e m
i
ni
m
i
z
ed by
about
52.
2
7% and b
i
t
er
r
or
s
ar
e
r
educ
e
d by
abo
ut
23.
9
3% t
h
an S
ub
-
P
a
c
k
e
t
bas
e
d M
ul
t
i
pat
h
Loa
d D
i
s
t
r
i
b
ut
i
on.
Ke
y
w
o
rd
s
:
o
v
er
l
ay
net
w
or
k
,
v
i
deo s
t
r
eam
i
n
g,
m
ul
t
i
pat
h,
w
i
r
e
l
es
s
net
w
or
k
C
o
p
y
r
i
g
h
t
©
20
16 U
n
i
ver
si
t
a
s A
h
mad
D
ah
l
an
.
A
l
l
r
i
g
h
t
s r
eser
ved
.
1
.
I
n
tr
o
d
u
c
ti
o
n
Mul
t
i
h
op
w
i
r
el
es
s
net
w
or
k
s
ar
e c
har
ac
t
er
i
z
ed b
y
t
he
i
r
l
o
w
de
pl
o
y
m
en
t
c
os
t
s
and f
l
ex
i
bl
e
i
nf
r
as
t
r
uc
t
ur
e.
Mu
l
t
i
pl
e
a
pp
l
i
c
at
i
ons
c
an
be
ex
ec
ut
ed
o
n m
ul
t
i
ho
p
w
i
r
el
es
s
n
et
w
or
k
s
.
T
he
m
os
t
c
o
m
pl
ex
and
c
hal
l
e
ng
i
ng
of
t
hes
e ap
pl
i
c
at
i
ons
ar
e
R
eal
-
t
i
m
e
m
ul
t
i
m
edi
a bas
e
d app
l
i
c
at
i
ons
.
T
hes
e appl
i
c
at
i
o
ns
dem
and s
uf
f
i
c
i
ent
ban
d
w
i
dt
h
t
hat
v
ar
i
es
f
r
o
m
4 t
o 25Mbps
.
R
i
gi
d del
a
y
r
equi
r
em
ent
s
[
1]
p
os
e a m
aj
or
c
hal
l
e
nge t
hat
nee
ds
t
o
be ad
dr
es
s
ed.
R
es
e
ar
c
her
s
i
n [
2
-
5]
h
a
v
e
adop
t
ed
t
h
e m
ul
t
i
p
at
h
t
r
an
s
m
i
s
s
i
on s
c
hem
e t
o de
l
i
v
e
r
m
ul
t
i
m
edi
a dat
a o
n
w
i
r
el
es
s
net
w
or
k
s
.
T
he r
es
ul
t
s
pr
es
ent
ed b
y
t
hem
pr
ov
e t
h
e ef
f
ec
t
i
v
en
es
s
of
t
he m
ul
t
i
pat
h s
c
he
m
e t
o t
r
ans
m
i
t
m
ul
t
i
m
edi
a
d
at
a.
Mu
l
t
i
pat
h
di
s
t
r
i
but
i
on
m
ec
hani
s
m
s
enab
l
e
t
o
ac
h
i
e
v
e
opt
i
m
um
bandw
i
dt
h
ut
i
l
i
z
at
i
on a
nd en
ab
l
e t
o f
ul
f
i
l
l
t
he de
l
a
y
r
equ
i
r
em
ent
s
pos
ed i
n d
i
s
t
r
i
bu
t
i
o
n of
m
ul
t
i
m
edi
a or
v
i
de
o
app
l
i
c
at
i
ons
[
6,
7]
.
A
not
her
s
o
l
ut
i
on c
o
ns
i
d
er
ed t
o s
u
pp
or
t
m
ul
t
i
m
edi
a
app
l
i
c
at
i
o
ns
i
s
t
o
ado
pt
ov
er
l
a
y
net
w
or
k
s
bas
ed
on
t
h
e
ap
p
l
i
c
at
i
on
l
a
y
er
[
8
-
1
0]
.
I
n
ov
er
l
a
y
net
w
or
k
s
t
he
i
nt
er
m
edi
at
e
ho
p
no
des
ai
d
m
ul
t
i
m
edi
a
dat
a
de
l
i
v
er
y
as
t
he
y
not
o
nl
y
beh
av
e
as
r
ec
ei
v
er
s
but
a
l
s
o
as
d
at
a
f
or
w
ar
d
i
n
g
nodes
.
T
he
ho
p
no
des
,
buf
f
er
v
i
deo
dat
a
and
c
a
n
r
et
r
ans
m
i
t
i
ndep
end
ent
l
y
w
hen
pac
k
et
i
s
l
os
t
r
at
her
t
ha
n
r
equ
es
t
i
n
g
t
he
s
our
c
e
node
f
or
t
he
l
os
t
d
at
a
en
ab
l
i
n
g
t
o
r
ed
uc
e
end
t
o
end
d
el
a
y
s
and
per
h
op d
el
a
y
s
.
I
ni
t
i
a
l
l
y
r
es
ear
c
her
s
c
ons
i
der
e
d t
he us
e
of
t
r
ee s
t
r
uc
t
ur
es
f
or
v
i
de
o da
t
a
di
s
t
r
i
b
ut
i
on
[
8,
11]
w
h
i
c
h
w
as
p
r
o
v
ed
t
o
be
ef
f
i
c
i
ent
.
T
he m
aj
or
dr
aw
bac
k
of
s
uc
h s
i
n
gl
e
t
r
ee
m
e
c
hani
s
m
s
i
s
t
he
ov
er
h
ead
ar
i
s
i
ng f
r
om
t
r
ee up
d
at
i
o
n
due
t
o
t
he
d
y
n
am
i
c
nat
ur
e
of
t
he
net
w
or
k
.
T
he
ov
er
h
ead
obs
er
v
ed
,
es
p
ec
i
a
l
l
y
i
n
c
as
e
o
f
w
i
r
e
l
es
s
ne
t
w
or
k
s
due
t
o
node
m
obi
l
i
t
y
,
j
oi
n,
l
e
av
e et
c
.
,
n
egat
ed
i
t
s
pur
pos
e
of
ad
opt
i
on.
C
o
ns
i
der
i
n
g
w
i
r
el
es
s
o
v
er
l
a
y
net
w
or
k
s
t
he t
r
ee
m
ai
nt
enanc
e c
a
nno
t
be
c
o
m
pl
et
e
l
y
el
i
m
i
nat
ed
,
bu
t
b
y
r
e
duc
i
ng e
nd t
o e
n
d de
l
a
y
s
and
r
et
r
ans
m
i
s
s
i
on
ov
er
he
ad
s
an
opt
i
m
u
m
bal
anc
e
c
an
b
e
ac
hi
e
v
e
d.
T
hi
s
w
or
k
i
nt
r
oduc
es
an
E
r
r
o
r
C
om
pens
at
ed
D
at
a
D
i
s
t
r
i
but
i
on
Mod
el
(
E
C
D
D
)
f
or
m
ul
t
i
m
edi
a
dat
a
d
el
i
v
er
y
ov
er
w
i
r
e
l
es
s
net
w
or
k
s
us
i
ng m
ul
t
i
pat
h
t
r
ans
m
i
s
s
i
on.
T
he E
C
D
D c
ons
i
d
er
s
app
l
i
c
a
t
i
o
n l
a
y
er
bas
ed ov
er
l
a
y
net
w
or
k
.
T
he m
T
r
eeB
one
[
12]
m
ec
hani
s
m
i
s
adopt
ed
i
n t
he
E
C
D
D
t
o i
dent
i
f
y
t
he
u
ns
t
ab
l
e no
de
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
E
r
r
or
R
es
i
l
i
e
nt
Mul
t
i
p
at
h
V
i
deo D
el
i
v
er
y
on W
i
r
el
es
s
O
v
er
l
ay
N
et
w
or
k
s
(
B
U
ma
Mahes
w
ar
i
)
895
and
gen
er
at
e
t
he
ov
er
l
a
y
t
r
ee.
T
he t
r
ee
gen
er
at
e
d i
s
f
ur
t
her
s
pl
i
t
i
nt
o m
ul
t
i
p
l
e
s
ub
-
t
r
ees
t
o
s
uppor
t
m
ul
t
i
pat
h
v
i
d
eo
d
at
a
di
s
t
r
i
but
i
o
n.
I
n
E
C
D
D
v
i
d
eo
d
at
a
pac
k
et
s
ar
e
f
ur
t
her
s
pl
i
t
i
nt
o
s
u
b
pac
k
et
s
,
t
hat
ar
e de
l
i
v
er
e
d
ov
er
m
ul
t
i
p
l
e p
at
hs
.
E
f
f
i
c
i
ent
m
ul
t
i
m
edi
a dat
a de
l
i
v
e
r
y
on
w
i
r
el
es
s
net
w
or
k
s
c
an be
ac
hi
ev
ed
b
y
r
e
duc
i
ng t
he
end
t
o
end
del
a
y
s
.
E
n
d t
o
en
d d
el
a
y
s
ar
e de
pen
de
nt
on
t
he
per
ho
p
de
l
a
y
s
i
n
w
i
r
el
es
s
net
w
or
k
s
.
T
o
m
i
ni
m
i
z
e
t
r
a
ns
m
i
s
s
i
on
er
r
or
s
and
adher
e
t
o
t
h
e
Q
oS
r
eq
ui
r
em
ent
s
of
t
he
m
u
l
t
i
m
edi
a
d
at
a,
t
he
E
C
D
D
c
ons
i
d
er
s
t
he
us
e
of
f
or
w
ar
d
er
r
or
c
or
r
ec
t
i
on F
E
C
m
ec
hani
s
m
and ov
er
l
a
y
bas
e
d r
et
r
ans
m
i
s
s
i
on t
ec
hni
qu
es
di
s
c
us
s
ed i
n t
hi
s
w
or
k
.
T
he
no
v
e
l
t
y
of
t
he
E
C
D
D
i
s
t
he
a
dopt
i
o
n
of
t
he
m
T
r
eebone
t
o
w
i
r
el
es
s
n
et
w
or
k
s
(
pr
es
ent
l
y
i
t
has
be
en
c
on
s
i
der
ed
f
or
w
i
r
ed
net
w
or
k
s
)
.
N
ov
e
l
s
ub
t
r
ee
gen
er
at
i
o
n
t
ec
hn
i
qu
e t
o
s
uppor
t
m
ul
t
i
pa
t
h t
r
ans
m
i
s
s
i
on,
s
ub
-
p
ac
k
et
i
z
at
i
on
ba
s
ed m
ul
t
i
pat
h d
i
s
t
r
i
b
ut
i
on
[
1
4]
,
ad
opt
i
on of
F
E
C
t
o r
e
duc
e
er
r
or
s
an
d
r
et
r
ans
m
i
s
s
i
on t
ec
h
ni
qu
e
s
t
o r
e
duc
e
end
t
o
e
nd
d
el
a
y
s
.
R
es
ul
t
s
pr
es
ent
e
d i
n t
h
e l
at
t
er
s
ec
t
i
ons
pr
o
v
e ef
f
i
c
i
enc
y
of
EC
D
D
o
v
er
ex
i
s
t
i
ng
s
t
at
e
of
ar
t
s
y
s
t
em
s
.
2.
L
i
te
r
a
tu
r
e
S
u
r
v
e
y
W
an
g
,
et
a
l
.,
[
12]
pr
op
o
s
ed m
T
r
eebone o
v
er
l
a
y
c
r
eat
i
on
m
ec
hani
s
m
t
o s
uppor
t
m
ul
t
i
c
as
t
v
i
deo s
t
r
e
am
i
ng.
B
ot
h t
r
ee
an
d m
es
h ov
er
l
a
y
des
i
g
ns
ar
e c
om
bi
ne
d
t
o f
or
m
a t
r
ee
bone
w
hi
c
h
w
i
l
l
ac
t
as
a
ba
c
k
bone
t
ha
t
w
i
l
l
pus
h
a
l
l
d
a
t
a
t
o
t
h
e
o
v
er
l
a
y
n
et
w
or
k
.
I
n
t
hi
s
a
s
i
ng
l
e
s
our
c
e nod
e s
en
ds
dat
a t
o
al
l
ot
h
er
no
des
.
I
n m
T
r
eebone t
h
e o
v
er
l
a
y
f
or
m
ul
at
i
on
i
s
pr
opos
ed
t
o
ef
f
ec
t
i
v
el
y
han
dl
e t
he n
ode
s
j
oi
n and l
ea
v
e e
v
e
nt
s
i
n t
he o
v
er
l
a
y
.
T
he m
odel
m
ai
nl
y
s
t
r
u
ggl
ed t
o
obt
a
i
n
l
o
w
o
v
er
hea
d a
nd
al
s
o s
h
or
t
d
el
a
y
.
T
o t
h
e
bes
t
of
o
ur
k
now
l
e
dge
t
h
e a
dopt
i
o
n of
m
T
r
eebone
f
or
w
i
r
e
l
es
s
net
w
or
k
s
has
not
been
c
ons
i
der
e
d.
T
he
aut
hor
s
i
n
[
13]
pr
opos
e
c
o
m
bi
ned
s
c
a
l
ab
l
e V
i
de
o C
odi
ng
m
et
hod t
o i
m
pr
ov
e
t
he v
i
de
o
s
t
r
eam
i
ng
q
ua
l
i
t
y
i
n w
i
r
el
es
s
net
w
or
k
.
A
n
ov
el
s
ub
-
pac
k
et
i
z
at
i
on
bas
ed m
ul
t
i
pat
h
l
oa
d d
i
s
t
r
i
but
i
on
t
ec
hn
i
q
ue f
or
m
ul
t
i
m
edi
a
dat
a
i
s
pr
o
pos
ed
i
n
[
14
]
.
T
he t
ec
hn
i
qu
e pr
op
os
ed i
n t
hi
s
pap
er
i
s
r
ef
er
r
ed t
o as
“
S
ub
-
P
ac
k
et
bas
ed
M
ul
t
i
p
at
h
L
oad
D
i
s
t
r
i
but
i
on
f
or
R
ea
l
-
T
i
m
e
Mul
t
i
m
edi
a
T
r
a
f
f
i
c
”
(
S
P
MLD
)
.
T
he
m
odel
pr
opos
e
d i
n
[
14]
ef
f
ec
t
i
v
el
y
ac
h
i
e
v
ed
pac
k
et
del
a
y
m
i
ni
m
i
z
at
i
on b
y
a
ggr
e
gat
i
ng t
h
e m
ul
t
i
pl
e
av
a
i
l
ab
l
e
par
al
l
e
l
p
at
hs
as
a s
i
n
gl
e v
i
r
t
u
al
pat
h
f
or
t
r
ans
m
i
s
s
i
on.
A
p
ac
k
et
s
pl
i
t
t
i
ng s
t
r
a
t
eg
y
i
s
adop
t
ed
i
n
S
P
M
LD
.
T
he
D
/
M/
1
m
odel
i
nt
r
od
uc
ed
i
s
u
s
ed
t
o
an
al
y
z
e
t
h
e
pac
k
et
queu
i
n
g
de
l
a
y
and der
i
v
e t
he d
y
nam
i
c
pac
k
et
s
pl
i
t
t
i
ng r
at
i
o f
or
eac
h pat
h.
I
n
S
P
MLD
,
s
c
hed
ul
i
n
g of
s
ub
-
pac
k
et
di
s
t
r
i
b
ut
i
on w
as
ac
h
i
e
v
e
d b
y
d
ev
el
op
i
ng
i
n
de
pen
dent
s
c
hedu
l
i
n
g
al
g
or
i
t
hm
s
f
or
t
he s
our
c
e
nod
e
and
t
he
d
es
t
i
n
at
i
on
n
ode
.
T
he
r
es
ul
t
s
pr
ov
e
t
ha
t
S
P
MLD
o
ut
p
er
f
or
m
s
ex
i
s
t
ent
a
l
gor
i
t
hm
s
pr
es
ent
e
d i
n
[2
-
5]
.
T
he pe
r
f
or
m
anc
e of
S
P
ML
D
c
oul
d be i
m
pr
ov
e
d f
ur
t
her
b
y
adop
t
i
n
g er
r
or
co
r
r
ec
t
i
on m
ec
hani
s
m
s
t
o r
educ
e
pac
k
et
r
et
r
ans
m
i
s
s
i
o
n o
v
er
he
ad.
3.
Er
ro
r
C
o
m
p
e
n
s
a
te
d
D
a
ta
D
i
s
tr
i
b
u
ti
o
n
M
o
d
e
l
-
E
CDD
A
w
i
r
el
es
s
o
v
er
l
a
y
n
et
w
or
k
c
ons
i
der
ed
o
v
er
an ar
e
a
of
s
quar
e m
et
er
s
def
i
n
e
d as
W
=
(
B
,
F
)
,
wi
t
h
R
=
|
B
|
v
er
t
i
c
es
and
F
edge
s
.
Let
us
den
ot
e
S
o
as
t
he
s
o
ur
c
e
of
t
he
o
v
er
l
a
y
net
w
or
k
and
t
r
ee
on
w
hi
c
h
t
he
s
our
c
e
i
s
r
oot
ed
i
s
a
t
r
ee
-
m
es
h
denot
ed
as
T
r
e
e
j
w
hi
c
h
pas
s
es
c
opi
es
of
v
i
d
eo p
ac
k
et
j
t
o t
he no
des
i
n
B
.
A
n
y
pac
k
et
j
t
hat
i
s
t
r
ans
m
i
t
t
ed f
r
o
m
s
our
c
e
S
o
t
o
a
node
b
in
B
c
r
os
s
i
ng o
v
er
a
l
l
t
h
e p
er
ho
p de
l
a
y
s
w
h
i
c
h
i
s
t
he t
ot
a
l
d
i
s
t
anc
e
of
ov
e
r
l
a
y
pat
h
i
s
gi
v
en b
y
l
0
.
T
ot
al
end
t
o
end
del
a
y
i
s
g
i
v
en b
y
:
ED
j
(
b
,
l
0
)
=
D
i
(
b
)
l
0
i
=
1
(
1)
W
h
er
e
D
i
(
b
)
ar
e
t
h
e
p
er
h
op
d
el
a
y
s
.
T
he
pr
opos
e
d
m
odel
c
ons
i
d
er
s
t
he
i
nd
i
v
i
du
al
ho
p
de
l
a
y
s
am
ongs
t
t
he
no
des
as
i
nd
e
pend
ent
a
nd
f
ol
l
o
w
s
t
h
e
pr
obab
i
l
i
t
y
di
s
t
r
i
but
i
on
r
epr
es
ent
e
d
as
D
i
(
b
)
.
T
hus
t
he i
n
dep
end
e
nt
end
t
o en
d de
l
a
y
i
s
gi
v
e
n b
y
:
ED
j
(
b
)
=
ED
j
(
b
,
l
0
)
Q
S
j
(
b
)
=
l
0
R
−
1
i
=
1
(
2)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
8
94
–
90
3
896
w
her
e
Q
S
j
(
b
)
=
l
0
r
epr
es
ent
s
t
he
pr
o
b
abi
l
i
t
y
t
hat
t
h
e
j
th
v
i
deo
pac
k
et
i
s
l
os
t
du
e t
o
er
r
or
s
at
node
b
.
R
r
epr
es
ent
s
t
he
v
er
t
i
c
es
.
T
he v
ar
i
ab
l
e
S
j
(
b
)
i
s
t
h
e l
en
gt
h of
t
h
e pa
t
h f
r
om
s
our
c
e
S
o
t
o
a no
de
b
in
T
r
e
e
j
.
Let
us
c
ons
i
d
er
r
i
nd
epe
n
dent
a
nd
i
de
n
t
i
c
a
l
l
y
d
i
s
t
r
i
but
e
d
r
and
om
v
ar
i
abl
es
D
(
d
=
1,2,3…
r
, r
>0
h
av
i
n
g
a
m
a
r
gi
na
l
pr
o
bab
i
l
i
t
y
de
ns
i
t
y
f
u
nc
t
i
on
(
p
df
)
r
epr
es
ent
ed
as
x
i
y
and
al
s
o
j
oi
nt
p
df
as
x
(
d
1
,
…
…
…
…
…
.
d
n
)
s
uc
h t
hat
F
(
e
α
D
r
)
<
∞
.
α
r
epr
es
ent
s
a v
ar
i
abl
e and
Q
r
epr
es
ent
s
pr
o
bab
i
l
i
t
y
.
T
he
av
er
a
ge
of
D
c
ons
i
der
i
ng
d
>
[
D
]
c
an be
d
ef
i
ned as
:
Q
∑
D
k
r
k
=
1
r
≥
d
≤
e
−
r
(
d
)
(
3)
W
h
er
e
J
(
d
)
i
s
t
he r
at
e f
unc
t
i
o
n
gi
v
en
as
:
(
d
)
=
ma
x
α
>
0
(
α
d
)
−
ln
F
(
e
α
D
)
(
4)
T
he f
unc
t
i
on
(
d
)
i
s
c
onv
ex
f
or
t
he r
a
ndom
v
ar
i
ab
l
es
w
h
i
c
h i
nc
r
eas
es
dep
end
i
ng
o
n
(
F
[
D
]
,
∞
)
and
(
F
[
D
]
)
=
0
bas
ed on
[
15]
.
F
or
po
s
i
t
i
v
e r
and
om
v
ar
i
ab
l
es
D
wi
t
h
F
[
D
]
≥
0
,
t
hen t
he
der
i
v
at
i
v
e
∂
(
d
)
∂
x
|
d
0
≥
(
d
0
)
d
0
⁄
f
or
al
l
d
0
>
[
D
]
.
B
as
ed o
n
[
15]
f
or
d
>
[
]
and
c
>
1
,
w
e c
an
wr
i
t
e
:
(
cd
)
≥
(
(
d
)
+
(
cd
−
d
)
(
d
)
d
⁄
)
=
c
(
d
)
(
5)
B
as
ed on
(
cd
)
w
e c
an
def
i
n
e pr
obab
i
l
i
t
y
as
:
F
or
end
t
o e
nd
del
a
y
ED
j
(
b
)
bas
ed
on
E
q
u
at
i
on
(6
),
E
q
uat
i
on
(
3)
and
r
=
u
n
ity
w
e g
et
:
E
qu
at
i
on
(
8)
gi
v
es
t
h
e pr
o
b
abi
l
i
t
y
of
l
os
i
n
g da
t
a
i
n t
he
net
w
or
k
bas
ed
on d
el
a
y
.
T
he
v
i
de
o
da
t
a
ar
e
l
ar
g
e
a
nd
ar
e
m
odel
e
d
us
i
ng
h
ea
v
y
t
ai
l
e
d
di
s
t
r
i
but
i
o
ns
[
16,
17]
L
et
us
c
ons
i
der
i
n
dep
end
ent
and i
de
nt
i
c
a
l
l
y
di
s
t
r
i
but
ed r
andom
v
ar
i
a
bl
es
D
j
w
it
h
t
h
e
d
is
t
r
i
b
ut
i
on
f
unc
t
i
on
E
(
y
)
<
1
,
∀
>
0
and
D
j
:
j
∈
R
.
Let
t
h
e t
a
i
l
of
E
be
E
(
y
)
=
1
−
E
(
y
)
and b
y
E
l
0
∗
(
y
)
=
Q
Y
1
+
⋯
+
Y
l
0
>
is
t
h
e
t
a
il
o
f
t
h
e
l
0
t
he par
t
of
E
.
F
or
s
ub ex
po
nen
t
i
a
l
d
i
s
t
r
i
b
ut
i
o
n
:
T
he
t
ot
al
s
um
ex
ponent
i
a
l
di
s
t
r
i
but
i
ons
of
t
he
v
i
d
e
o
pac
k
et
s
t
r
ans
m
i
t
t
ed
ov
er
t
he
o
v
er
l
a
y
net
w
or
k
W
i
s
c
har
ac
t
er
i
z
ed
b
y
:
Q
∑
D
k
r
k
=
1
r
≥
cd
≤
e
−
r
(
cd
)
≤
e
−
cr
(
d
)
=
e
−
r
(
d
)
c
(
6)
Q
ED
j
(
b
)
≥
c
≤
e
−
(
c
)
(
7)
Q
ED
j
(
b
)
≥
ca
≤
e
−
(
ca
)
≤
e
−
(
a
)
c
(
8)
E
l
0
∗
(
y
)
E
(
y
)
≈
l
0
(
9)
q
l
0
∞
l
0
=
0
(
1
+
ω
)
l
0
<
∞
(
10)
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
E
r
r
or
R
es
i
l
i
e
nt
Mul
t
i
p
at
h
V
i
deo D
el
i
v
er
y
on W
i
r
el
es
s
O
v
er
l
ay
N
et
w
or
k
s
(
B
U
ma
Mahes
w
ar
i
)
897
w
her
e
t
he
v
ar
i
ab
l
e
ω
>
0
,
q
l
0
r
epr
es
ent
s
t
h
e
pr
o
bab
i
l
i
t
y
s
uc
h
t
h
at
Q
S
j
(
b
)
=
l
0
.
Let
us
def
i
ne
a
f
unc
t
i
on
W
(
y
)
=
q
l
0
∞
l
0
=
0
E
l
0
∗
(
y
)
.
A
pp
l
y
i
n
g
a
s
um
mat
i
o
n
r
ang
i
n
g
f
r
o
m
0
t
o
∞
in
E
q
uat
i
on
9
w
e
get
:
W
(
y
)
E
(
y
)
≈
∑
l
0
q
l
0
∞
l
0
=
0
(
11)
As
q
l
0
r
epr
es
ent
s
t
h
e pr
ob
abi
l
i
t
y
s
uc
h t
hat
Q
S
j
(
b
)
=
l
0
,
s
i
m
ila
r
i
l
y
W
(
y
)
=
Q
ED
j
(
b
)
>
and
a
ls
o
E
(
y
)
=
Q
D
j
(
b
)
>
.
T
her
ef
or
e
:
Q
ED
j
(
b
)
>
≈
F
S
j
(
b
)
Q
D
j
(
b
)
>
(
12)
B
as
ed
on
E
q
u
at
i
on
12
i
t
c
a
n
be
obs
er
v
ed
t
hat
as
t
he
per
ho
p
de
l
a
y
i
nc
r
eas
es
t
h
e
pr
ob
abi
l
i
t
y
of
t
he m
i
s
s
i
ng pac
k
et
s
al
s
o i
nc
r
eas
es
.
S
i
m
i
l
ar
l
y
a
s
t
he per
hop
del
a
y
s
dec
r
eas
es
s
ub
ex
pone
nt
i
al
l
y
,
t
he pr
o
bab
i
l
i
t
y
of
pac
k
et
s
l
os
t
i
n t
he ov
er
l
a
y
n
et
w
or
k
al
s
o r
educ
es
s
ub
ex
pone
nt
i
a
lly
.
F
i
gur
e
1
.
S
am
pl
e W
i
r
el
es
s
N
et
w
or
k
F
i
gur
e
2
.
O
v
er
l
a
y
F
or
m
ul
at
i
on us
i
ng
m
T
r
eeB
one
F
i
gur
e
3
.
O
v
e
r
l
a
y
S
u
b
-
t
r
ee
F
or
m
ul
at
i
on t
o
S
upp
or
t
M
ul
t
i
pa
t
h T
r
ans
m
i
s
s
i
ons
4.
P
r
o
p
o
s
e
d
E
C
D
D
M
o
d
e
l
W
e
di
s
c
us
s
ed
i
n
pr
ev
i
o
us
s
ec
t
i
on
,
r
ed
uc
i
ng
p
er
-
hop
del
a
y
r
ed
uc
es
t
he
m
i
s
s
i
ng
pac
k
et
pr
oba
bi
l
i
t
y
.
C
ons
i
der
a
w
i
r
e
l
es
s
ne
t
w
or
k
as
s
ho
w
n
i
n
F
i
gur
e
1.
I
t
i
s
as
s
um
ed
t
ha
t
eac
h
n
ode
i
n
t
he
w
i
r
e
l
es
s
ne
t
w
or
k
hav
e eq
ua
l
b
and
w
i
dt
h c
a
pa
c
i
t
y
c
a
pab
l
e
of
s
upp
or
t
i
n
g l
ar
ge
v
i
de
o
t
r
ans
m
i
s
s
i
on s
t
r
ea
m
s
.
T
he E
C
D
D
m
odel
pr
opos
e
d
c
ons
i
der
s
t
he m
T
r
eebon
e al
gor
i
t
hm
f
o
r
det
ec
t
i
ng
t
he
s
t
ab
l
e,
uns
t
a
b
l
e n
od
es
an
d o
v
er
l
a
y
f
or
m
at
i
on
.
T
he s
our
c
e n
od
e (
r
epr
es
ent
e
d as
S
)
i
s
c
ons
i
der
ed
as
t
h
e
r
oot
n
ode
a
nd
t
he
d
es
t
i
n
at
i
o
n
(
r
e
pr
es
ent
e
d
as
D
)
i
s
c
ons
i
de
r
ed
as
t
he
l
e
af
node
i
n
t
h
e
ov
er
l
a
y
c
o
ns
t
r
uc
t
i
on
.
T
he
E
C
D
D
m
odel
p
r
opos
ed
i
n
t
h
i
s
pa
per
c
o
ns
i
der
s
m
ul
t
i
pat
h
t
r
ans
m
i
s
s
i
ons
.
T
he ov
er
l
a
y
c
ons
t
r
uc
t
ed
us
i
ng
t
h
e m
T
r
e
ebon
e a
l
g
or
i
t
hm
f
or
t
he
w
i
r
el
es
s
t
o
po
l
og
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
8
94
–
90
3
898
i
s
as
s
ho
w
n
i
n F
i
gur
e
2
f
or
t
he
s
am
pl
e
w
i
r
e
l
es
s
net
w
or
k
o
f
F
i
gur
e.
1
a
nd
i
s
den
ot
ed
as
T
r
ee.
T
he ov
er
l
a
y
n
et
w
or
k
gener
at
ed
i
s
f
ur
t
her
s
p
l
i
t
t
o c
o
ns
t
r
uc
t
s
ubt
r
ees
as
s
h
o
w
n
i
n F
i
gur
e
3.
N
od
es
l
ab
el
e
d S
,
a,
b,
c
an
d D
f
or
m
t
he
f
i
r
s
t
s
ub t
r
ee i
.
e.
T
r
ee
1
.
N
odes
l
ab
el
ed
S
,
e,
f
,
g an
d D
c
ons
t
i
t
ut
e t
he s
ec
ond s
ub t
r
ee T
r
ee
2
.
T
he nodes
S
,
h,
i
,
j
,
k
and D
f
or
m
t
he t
hi
r
d s
ub t
r
ee T
r
ee
3
.
Let
us
c
ons
i
der
o
v
er
l
a
y
n
e
t
w
or
k
r
epr
es
ent
ed
as
T
r
ee
co
n
si
st
s
o
f
R
w
i
r
el
es
s
nodes
.
T
o
r
educ
e
t
he
hop
de
l
a
y
s
,
E
C
D
D
ado
pt
s
a
m
ul
t
i
p
at
h
di
s
t
r
i
but
i
o
n
m
odel
.
T
he
ov
er
l
a
y
net
w
or
k
T
r
ee
f
or
m
ed
i
s
s
pl
i
t
t
o
a n
um
ber
of
s
ub t
r
ees
.
T
he
k
th
s
ubt
r
ee
i
s
de
not
e
d as
T
r
ee
k
.
T
he end t
o end d
el
a
y
E
D
obs
er
v
e
d per
pa
t
h i
s
de
pe
nden
t
on t
h
e s
i
z
e of
t
he s
ubt
r
ee
f
or
m
ed.
Let
T
l
k
(
b
)
r
epr
es
ent
t
he l
ev
e
l
of
a nod
e
b
i
n s
ub t
r
ee
k
.
T
he l
e
ngt
h
N
L
of
a
s
ub t
r
ee c
an
be c
om
put
ed us
i
ng
:
N
L
=
ma
x
k
,
b
T
l
k
(
b
)
(
13)
T
he pr
obab
i
l
i
t
y
t
hat
t
he
b
th
nod
e w
i
l
l
r
ec
ei
v
e
C
b
(
a
)
v
i
de
o s
ub
-
pa
c
k
et
out
of
t
he t
ot
al
r
v
i
deo s
ub
-
pac
k
et
di
s
t
r
i
bu
t
ed
i
n
a m
ul
t
i
pat
h
net
w
or
k
of
R
ov
er
l
a
y
no
des
i
s
d
ef
i
ned
as
:
Q
V
P
(
a
)
=
F
∑
C
b
(
a
)
r
R
b
(
14)
S
i
m
i
l
ar
l
y
m
i
s
s
i
ng pr
ob
ab
i
l
i
t
y
c
an
be
def
i
ne
d as
:
Q
ED
j
(
b
)
≥
a
=
1
−
Q
V
P
(
a
)
=
1
−
F
∑
C
b
(
a
)
r
R
b
(
15)
Let
Q
k
,
l
0
V
P
R
(
T
)
r
epr
es
ent
t
h
e
pr
ob
ab
i
l
i
t
y
t
ha
t
an
o
v
er
l
a
y
no
de
at
l
0
l
ev
e
l
i
n
t
he
T
r
e
e
k
w
ou
l
d
r
ec
e
i
v
e
a
v
i
de
o
s
ub
-
pac
k
et
(
pos
t
t
he
t
r
ans
m
i
s
s
i
on c
om
m
e
nc
em
ent
f
r
o
m
t
he
s
our
c
e
node)
w
i
t
h
i
n
T
s
ec
onds
.
Q
k
,
l
0
V
P
R
(
T
)
i
s
def
i
ne
d as
:
Q
k
,
l
0
V
P
R
(
T
)
=
Q
BR
k
,
l
0
(
T
)
=
1
,
(
16)
W
h
er
e
BR
k
,
l
0
(
T
)
i
s
a
b
i
n
ar
y
r
and
o
m
v
ar
i
abl
e.
T
he pr
o
ba
bi
l
i
t
y
k
,
l
0
(
T
)
t
hat
t
he
e
v
e
nt
ℰ
r
s
w
ou
l
d
oc
c
ur
w
i
t
h
i
n
T
s
ec
onds
i
s
g
i
v
en b
y
:
∂
k
,
l
0
(
T
)
∂
T
=
∫
∂
Q
k
,
l
0
−
1
VP
x
(
T
−
b
)
∂
T
x
r
s
T
0
(
b
)
d
b
(
17)
W
h
er
e
x
r
s
r
epr
es
ent
s
t
h
e pr
o
bab
i
l
i
t
y
dens
i
t
y
f
unc
t
i
on of
t
he t
i
m
e i
ns
t
anc
e
at
w
h
i
c
h t
he e
v
e
nt
ℰ
r
s
oc
c
ur
s
and
∫
x
r
s
(
T
)
d
T
∞
0
=
1
−
q
.
I
n
E
qua
t
i
o
n
17
,
Q
k
,
l
0
−
1
V
P
x
(
T
−
b
)
r
epr
es
ent
s
t
h
e
pr
ob
abi
l
i
t
y
t
hat
t
he pr
e
v
i
ous
hop
nod
e i
.
e
.
at
t
he l
ev
el
l
0
−
1
i
n t
he s
ub
T
r
e
e
k
h
as
t
he v
i
de
o s
ub
-
pac
k
et
.
T
he
pr
oba
bi
l
i
t
y
de
ns
i
t
y
f
unc
t
i
on
k
,
l
0
(
T
)
of
t
he t
i
m
e el
aps
ed t
o
d
et
ec
t
t
he
oc
c
ur
r
enc
e of
t
he s
ub
-
pac
k
et
l
os
s
ev
e
nt
ℰ
r
f
i
s
def
i
n
e
d as
:
k
,
l
0
(
T
)
=
∫
∂
Q
k
,
l
0
−
1
VP
x
(
T
−
b
)
∂
T
T
0
x
r
f
(
b
)
d
b
(
18)
W
he
r
e
x
r
f
r
epr
es
ent
s
t
he
pr
o
bab
i
l
i
t
y
d
ens
i
t
y
f
unc
t
i
on
of
t
he
t
i
m
e i
ns
t
anc
e
at
w
h
i
c
h t
h
e
ev
ent
ℰ
r
f
oc
c
ur
s
and
∫
x
r
f
(
T
)
d
T
∞
0
=
q
.
A
n ov
er
l
a
y
no
de
at
l
0
l
ev
el
i
n t
he
T
r
ee
k
,
o
n det
ec
t
i
n
g
a
l
os
t
or
i
r
r
ec
ov
er
abl
e er
r
or
s
ub
-
pac
k
et
s
i
ni
t
i
at
es
a r
et
r
ans
m
i
s
s
i
on r
eq
ues
t
t
o t
he pr
e
v
i
ous
hop nod
e or
nodes
bas
ed on
t
he av
ai
l
ab
i
l
i
t
y
of
t
he s
u
b
-
pac
k
et
and
i
s
gi
v
en
as
:
ℂ
k
,
l
0
(
T
)
=
k
,
l
0
(
T
)
+
k
,
l
0
(
T
)
+
k
,
l
0
(
T
)
(
19)
W
h
er
e
k
,
l
0
(
T
)
t
he pr
ob
abi
l
i
t
y
t
h
a
t
t
he
no
de
det
ec
t
s
a s
u
b
-
pac
k
et
l
os
s
f
r
o
m
i
t
s
pr
ev
i
o
us
hop
node
,
k
,
l
0
(
T
)
i
s
t
he
pr
ob
abi
l
i
t
y
w
hen
t
h
e
r
et
r
ans
m
i
s
s
i
on
f
r
o
m
t
he
pr
ev
i
ous
ho
p
nod
e/
n
odes
i
n
t
he
T
r
e
e
k
i
s
uns
uc
c
es
s
f
ul
and
k
,
l
0
(
T
)
i
s
t
he pr
o
bab
i
l
i
t
y
w
hen
t
he
r
et
r
ans
m
i
s
s
i
on r
eques
t
ha
s
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
E
r
r
or
R
es
i
l
i
e
nt
Mul
t
i
p
at
h
V
i
deo D
el
i
v
er
y
on W
i
r
el
es
s
O
v
er
l
ay
N
et
w
or
k
s
(
B
U
ma
Mahes
w
ar
i
)
899
been
l
os
t
d
ue t
o c
h
ann
el
noi
s
e.
T
he
pr
ob
abi
l
i
t
y
k
,
l
0
(
T
)
t
hat
t
he
e
v
ent
ℰ
t
r
s
w
o
ul
d oc
c
ur
w
i
t
hi
n
T
s
ec
onds
i
s
:
W
h
er
e
x
t
r
s
r
epr
es
ent
s
t
he
pr
o
bab
i
l
i
t
y
de
ns
i
t
y
f
unc
t
i
o
n of
t
he t
i
m
e i
ns
t
anc
e
at
w
h
i
c
h t
he
ev
ent
ℰ
t
r
s
oc
c
ur
s
and
∫
x
t
r
s
(
T
)
d
T
∞
0
=
1
−
q
.
T
he pr
o
bab
i
l
i
t
y
k
,
l
0
(
T
)
th
a
t
th
e
e
v
e
n
t
ℰ
t
r
r
(
i
.
e.
r
ec
ei
v
i
n
g
t
he
r
et
r
ans
m
i
s
s
i
on
s
ub
-
pac
k
et
s
uc
c
es
s
f
ul
l
y
f
r
o
m
t
he
node
at
l
0
∗
le
v
e
l
in
T
r
e
e
k
)
of
w
ou
l
d
oc
c
ur
w
it
h
in
T
s
ec
onds
i
s
:
W
h
er
e
x
t
r
r
r
epr
es
ent
s
t
he
pr
o
bab
i
l
i
t
y
de
ns
i
t
y
f
unc
t
i
o
n of
t
he t
i
m
e i
ns
t
anc
e
at
w
hi
c
h t
he
e
v
ent
ℰ
t
r
r
oc
c
ur
s
and
∫
x
t
r
r
(
T
)
d
T
∞
0
=
1
−
q
.
T
he pr
obab
i
l
i
t
y
dens
i
t
y
f
unc
t
i
on
k
,
l
0
(
T
)
of
t
he t
i
m
e el
a
ps
ed
t
o d
et
ec
t
t
h
e
oc
c
ur
r
enc
e of
t
he r
et
r
ans
m
i
s
s
i
on of
d
at
a
s
ub
-
pac
k
et
una
v
a
i
l
ab
i
l
i
t
y
w
i
t
h t
he
no
de
at
l
0
∗
le
v
e
l i
n
T
r
e
e
k
i
s
d
ef
i
ned
as
.
T
he pr
obab
i
l
i
t
y
k
,
l
0
(
T
)
t
hat
t
he e
v
ent
ℰ
tl
w
oul
d oc
c
ur
w
i
t
h
i
n
T
s
ec
onds
i
s
:
W
h
er
e
x
tl
r
epr
es
en
t
s
t
he pr
o
bab
i
l
i
t
y
dens
i
t
y
f
unc
t
i
on
of
t
he t
i
m
e i
ns
t
anc
e at
w
hi
c
h t
he
ev
ent
ℰ
tl
oc
c
ur
s
and
∫
x
tl
(
T
)
d
T
∞
0
=
1
−
(
1
−
q
)
2
.
T
he pr
oba
bi
l
i
t
y
t
hat
t
h
e o
v
er
l
a
y
n
ode
at
l
0
l
ev
el
i
n
t
h
e
T
r
ee
k
w
o
ul
d r
ec
ei
v
e
a s
ub
-
pac
k
et
di
r
ec
t
l
y
or
t
hr
oug
h r
et
r
ans
m
i
s
s
i
ons
i
s
def
i
ne
d a
s
:
B
as
ed
o
n
t
he
pac
k
et
r
ec
ei
v
i
ng
pr
ob
abi
l
i
t
y
g
i
v
en
b
y
E
quat
i
on
2
4
,
t
he
pr
o
bab
i
l
i
t
y
t
hat
an
o
v
er
l
a
y
node
at
l
0
l
e
v
e
l
i
n
t
he
T
r
ee
k
w
i
l
l
be
abl
e
t
o
r
ec
ons
t
r
uc
t
t
he
r
ec
ei
v
ed
er
r
or
s
ub
-
p
a
ck
e
t
f
r
o
m
i
t
s
pr
ev
i
ous
hop
nod
e/
n
odes
u
s
i
ng t
he
F
EC
(
r
,
D
EC
)
f
unc
t
i
on
i
s
def
i
n
ed
as
:
W
h
er
e
l
0
j
=
l
0
f
or
ov
er
l
a
y
n
od
es
t
hat
t
r
ans
m
i
t
t
he d
at
a p
a
c
k
et
s
w
i
t
ho
u
t
e
rro
rs
,
l
0
j
=
N
L
l
0
f
o
r
ov
er
l
a
y
no
des
i
n
w
hi
c
h
t
r
ans
m
i
s
s
i
on pac
k
et
er
r
or
s
hav
e be
en d
et
ec
t
e
d.
T
he pr
obab
i
l
i
t
y
of
r
ec
ov
er
i
ng t
he
er
r
or
s
ub
-
pa
c
k
et
f
r
o
m
t
he s
our
c
e nod
e t
r
ans
m
i
t
t
i
ng at
a r
at
e of
c
r
b
i
s
d
ef
i
ned
as
:
k
,
l
0
(
T
)
=
ℂ
k
,
l
0
(
T
−
b
)
x
t
r
s
(
b
)
d
b
.
T
0
(
20)
∂
k
,
l
0
(
T
)
∂
T
=
k
,
l
0
(
T
−
b
)
Q
k
,
l
0
∗
V
P
T
0
(
T
−
b
)
x
t
r
r
(
b
)
d
b
(
21)
k
,
l
0
(
T
)
=
k
,
l
0
(
T
−
b
)
1
−
Q
k
,
l
0
∗
V
P
(
T
−
b
)
T
0
x
t
r
r
(
b
)
d
b
(
22)
k
,
l
0
(
T
)
=
ℂ
k
,
l
0
(
T
−
b
)
x
tl
(
b
)
d
b
T
0
(
23)
∂
q
k
,
l
0
(
T
)
∂
T
=
∂
k
,
l
0
(
T
)
∂
T
+
∂
k
,
l
0
(
T
)
∂
T
.
(
24)
Q
k
,
l
0
V
P
x
(
T
)
=
q
k
,
l
0
(
T
)
+
1
−
q
k
,
l
0
(
T
)
Q
BR
j
,
l
0
j
(
T
)
≥
j
≠
k
D
EC
(
25)
Q
k
,
0
V
P
x
(
T
)
=
ℍ
T
−
(
k
−
1
)
c
S
(
26)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
8
94
–
90
3
900
W
h
er
e
c
i
s
t
he
a
v
er
a
ge
s
u
b
-
pac
k
et
s
i
z
e
of
t
h
e
s
ub
-
pa
c
k
et
s
i
s
,
S
is
t
h
e
s
t
r
eam
bi
t
r
a
t
e
of
v
i
de
o
s
t
r
eam
and
ℍ
i
s
t
he
H
ea
v
i
s
i
de
s
t
ep
f
unc
t
i
on
.
T
he
pr
obabi
l
i
t
y
t
h
at
an
o
v
er
l
a
y
n
od
e
at
l
0
l
ev
el
i
n
T
r
ee
k
hav
i
ng
a
r
eq
ui
r
e
d p
ac
k
et
and c
a
n s
en
d
i
t
t
o
a r
eq
ues
t
ed
n
ode
w
i
t
h
i
n
a
t
i
m
e f
r
a
m
e of
a
s
ec
onds
i
s
g
i
v
en
as
:
I
f
t
he
s
ub
-
pac
k
et
r
eques
t
e
d
i
s
una
v
ai
l
ab
l
e
i
n
t
h
e
s
u
b
-
t
r
ee
T
r
e
e
k
t
hen
t
he
pr
ob
ab
i
l
i
t
y
of
t
he
s
u
b
-
pac
k
et
av
a
i
l
a
bi
l
i
t
y
w
i
t
hi
n
t
h
e ent
i
r
e n
et
w
or
k
T
r
ee i
s
gi
v
en a
s
:
T
he E
C
D
D
ado
pt
s
a m
ul
t
i
pat
h di
s
t
r
i
but
i
o
n m
odel
f
or
v
i
de
o or
m
ul
t
i
m
edi
a dat
a.
T
o
r
educ
e
per
ho
p
de
l
a
y
,
D
a
nd
end
t
o
end
d
el
a
y
,
E
D
t
he
E
C
D
D
ado
pt
s
a
F
E
C
t
e
c
hni
q
ue
w
i
t
h
a
nov
el
r
et
r
ans
m
i
s
s
i
on t
ec
h
ni
que
t
o
ens
ur
e
dat
a d
el
i
v
er
y
w
i
t
hi
n
t
h
e d
el
a
y
bo
unds
.
T
o e
v
a
l
uat
e t
he
per
f
or
m
anc
e of
E
C
D
D
a
n
ex
per
i
m
ent
al
s
t
u
d
y
i
s
c
on
duc
t
ed
and
i
s
pr
es
ent
e
d i
n t
he
f
ol
l
o
w
i
ng
s
ec
t
i
on
of
t
hi
s
pa
per
.
5.
E
x
p
e
r
i
m
e
n
ta
l
S
tu
d
y
a
n
d
S
i
m
u
l
a
ti
o
n
R
e
s
u
l
ts
T
o ev
al
ua
t
e t
he p
er
f
or
m
anc
e of
t
he pr
opos
ed E
C
D
D
m
ode
l
w
e c
ons
i
der
m
ul
t
i
pa
t
h v
i
deo
t
r
ans
m
i
s
s
i
on
s
c
enar
i
os
ov
e
r
w
i
r
el
es
s
no
des
i
n
a
s
i
m
ul
at
i
o
n
p
l
at
f
or
m
.
T
he
s
i
m
ul
at
i
on
p
l
at
f
or
m
i
s
dev
el
ope
d
us
i
n
g
MA
T
LA
B
and
C
+
+
.
A
n
ar
ea
i
.
e.
=
1
0
0
m
×
1
0
0
m
w
as
c
ons
i
der
e
d
t
o
de
pl
o
y
100
w
i
r
e
l
es
s
n
odes
.
T
he
band
w
i
d
t
h
as
s
i
gn
ed t
o
ea
c
h
w
i
r
e
l
es
s
n
ode
i
s
1
M
bp
s
.
T
he w
i
r
el
es
s
c
hanne
l
s
ar
e
m
odel
ed
us
i
ng
t
h
e
A
dd
i
t
i
v
e
w
hi
t
e
G
aus
s
i
an
no
i
s
e
bas
e
d
noi
s
e
m
odel
.
T
he
ph
y
s
i
c
al
l
a
y
er
de
v
el
op
ed
f
or
t
he
s
i
m
ul
at
i
o
n
en
v
i
r
o
nm
ent
i
s
i
n
ac
c
or
danc
e
t
o
t
he
I
E
E
E
802
.
11
g
s
t
andar
d.
O
r
t
ho
gon
al
F
r
e
quenc
y
D
i
v
i
s
i
on
Mu
l
t
i
pl
ex
i
ng
i
s
c
o
ns
i
der
ed f
or
t
r
a
ns
m
i
t
t
i
ng
a
nd
r
ec
ei
v
i
n
g v
i
deo d
at
a.
T
he v
i
deo d
at
a t
o be t
r
a
ns
m
i
t
t
ed i
s
i
ni
t
i
al
l
y
e
nc
ode
d us
i
ng H
.
26
4/
S
V
C
J
SVM
[
18]
.
T
he enc
od
ed
v
i
deo
dat
a i
s
t
r
a
ns
m
i
t
t
e
d v
i
a
a m
ul
t
i
pa
t
h c
hann
el
.
T
he
m
T
r
eeB
one
al
gor
i
t
hm
i
s
ad
opt
e
d t
o
c
ons
t
r
uc
t
t
he
ov
er
l
a
y
and
e
l
i
m
i
nat
e
uns
t
a
bl
e
w
i
r
el
es
s
nod
es
.
T
he ov
er
l
a
y
t
r
ee
i
.
e.
T
r
ee
c
ons
t
r
uc
t
ed c
ons
i
d
er
s
s
our
c
e node S
as
t
h
e r
oot
nod
e
and
t
h
e
l
eaf
node
as
des
t
i
na
t
i
o
n n
ode
D
.
T
he
s
our
c
e
and
des
t
i
nat
i
o
n
ar
e
s
e
l
ec
t
e
d
s
o
as
t
o
s
u
ppor
t
5
m
ul
t
i
pat
h
t
r
a
ns
m
i
s
s
i
ons
of
t
he
v
i
de
o
dat
a.
T
he
T
r
ee
obt
a
i
ne
d f
r
om
adopt
i
ng
t
he
m
T
r
eebone
al
gor
i
t
hm
i
s
f
u
r
t
her
s
p
l
i
t
i
nt
o
f
i
v
e s
ub
-
t
r
e
es
nam
el
y
T
r
ee
1
,
T
r
ee
2
,
T
r
ee
3
,
T
r
ee
4
and T
r
ee
5
.
E
ac
h s
u
b
-
t
r
ee
ob
t
a
i
n
ed h
as
l
0
=
7
i
.
e.
S
,
5
ho
p
n
od
es
an
d
t
h
e
d
es
t
i
n
at
i
on
D
.
T
he
per
f
or
m
anc
e
of
E
C
D
D
i
s
c
om
par
ed
w
i
t
h
t
he
s
t
at
e
of
ar
t
S
P
MLD
a
l
g
or
i
t
hm
.
J
i
y
a
n
[
1
4]
h
as
r
epo
r
t
ed
t
h
e
s
up
er
i
or
per
f
or
m
anc
e
of
S
P
ML
D
ov
er
t
he ex
i
s
t
i
n
g E
f
f
ec
t
i
v
e
D
el
a
y
-
C
o
nt
r
ol
l
e
d Lo
ad D
i
s
t
r
i
but
i
on
m
odel
[
2]
,
F
L
A
R
E
[
3]
,
L
e
as
t
-
L
oad
ed
-
F
i
r
s
t
a
ppr
oac
h
[
4]
,
Load
B
a
l
a
nc
i
ng
P
ar
al
l
el
F
or
w
ar
di
ng
[5
]
an
d
t
h
e
F
l
o
w
S
l
i
c
e
s
c
hem
e
[
19]
.
I
n t
h
i
s
pap
er
,
t
he aut
h
or
s
pr
es
ent
t
he c
om
par
i
s
on r
es
ul
t
s
bet
w
ee
n t
he pr
o
po
s
ed E
C
D
D
an
d
t
he S
P
MLD
al
gor
i
t
hm
.
A
s
c
enar
i
o
c
ons
i
der
i
ng
5
00
v
i
deo
dat
a
pac
k
et
s
ar
e
s
i
m
ul
at
ed
us
i
n
g
t
h
e
E
C
D
D
a
nd
t
h
e
S
P
MLD
al
gor
i
t
hm
s
.
T
he
s
y
m
m
et
r
i
c
S
i
g
nal
t
o
N
o
i
s
e
R
at
i
o
i
s
c
ons
i
d
er
ed
t
o
be
5
d
b
.
T
he
num
ber
of
r
et
r
ans
m
i
s
s
i
on r
eques
t
s
obs
er
v
e
d b
y
a
l
l
t
he no
de
s
i
n t
he T
r
ee c
ons
t
r
uc
t
ed i
s
as
s
how
n i
n
F
i
gur
e
4.
F
r
om
t
he
gr
aph
i
t
c
an
be o
bs
er
v
e
d
t
h
at
t
h
e
r
et
r
ans
m
i
s
s
i
on r
eques
t
s
a
r
e
r
educ
ed
b
y
abou
t
52
.
27%
c
ons
i
d
er
i
n
g t
he E
C
D
D
w
hen
c
om
par
ed t
o S
P
ML
D
a
l
gor
i
t
hm
.
R
educ
i
ng E
D
al
s
o h
el
ps
i
n r
educ
i
n
g t
he l
os
t
pac
k
et
pr
obab
i
l
i
t
i
es
.
R
ed
uc
t
i
o
n i
n l
os
t
pac
k
et
pr
obabi
l
i
t
i
es
i
n
t
ur
n
hel
ps
i
n
ef
f
i
c
i
ent
v
i
d
eo
t
r
an
s
m
i
s
s
i
on
and
r
ec
ons
t
r
uc
t
i
o
n.
T
o
m
eas
ur
e
t
he qu
al
i
t
y
of
r
ec
ons
t
r
uc
t
i
on,
t
he a
ut
h
or
s
hav
e a
do
pt
ed
v
i
d
eo P
S
N
R
c
om
put
at
i
o
ns
at
f
r
a
m
e
l
e
ve
l
.
T
o
m
eas
ur
e
t
he
r
ec
o
ns
t
r
uc
t
i
on
qua
l
i
t
y
i
n
t
er
m
s
of
P
S
N
R
a
nd
t
o
m
eas
ur
e
t
he
en
d
t
o
e
nd
del
a
y
,
v
i
deo d
at
a of
N
+
x
f
r
am
es
ar
e c
ons
i
der
ed.
T
he s
ub pac
k
et
i
z
at
i
o
n s
c
hem
e i
s
adopt
e
d
and
t
he
s
ub
pac
k
et
s
ar
e
t
r
ans
m
i
t
t
ed
f
r
o
m
S
node
t
o
D
n
ode
t
hr
oug
h
5
m
ul
t
i
pat
hs
.
T
he
s
y
m
m
et
r
i
c
S
i
g
na
l
t
o N
oi
s
e R
at
i
o i
n t
hi
s
c
as
e i
s
t
ak
en as
5db.
T
he v
i
deo
r
ec
ons
t
r
uc
t
i
on r
es
ul
t
s
t
hus
obt
a
i
ne
d c
ons
i
d
er
i
n
g E
C
D
D
and S
P
ML
D
s
c
hem
es
ar
e s
ho
w
n i
n F
i
gur
e 5.
I
n
S
P
MLD
t
he h
i
g
hes
t
Q
k
,
l
0
V
P
(
a
)
=
q
k
,
l
0
(
T
k
)
+
1
−
q
k
,
l
0
(
T
k
)
Q
BR
j
,
l
0
j
(
T
k
)
≥
j
≠
k
D
EC
(
27)
Q
V
P
(
a
)
=
1
r
1
R
r
k
=
1
Q
k
,
l
0
V
P
(
a
)
(
T
)
R
l
0
N
l
l
0
=
1
(
28)
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
E
r
r
or
R
es
i
l
i
e
nt
Mul
t
i
p
at
h
V
i
deo D
el
i
v
er
y
on W
i
r
el
es
s
O
v
er
l
ay
N
et
w
or
k
s
(
B
U
ma
Mahes
w
ar
i
)
901
PSN
R
v
al
u
e
obs
er
v
ed
i
s
46
.
177
0
d
B
w
h
i
c
h
i
s
i
n
c
onc
u
r
r
enc
e
w
i
t
h
t
he
r
es
u
l
t
s
p
ub
l
i
s
hed
i
n
[
14]
.
F
r
o
m
F
i
gur
e
5
i
t
i
s
ev
i
d
ent
t
hat
E
C
D
D
ex
h
i
b
i
t
s
h
i
g
h
e
r
PSN
R
v
al
ues
w
hen
c
om
par
ed
t
o
S
P
M
LD
and
t
he hi
g
hes
t
v
a
l
ue obs
e
r
v
ed i
n E
C
D
D
i
s
4
8.
68
31 d
B
.
F
i
gur
e
4
.
N
um
ber
of
R
et
r
ans
m
i
s
s
i
on R
eq
ues
t
s
in
E
C
D
D
and
S
P
ML
D
F
i
gur
e
5
.
V
i
d
eo f
r
am
e r
ec
ons
t
r
uc
t
i
on
at
t
he
des
t
i
n
at
i
on bas
ed on P
S
N
R
T
he
end
t
o
end
de
l
a
y
i
s
m
eas
ur
ed
an
d
t
he
r
es
ul
t
s
ob
t
ai
n
ed
i
s
s
ho
w
n
i
n
F
i
gur
e
6
.
I
t
c
a
n
be
obs
er
v
ed
t
hat
t
he
E
D
f
or
E
C
D
D
an
d
S
P
ML
D
ar
e
ex
pone
nt
i
al
i
n
n
at
ur
e.
T
hi
s
i
s
i
n
c
ons
i
s
t
e
nt
w
i
t
h
m
odel
obt
ai
n
ed i
n E
q
ua
t
i
on
.
1
2 f
or
per
ho
p de
l
a
y
s
.
T
he c
u
m
ul
at
i
v
e
per
h
op d
e
l
a
y
s
ar
e us
e
d
t
o ob
t
ai
n t
h
e e
nd t
o en
d d
el
a
y
i
.
e.
E
D
.
T
he
end
t
o e
nd d
el
a
y
ov
er
x
num
ber
of
w
i
r
el
es
s
hop
nodes
f
or
E
C
D
D
i
.
e.
ED
EC
D
D
i
s
f
ou
nd t
o
be
:
T
he end t
o
en
d d
el
a
y
f
or
S
P
ML
D
i
s
f
ound t
o
be
:
B
as
ed
on t
he d
ef
i
ni
t
i
o
ns
f
or
ED
EC
D
D
and
ED
S
P
M
L
D
i
t
c
an
be c
o
nc
l
ud
e
d t
hat
t
he
E
C
D
D
ex
hi
b
i
t
s
a
l
o
w
er
end
t
o
e
nd d
el
a
y
a
nd t
he
per
f
or
m
anc
e i
m
pr
ov
e
m
ent
i
s
hi
gher
f
or
l
ar
ger
net
w
or
k
s
.
E
x
per
i
m
ent
al
s
t
ud
y
i
s
c
on
duc
t
ed
t
o s
t
u
d
y
t
he ef
f
ec
t
of
w
i
r
e
l
es
s
c
ha
nne
l
no
i
s
e a
nd
v
i
deo
dat
a
t
r
ans
m
i
s
s
i
on
on
E
C
D
D
a
nd
S
P
MLD
m
ec
hani
s
m
s
.
E
x
per
i
m
ent
al
s
t
ud
y
i
s
c
ond
uc
t
ed
t
o s
t
ud
y
t
he ef
f
ec
t
of
w
i
r
el
e
s
s
c
hannel
n
oi
s
e a
nd
v
i
d
eo
dat
a t
r
ans
m
i
s
s
i
on on
E
C
D
D
and S
P
M
LD
m
e
c
hani
s
m
s
.
T
he
s
i
gnal
t
o
no
is
e
r
a
t
io
is
v
a
r
ie
d
f
r
o
m
0
d
B
to
4
d
B
.
T
he
bi
t
er
r
or
s
obs
er
v
ed
at
t
he
s
ub pac
k
et
l
e
v
e
l
ar
e m
eas
u
r
ed an
d t
h
e r
es
ul
t
s
obt
ai
ne
d ar
e s
ho
w
n
i
n F
i
gur
e
7.
F
i
gur
e
6
.
F
r
am
e bas
ed
E
nd
t
o
E
nd
D
e
l
a
y
m
eas
ur
e
d at
t
he des
t
i
nat
i
o
n
F
i
gur
e
7
.
B
i
t
E
r
r
or
R
at
e
m
eas
ur
ed w
i
t
h
v
ar
y
i
n
g S
N
R
c
ons
i
der
i
ng
E
C
D
D
an
d
S
P
MLD
0
1
0
0
2
0
0
3
0
0
4
0
0
5
0
0
6
0
0
1
3
5
7
9
1
1
1
3
1
5
1
7
1
9
2
1
2
3
2
5
# of
P
ac
k
et
R
et
r
ans
m
i
s
s
i
on
R
eques
t
s
S
i
m
ul
at
i
on T
i
m
e
(
m
s
)
P
ac
k
et
R
et
r
ans
m
i
s
s
i
on R
eques
t
s
E
C
DD
SPM
L
D
3
5
4
0
4
5
5
0
N+
1
N+
6
N+
1
1
N+
1
6
N+
2
1
PSN
R
(
d
B)
V
i
deo F
r
am
e
N
um
ber
V
i
deo
R
ec
ons
t
r
uc
t
i
on P
er
f
or
m
anc
e
-
PSN
R
E
C
DD
SPM
L
D
0
5
0
1
0
0
1
5
0
2
0
0
N+
1
N+
2
N+
3
N+
4
N+
5
N+
6
N+
1
1
N+
2
1
E
nd
T
o
E
nd D
el
ay
(
m
s
)
V
i
deo F
r
am
e N
um
ber
s
F
r
am
e
E
nd T
o
E
nd D
el
ay
O
bs
er
v
ed
E
C
DD
SPM
L
D
E
xp
on
.
(
E
C
D
D
)
E
xp
on
.
(
S
P
M
L
D
)
ED
EC
D
D
=
1
6
.
4
8
4
e
(
x
×
0
.
3
0
)
ED
S
P
M
L
D
=
1
4
.
9
3
6
e
(
x
×
0
.
3
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
8
94
–
90
3
902
T
he
r
es
ul
t
s
obt
ai
n
ed
s
ho
w
t
hat
t
h
e
bi
t
er
r
or
s
i
nd
uc
ed
b
y
t
h
e
w
i
r
e
l
es
s
c
hanne
l
no
i
s
e
ar
e
r
educ
ed b
y
ab
out
23.
9
3%
us
i
ng
pr
op
os
ed E
C
D
D
.
B
as
ed o
n t
h
e ex
per
i
m
ent
al
s
t
ud
y
and
t
he
c
or
r
es
pondi
ng
r
es
ul
t
s
pr
es
ent
e
d i
n t
hi
s
pap
er
i
t
c
an
be c
o
nc
l
u
ded
t
h
at
t
he
E
C
D
D a
ch
i
e
v
e
s
bet
t
er
per
f
or
m
anc
e ow
i
ng
t
o t
h
e a
dop
t
i
o
n
of
t
he F
E
C a
n
d p
ac
k
et
r
et
r
ans
m
i
s
s
i
on t
ec
hni
ques
i
nc
or
por
a
t
ed.
6.
C
o
n
c
l
u
s
i
o
n
I
n t
h
i
s
pap
er
t
h
e i
s
s
ue
s
w
i
t
h
s
up
por
t
i
ng r
eal
-
t
i
m
e
m
ul
t
i
m
edi
a or
v
i
deo
bas
ed
app
l
i
c
at
i
ons
o
v
er
w
i
r
el
es
s
net
w
or
k
ar
e
pr
es
ent
e
d.
T
he
w
or
k
c
ar
r
i
ed
out
s
o
f
ar
t
o
addr
es
s
t
hes
e
i
s
s
ues
and
t
h
ei
r
dr
a
w
bac
k
s
hav
e
b
een
c
l
ear
l
y
d
i
s
c
us
s
ed.
I
t
i
s
obs
er
v
ed
t
h
at
s
u
b
pac
k
et
i
z
at
i
o
n
t
ec
hni
ques
pr
o
v
e t
o
be ef
f
i
c
i
ent
f
or
v
i
de
o del
i
v
er
y
o
v
er
m
ul
t
i
pat
hs
.
O
v
er
l
a
y
net
w
or
k
s
c
an ai
d
i
nf
or
m
at
i
on
f
eedbac
k
es
s
e
nt
i
a
l
f
or
v
i
d
eo
da
t
a
de
l
i
v
er
y
.
T
he
ov
er
l
a
y
ne
t
w
or
k
s
i
nduc
e
ov
er
hea
ds
due t
o t
r
e
e m
ai
nt
en
anc
e.
T
o r
educ
e t
hes
e
o
v
er
he
ad
s
t
hi
s
w
or
k
pr
opos
ed
E
C
D
D
m
odel
.
F
or
e
f
f
i
c
i
ent
v
i
d
eo d
at
a de
l
i
v
e
r
y
a
nd t
o m
i
ni
m
i
z
e t
he d
el
a
y
,
E
C
D
D
ad
opt
s
a s
ub
pac
k
et
i
z
at
i
on
t
ec
hni
que
.
T
he
ov
er
h
eads
ar
e
pr
edom
i
nan
t
l
y
i
n
duc
ed
due
t
o p
ac
k
et
l
os
s
es
.
T
o
r
educ
e pac
k
et
l
os
s
es
E
C
D
D
a
do
pt
s
a
F
E
C
m
ec
hani
s
m
ai
di
ng
i
n
r
ec
ov
er
y
.
A
r
e
t
r
ans
m
i
s
s
i
on
t
e
c
hni
q
ue
i
s
a
l
s
o
di
s
c
us
s
ed t
o obt
ai
n t
h
e l
os
t
pac
k
et
s
i
n t
he net
w
or
k
.
T
h
e per
f
or
m
anc
e of
E
C
D
D
i
s
c
o
m
par
ed w
i
t
h
a s
t
at
e of
ar
t
ex
i
s
t
i
ng m
od
el
i
n t
h
e ex
per
i
m
ent
al
s
t
ud
y
.
T
he per
f
or
m
anc
e i
m
pr
o
v
em
ent
of
t
he
E
CDD
i
s
pr
ov
ed
i
n
t
er
m
s
of
t
he en
d t
o e
nd
de
l
a
y
,
v
i
de
o r
ec
ons
t
r
uc
t
i
on
qu
al
i
t
y
,
r
e
duc
ed
num
ber
of
r
et
r
ans
m
i
s
s
i
ons
and
r
ob
us
t
nes
s
t
o n
oi
s
e.
A
s
a f
ut
ur
e
w
or
k
,
t
he per
f
or
m
anc
e of
t
he
E
C
D
D
w
i
l
l
be e
v
a
l
ua
t
ed f
or
l
ar
ger
n
et
w
or
k
s
s
uppor
t
i
ng m
ul
t
i
p
l
e
v
i
de
o s
t
r
eam
s
.
R
ef
er
en
ces
[1
]
I
TU
-
T
.
G
.
114
.
O
n
e
-
w
ay
t
r
an
s
m
i
s
s
i
on
t
i
m
e
.
I
nt
er
nat
i
ona
l
T
e
l
e
c
om
m
uni
c
at
i
on U
ni
o
n
.
R
ec
om
m
endat
i
on
.
200
3.
[2
]
P
r
abhav
at
S
,
N
i
s
h
i
y
am
a H
,
A
ns
ar
i
N,
K
at
o N
.
E
f
f
ec
t
i
v
e
D
el
ay
-
C
ont
r
ol
l
ed Loa
d
D
i
s
t
r
i
b
ut
i
on ov
e
r
M
ul
t
i
pat
h N
et
w
or
k
s
.
I
E
E
E
T
r
ans
a
c
t
i
ons
o
n P
ar
al
l
el
and D
i
s
t
r
i
but
e
d S
y
s
t
em
s
.
20
11;
22
(
10
):
173
0
-
1741
.
[3
]
Sri
k
a
n
t
h
K
and
ul
a,
D
i
na K
at
a
bi
,
S
hant
anu
Si
n
h
a
,
Art
h
u
r Be
rg
e
r.
D
y
n
am
i
c
l
o
ad bal
anc
i
ng w
i
t
h
o
u
t
pac
k
et
r
e
or
der
i
ng.
A
C
M
S
I
G
C
O
M
M
C
om
put
e
r
C
o
mm
un
i
c
at
i
on R
ev
i
ew
.
2
007
:
51
-
62
.
[4
]
Ch
i
-
C
hung
H
ui
,
C
h
an
s
on S
.
T
.
H
y
dr
ody
nam
i
c
l
oad
bal
anc
i
ng
.
I
E
E
E
T
r
ans
ac
t
i
on
s
on
P
ar
al
l
el
an
d
D
i
s
tr
i
b
u
te
d
S
y
s
te
m
s
.
199
9;
10
(
11
):
1
118
-
1
137.
[5
]
W
eig
ua
n
g
S
hi
,
Ma
c
G
r
egor
M
H
,
G
bur
z
y
ns
k
i
P
.
Loa
d B
al
a
n
c
i
ng f
or
P
ar
al
l
el
F
or
w
ar
di
ng
.
I
EEE/
AC
M
T
r
ans
ac
t
i
on
s
on
N
et
w
or
k
i
ng
.
2
005;
13
(
4
):
790
-
801
.
[6
]
C
hangq
i
ao X
u,
Z
Li
,
J
Li
,
H
Z
hang,
G
M
M
unt
ean
.
Cr
o
s
s
-
l
ay
er
F
ai
r
ne
ss
-
dr
i
v
en C
on
c
ur
r
e
nt
M
ul
t
i
pa
t
h
V
i
deo D
el
i
v
er
y
ov
er
H
et
er
oge
nous
W
i
r
el
es
s
N
et
w
or
k
s
.
I
E
E
E
T
r
ans
a
c
t
i
ons
on C
i
r
c
ui
t
s
an
d S
y
s
t
em
s
f
or
V
i
d
eo T
e
c
hn
ol
o
gy
.
201
5;
25
(
7
):
1
175
-
1
189
.
[7
]
A
A
Al
m
a
z
ro
i
,
M
A
N
gadi
.
R
obus
t
P
at
h
C
on
s
t
r
u
c
t
i
o
n
f
or
R
el
i
a
bl
e
D
at
a
T
r
ans
m
i
s
s
i
ons
i
n
N
ode
D
i
s
j
oi
n
t
M
ul
t
i
pat
h R
out
i
ng
f
or
W
i
r
el
e
s
s
S
e
ns
or
N
et
w
or
k
.
T
EL
KO
M
N
I
KA
T
el
ec
om
m
uni
c
at
i
on,
C
om
put
i
ng
,
E
l
ec
t
r
oni
c
s
and
C
o
nt
r
o
l
.
20
15;
13
(3
):
904
-
921
.
[8
]
Ch
u
YH
,
R
ao S
G
,
S
es
han
S
,
H
ui
Z
h
ang
.
A
c
a
s
e f
o
r
e
n
d
sy
st
e
m
mu
l
t
i
ca
st
.
I
EEE
T
r
a
ns
a
c
t
io
n
s
on
S
el
ec
t
ed A
r
eas
i
n C
om
m
uni
c
a
t
i
on
s
.
20
02;
20
(
8
)
:
145
6
-
14
71
.
[9
]
M
aghar
ei
N
,
R
ej
ai
e R
.
P
R
I
M
E
:
P
eer
-
to
-
P
e
er
R
ec
e
i
v
er
-
D
r
i
v
en
M
es
h
-
B
as
ed S
t
r
ea
m
i
ng
.
I
EEE/
AC
M
T
r
ans
ac
t
i
on
s
on
Ne
t
wo
r
k
i
n
g.
2
009;
17
(
4
):
105
2
-
106
5
.
[1
0
]
Y
ong C
hen,
W
ei
-
z
hong X
i
a
o,
H
uan
-
l
i
n
L
i
u,
Long
-
z
h
ao
S
un.
E
f
f
i
c
i
e
nt
C
on
t
ent
Loc
at
i
on U
s
i
n
g
S
em
an
t
i
c
S
m
al
l
W
o
r
l
d i
n P
e
er
-
to
-
Pe
e
r N
e
t
w
o
rk
s
.
T
EL
KO
M
N
I
KA
T
el
ec
om
m
uni
c
at
i
on,
C
om
put
i
n
g
,
E
l
ec
t
r
oni
c
s
and
C
o
nt
r
o
l
.
20
13;
11
(2
):
223
-
230
.
[1
1
]
S
B
aner
j
ee,
B
B
hat
t
ac
h
ar
j
e
e,
C
K
om
m
ar
e
ddy
.
S
c
a
l
ab
l
e appl
i
c
at
i
on l
a
y
er
m
ul
t
i
c
a
s
t
.
i
n
t
he
P
r
oc
ee
di
n
gs
of
A
C
M
S
I
G
C
O
M
M
C
onf
er
enc
e on A
ppl
i
c
at
i
on
s
,
T
ec
hno
l
og
i
e
s
,
A
r
c
hi
t
e
c
t
ur
e
s
,
a
n
d
Pro
t
o
c
ol
s
f
or
C
o
m
pu
t
er
C
o
m
m
uni
c
at
i
o
n
.
2
002;
32
:
205
-
217.
[1
2
]
F
eng
W
a
n
g
,
Y
ongqi
ang
Xi
o
n
g
,
J
i
a
ng
c
hua
n Li
u
.
m
T
r
eebo
ne:
A
C
ol
l
abor
a
t
i
v
e T
r
ee
-
Me
s
h
O
v
e
r
l
a
y
N
et
w
or
k
f
or
M
ul
t
i
c
a
s
t
V
i
deo
S
t
r
eam
i
ng
.
I
E
E
E
T
r
an
s
ac
t
io
ns
o
n P
a
r
al
l
el
a
nd
D
is
t
r
i
bu
t
ed
S
y
s
t
e
m
s
.
2010;
21
(
3
):
3
79
-
392
.
[1
3
]
K
al
v
ei
n R
ant
e
l
obo
,
W
i
r
aw
an
W
i
r
aw
an,
G
am
ant
y
o H
e
ndr
a
nt
or
o,
A
c
hm
ad A
f
f
an
di
.
C
om
b
i
n
ed S
c
al
a
bl
e
V
i
deo
C
odi
ng M
et
hod
f
or
W
i
r
el
es
s
T
r
an
s
m
i
s
s
i
on
.
T
E
LK
O
M
N
I
K
A
T
el
ec
om
m
uni
c
at
i
o
n,
C
om
put
i
n
g
,
E
l
ec
t
r
oni
c
s
and
C
o
nt
r
o
l
.
20
11;
9(
2)
:
2
95
-
3
02.
[1
4
]
J
i
y
an
W
u
,
J
i
ngq
i
Y
ang
,
Y
an
l
ei
S
han
g
,
B
o C
heng
,
J
unl
i
ang C
hen
.
SPM
L
D
:
Su
b
-
pac
k
et
ba
s
e
d
m
ul
t
i
pat
h l
oa
d di
s
t
r
i
but
i
on f
or
r
eal
-
t
i
m
e
m
ul
t
i
m
edi
a
t
r
af
f
i
c
.
J
o
ur
nal
o
f
C
om
m
uni
c
at
i
on
s
an
d N
et
w
or
k
s
.
2014;
16
(
5
):
5
48
-
558
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
E
r
r
or
R
es
i
l
i
e
nt
Mul
t
i
p
at
h
V
i
deo D
el
i
v
er
y
on W
i
r
el
es
s
O
v
er
l
ay
N
et
w
or
k
s
(
B
U
ma
Mahes
w
ar
i
)
903
[1
5
]
D
enui
t
M
,
G
enes
t
C
,
M
ar
c
eau
É
.
S
t
oc
ha
s
t
i
c
boun
ds
on
s
um
s
of
de
pend
ent
r
i
s
k
s
,
I
ns
ur
a
nc
e
.
M
at
hem
at
i
c
s
a
nd E
c
onom
i
c
s
.
1999;
25
(
1
):
85
-
104
.
[1
6
]
P
u
W
a
ng
,
A
k
yi
l
d
i
z I
F
.
O
n
t
he
O
r
i
gi
ns
of
H
e
av
y
-
T
ai
l
ed D
el
ay
i
n D
y
nam
i
c
S
pec
t
r
um
A
c
c
es
s
N
et
w
or
k
s
.
I
E
E
E
T
r
an
s
ac
t
i
on
s
o
n M
obi
l
e
C
om
put
i
ng
.
201
2;
11
(2
):
204
-
21
7
.
[1
7
]
C
r
ov
el
l
a
M
E
,
T
aqqu
M
S,
Be
s
t
a
v
ro
s
A
.
H
eav
y
-
t
a
i
l
ed
p
r
o
b
a
bi
li
t
y
d
is
t
r
i
bu
t
i
on
s
i
n t
h
e
W
or
l
d
W
ide
W
eb.
In
:
A
dl
e
r
R
J
,
F
e
ld
m
an
R
E
,
T
aqqu
M
S
.
E
di
t
or
s
.
A
P
r
ac
t
i
c
al
G
ui
de
T
o
H
eav
y
T
ai
l
s
.
N
ew
Y
or
k
:
C
hapm
a
n a
nd H
al
l
;
1
998
:
3
-
2
6.
[1
8
]
H
.
2
6
4
/
SVC
J
SVM
s
o
f
t
w
a
re
.
ht
t
p:
/
/
i
p.
hhi
.
de/
i
m
a
ge
c
om
-
G
1/
s
av
c
e/
d
ow
nl
oad
s
/
S
V
C
-
R
ef
er
en
c
e
-
S
o
ftw
a
r
e
.h
tm
.
[1
9
]
Sh
i
L
,
B
in
L
iu
,
C
h
angh
ua S
u
n
,
Z
he
ngy
u Y
i
n
,
Lax
m
i
B
huy
an
,
C
ha
o H
J.
Load
-
B
al
a
nc
i
ng
M
ul
t
i
pat
h
S
w
i
t
c
hi
ng S
y
s
t
em
w
i
t
h F
l
ow
S
l
i
c
e
.
I
E
E
E
T
r
an
s
a
c
t
i
ons
on C
o
m
put
er
s
.
201
2;
61
(3
)
:
35
0
-
36
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.