TELKOM
NIKA
, Vol.11, No
.2, Februa
ry 2013, pp. 80
9
~
81
8
ISSN: 2302-4
046
809
Re
cei
v
ed Au
gust 5, 201
2; Re
vised Decem
ber
28, 20
12; Accepted
Jan
uary 13, 2
013
Key Agreement Procotol in DSN
Jian Zhou*
1,2
, Xian
w
e
i Zhou
1
1
School of Co
mputer an
d Co
mmunicati
on E
ngi
neer
in
g,
Uni
v
ersit
y
of Sci
e
n
c
e and T
e
chno
log
y
Be
iji
ng,
Beiji
ng, Ch
in
a, 100
08
3
2
School of Man
agem
ent Scie
n
c
e and En
gi
ne
er, Anhui U
n
iv
ersit
y
of F
i
n
anc
e and Eco
nom
i
cs, Anhui, chin
a,
233
04
1
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: ac_zj_c
ourse
@16
3
.com
A
b
st
r
a
ct
In spac
e
netw
o
rks, the
lo
ng
time
del
ay, l
a
rg
e sca
le
an
d
mainta
ina
b
il
ity of
difficu
ltly
natur
e
mak
e
s
key ma
na
ge
ment mor
e
diffic
u
lt than gro
u
n
d
w
i
rele
ss net
w
o
rks. T
he main cha
lle
ng
e is how
to handl
e 1
-
affects-n probl
em th
at beco
m
e mor
e
seri
ous
as entities spr
ead ov
er a w
i
d
e
geo
gra
phic a
r
ea. T
o
solve it
,
this p
aper
pro
poses
a
on
e-to-many
map
p
i
ng s
har
e
d
k
e
y
agr
ee
me
nt,
w
h
ich is base
d
on one-to-
m
an
y
encrypti
on
mec
han
is
m mode
l. In t
he prop
ose
d
key agre
e
m
e
n
t, each entity
has differe
nt d
e
cryptio
n
key a
n
d
shares a
n
e
n
cryption key. W
h
en an
entity jo
i
n
s or le
aves n
e
tw
ork, update
d
keys on
ly are
publ
ic encry
pti
o
n
key and its d
e
c
ryption key.
How
e
ver,
the other entiti
e
s
’
secret key
re
ma
ins u
n
cha
n
ged, so as to
each
me
mber h
a
s the ab
ility to up
date ke
y auto
n
o
mous
ly. Cons
equ
ently the p
e
rformanc
e of the prop
ose
d
key
ma
na
ge
me
nt sche
m
e is
unre
l
ated to the n
e
t
w
ork scale
, no
de
mob
ility a
n
d
topol
ogy
stru
cture. It is show
n
that our
pr
op
o
s
ed k
e
y
mana
ge
me
nt sch
e
m
e n
o
t o
n
ly
i
m
proves
the
effi
ciency
an
d fl
e
x
ibil
ity for s
p
a
c
e
netw
o
rks, but also ach
i
eves
g
ood sec
u
rity pr
operti
es,
inclu
d
i
ng forw
ard sec
u
rity and b
a
ck
w
a
rd security a
n
d
ma
ny more by
theoretic
al a
nal
yses.
Ke
y
w
ords
: Sp
ace Netw
ork, Key Mana
ge
me
nt, Reke
ying, 1-affects-n Problem
, Security
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
Similar to the
grou
nd
wirel
e
ss net
works, spa
c
e n
e
two
r
ks [1] also n
eed
key man
ageme
n
t
techn
o
logie
s
to prote
c
t th
e network
wi
th t
he qui
ck
developm
ent
of sp
ace te
chn
o
logie
s
a
n
d
wirel
e
ss
net
work [2], ke
y mana
gem
ent is a
pr
essing
issu
e
for
sp
ac
e netwo
rk se
curity.
Therefore, it
is very ne
ce
ssary to a
dopt
appr
op
riate
key man
age
ment sche
m
e
s to g
uarant
ee
netwo
rk
se
cu
rity [3]. However, spa
c
e
entitie
s spre
ad over a wi
de geog
ra
phi
c are
a
, so it is
difficult to build the powerful on
-line
key manag
e cente
r
to impleme
n
t key manage
m
ent.
Contrary to
grou
nd
wi
rel
e
ss n
e
two
r
ks, spa
c
e
net
works h
a
ve
some
di
sting
u
ishi
ng fe
atu
r
es.
Firstly, the
space n
e
two
r
k environme
n
t is mo
re
co
m
p
lex than
g
r
o
und,
su
ch
a
s
the
ch
annel
is
easily i
n
terfe
r
ed
so
as to th
e bit e
r
ror rat
e
s
of spa
c
e
chann
el i
s
hi
g
h
, so
the
com
m
unication l
o
ad
and
time del
ay
is strictly rest
ri
cte
d
in
desi
gning
se
curity p
r
oto
c
o
l
, in other
wo
rds re
du
cing
the
roun
d compl
e
xity and improving
re
ke
ying efficien
cy are imp
o
rt
ant target
s i
n
de
signin
g
key
manag
eme
n
t schem
e for
spa
c
e
network. Secondly,
in
view
of wi
reless a
nd m
obility, similar to
grou
nd wirele
ss entities
sp
ace ha
rd
ware
level
is also limited
in
cludi
ng
ha
rd
wa
re size,
ha
rd
ware
weight, energy, computat
ional
capability and mem
o
ry
and so
on.
However,
space entities have
stron
g
e
r
p
e
rf
orma
nce o
n
above
ca
pabi
lities tha
n
g
r
ound
wi
rele
ss e
n
tities. F
o
r in
stan
ce,
m
any
spa
c
e
entitie
s’ si
ze
and
weight are big
g
e
r tha
n
groun
d wirele
ss
en
tities, whi
c
h a
l
so all
o
w
sp
a
c
e
entity to equi
p with big
me
mory an
d po
werful
proc
essor [4, 5]. An
d most of
sp
ace
entities
h
a
ve
s
o
lar batteries
, s
o
entities in s
p
ac
e networks
not on
ly have long
er life-time, b
u
t also
sup
p
o
rt
more
co
mple
x operatio
n th
an groun
d wi
reless n
e
two
r
ks. At last, like
gro
und
wirel
e
ss net
work
all
entities
organizes into a
dynamitic
group in
space net
work,
so sp
ace
entities have the ability of
joining i
n
o
r
l
eaving
spa
c
e
netwo
rk. Ho
wever,
sp
ace
entities m
o
vement tre
nd
can
be fo
re
cast
rea
s
on
ably, su
ch a
s
m
a
ny satellite
s
and ai
rcra
ft
s run in
orb
i
t under the
law of
cele
stial
mech
ani
cs.
Acco
rdi
ng to
the regul
ari
t
y, spac
e en
tities perio
di
c time is an
advantage
for
desi
gning
ke
y manage
me
nt scheme. T
herefo
r
e,
tra
d
itional key manag
eme
n
t
solutio
n
s ca
nnot
be dire
ctly ap
plied in spa
c
e
networks.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No. 2, Februa
ry 2013 : 809 – 818
810
2. Related Work
In the literatu
r
e, the existin
g
key di
stribu
tion and
key establi
s
hm
en
t can be divid
ed into
four cla
s
se
s as follo
ws: (1
) Cent
rali
zed
key manag
e
m
ent. There is an entity having po
werf
ul
cap
ability an
d cove
ring
wi
th the wh
ole
netwo
rk,
so it
can
su
ppo
rt
the key m
a
n
ageme
n
t in ti
me
and effici
en
cy includi
ng g
enerating a
n
d
distri
buting
keys, Su
ch a
s
GKMP [6]
and Se
cu
re L
o
ck
[7] The powe
r
ful entities i
s
re
spo
n
si
ble
for key man
ageme
n
t as
a whole a
n
d
rekey ha
ppe
ns
betwe
en
upd
ate mem
b
e
r
and
key m
anag
e
se
rver, so
1-affect
s-
n
p
r
obl
em
is n
o
t exi
s
t; (2)
Contri
buto
r
y key mana
ge
ment. In public ch
annel, al
l
legitimate membe
r
s
contri
bute their sha
r
es
in a
ro
und
o
f
messa
ge
e
x
chan
ge
s to
gene
rate
a
common
key, those
sche
mes targ
et i
s
to
redu
ce th
e complexity of roun
d an
d calcul
ation, su
ch a
s
Ing
e
m
a
rson’
s sch
e
m
e [8], GDH [9],
Octop
u
s [1
0] and BD [1
1]. Whe
n
a me
m
ber le
aves
or joins, all m
e
mbers n
eed t
o
upd
ate thei
r
sha
r
ed
k
e
y
,
whi
c
h
re
sult
s
in 1
-
af
f
e
ct
s-
n
problem
[1
2], so the
sch
e
mes a
r
e
suit
able to th
e
small-
scale net
work; (3) Key manag
eme
n
t base
d
on so
me espe
cia
lly topology
stru
cture. Th
ose
scheme
s
de
al with the efficien
cy question in
large scale net
work with e
s
peci
a
lly topology
stru
ctures [1
3
]. For insta
n
ce,
in the literature, [14,15,
16] su
gge
ste
d
key m
anag
ement
sch
em
es
based on tre
e
stru
cture, such a
s
STR [17],
DH-LKH
[18] and LKH [17]; and [20] sugge
sted key
manag
eme
n
t sch
eme
s
ba
sed o
n
clu
s
te
r. These sch
e
mes all
e
viate 1-affe
cts
-n
probl
em, but the
sho
r
tage
of t
hese
schem
es i
n
cl
ude
limited mo
bility and th
e
cl
uster he
ad
a
nd root a
r
e
the
bottlene
ck of these p
r
oto
c
o
l
s; (4) Pre
-
co
nfigur
atio
n ke
y manageme
n
t, member can get a key or
some
keys in
advance. Such a
s
, kro
n
o
s
[21] is
put fo
rwa
r
d
s
to dist
ribute si
ngle
comm
on secret
key for mem
bers in
advan
ce
s, the n
e
xt perio
dic ti
me
key i
s
ge
nera
t
ed by the l
a
st perio
dic time
key. Ho
weve
r, in sch
eme [
22] netwo
rk
membe
r
s
sel
e
ct pa
rt of ke
ys into altern
ative key set
from
key p
ool
whi
c
h i
s
built by
the off-lin
e
key mana
gem
ent center (K
MC).
To
co
rrectly e
s
tabli
s
h a
se
cure
chan
n
e
l, comm
uni
cation entitie
s
sha
r
e
a com
m
on
key in a
l
ternative
key
set. Mem
bers
can
move fre
e
ly at rand
o
m
and
spe
n
d
less time
in
agre
e
ing
a share
d
key. But alternative
key
set shoul
d h
a
ve eno
ugh
keys to
sh
are a commo
n
key with
a
high p
r
ob
abil
i
ty. Therefore,
rekeying
be
comes a
difficult que
stion
as th
e up
dat
ed
key is sh
ared
by a
lot of mem
bers and
those mem
b
ers lo
catio
n
s are un
kno
w
n to the
KMC, thus in the pr
e-configuration key
manag
eme
n
t sch
eme, it is difficult to guarante
e
forward
se
curity a
nd ba
ckwa
rd
se
curity, and
1-
affec
t
s
-n
p
r
o
b
lem is
an p
r
essing i
s
sue i
n
these sche
mes fo
r space networks.
B
e
ca
us
e cu
r
r
e
n
t
l
y
spa
c
e
key m
anag
ement
schem
es i
s
b
a
s
ed
on
ab
ove
groun
d
key
manag
eme
n
t, su
ch
a
s
[2
3] is
based
on
LKH p
r
oto
c
ol, [
24] is the
co
mbine
of
LK
H a
nd
GDH.
These limitin
g facto
r
s mot
i
vate
the ne
ed fo
r
desi
gning
aut
onomo
u
s an
d secure
key mana
gem
en
t schem
es wi
thout 1
-
affect
s
-n
probl
em for space networks.
3. Auton
o
mo
us Shared
Ke
y
Managem
e
nt in Space
Net
w
o
r
ks
(AKMSN)
In AKMSN, there a
r
e thre
e kind
s of ent
ities: legitimate entity
i
u
{
1
,
2
,
...,
}
in
, bulletin
board
B
and
key mana
gem
ent serve
r
C
. Each
legitimat
e
entity ha
s a
com
m
on
en
cryption
key
and a
differe
nt de
cryption
key, bulletin
board
B
ha
s e
noug
h mem
o
ry to pu
blish i
n
formatio
n on
encryption
ke
y from legitim
a
te memb
ers
and the
C ,
al
l informatio
n
written i
n
B
is public for any
entity in spa
c
e networks, the
C
’s ca
pabili
ty is limited, it can n
o
t cover the
wh
ole
netwo
rk, it
s
manag
eme
n
t rang
e is o
n
l
y
overlap Bul
l
etin boa
rd
B
,
meanwhile it
is abso
lute
sec
u
rity to any
legitimate member.
Figure 1. An Example of Space Net
w
orks
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Key Agre
em
ent Protocol in
DSN (Jia
n Zh
ou)
811
For exam
ple,
as shown i
n
Figure 1,
the l
egitimat
e
entities are satellites, they are
composed into satellite network
around mars,
the
key
m
anagement
server
is built on earth
who
s
e contro
l area is limited and difficult to over
lap mars, the b
u
lletin boa
rd is spa
c
e
station.
Assu
me that
the pro
p
o
s
e
d
proto
c
ol i
s
based
o
n
the Diffie-Hell
man (DH) p
r
otocol
and
all
legitimate me
mbers
sele
ct
*
p
F
as finite g
r
oup a
nd
g
is a the gen
era
t
or of
*
p
F
. The
AKMSN
scheme in
clu
des
key agre
e
ment pha
se,
key use p
h
a
s
e an
d re
key pha
se.
3.1 Ke
y
Agreement Phas
e
In this
phas
e
, member
i
u
{
1
,
2
,
...,
}
in
agree
s the pu
blic en
cryptio
n
key and
se
cre
t
decryption
ke
y with the
C
, and e
a
ch m
e
mber tell
C
the pe
riodi
c m
o
tion that
mem
ber com
e
s
into the
C
control area n
e
xt time.
Step1:
mem
ber
i
u
{
1
,
2
,
...,
}
in
select
s
the ran
dom v
a
lue
i
x
{
1
,
2
,
...,
}
in
as the
decryption
key from the
domain
[1
,
2
]
p
, and s
e
nd it to
C
with a se
cu
rity cha
nnel, the
C
kee
p
s it
se
cr
et
ly
;
Step2:
After t
he
C
gathers al
l legitimate m
e
mbe
r
s
de
cry
p
tion key as the key set
{}
i
x
, it
sele
cts th
e random
value
0
tp
from
[1
,
2
]
p
, it comp
utes
0
mo
d
t
Pg
p
and
i
P
{
1
,
2
,
...,
}
in
, whic
h is
s
hown as
follows:
12
3
1
12
1
3
1
2
3
1
2
12
(
...
)
1
(
.
..
..
.
)
2
(
.
..
..
.
)
mo
d
m
o
d
mo
d
m
o
d
......
mo
d
m
o
d
n
nn
n
ni
n
tx
x
x
x
p
tx
x
x
x
x
x
x
x
x
x
p
pt
x
x
x
x
n
Pg
p
g
p
Pg
p
g
p
Pg
p
g
p
(1)
Step3:
Th
e
C
is
sues
the public
enc
r
yption key
{}
i
eK
ey
P
{
1
,
2
,
...,
}
in
in the
bulletin boa
rd
B
, each en
cry
p
tion key ha
s a launch time;
Step4:
Th
e
C
use
s
the
valu
e belo
ngin
g
to set
{}
i
x
to co
mpute
0
mo
d
i
xt
i
Pg
p
,
and pu
blish
e
s
the value of
0
i
P
{
1
,
2
,
...,
}
in
in the bullet
i
n board
B
.
3.2 Ke
y
Use Phase
3.2.1 Encr
y
p
tion Phase
In the encrypt
ion pha
se, fro
m
the bulletin
board
B
member
k
u
{
1
,
2
,
...,
}
kn
gets up
-
to-date
{}
i
P
{
1
,
2
,
...,
}
in
a
nd
0
k
P
. When
k
u
need
s com
m
unicate with
memb
ers
j
u
{
1
,
2
,
...,
1
,
1
,
...,
}
jk
k
n
, it select
s two rand
om
values
s
and
r
from
[1
,
2
]
p
, and
comp
utes
mo
d
n
ps
gp
to repla
c
e the p
a
ram
e
ter
n
P
whi
c
h is the la
st
item in encryption key
{}
i
P
. Member
k
u
compute
s
the set
{m
o
d
}
i
rp
gp
{1
,
2
,
.
.
.
,
1
}
in
with
{}
i
P
{
1
,
2
,
...,
1
}
in
and
r
. At last,
k
u
en
crypt
s
t
he pl
aintext
m
into c
i
xphertext
mo
d
rs
mg
p
and sen
d
s
mo
d
rs
mg
p
and
{m
o
d
|
0
}
i
rp
gp
i
n
to
j
u
. So the process is
sh
own
as
follows
:
,,
{
}
()
m
o
d
,
{
m
o
d
}
i
p
i
rp
rs
rs
g
E
mm
g
p
g
p
{
1
,
2
,
...,
}
in
(2)
3.2.2 Decry
p
tion Phase
In the de
cry
p
tion ph
ase, the mem
b
e
r
j
u
{
1
,
2
,
...,
}
jn
use
s
t
h
e
se
cr
et
key
j
x
{1
,
2
,
.
.
.
,
}
jn
to decrypt th
e cixphe
rtext.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No. 2, Februa
ry 2013 : 809 – 818
812
Member
j
u
com
putes
set
{m
o
d
}
ni
ij
rp
x
gp
and
(1
)
0
mo
d
in
i
ij
n
rp
x
i
gp
{
1
,
2
,
...,
}
in
. We have that
(1
)
0
mo
d
in
i
ij
n
rp
x
i
gp
0
(1
)
mo
d
n
in
i
ij
i
rp
x
gp
(3)
1
()
mo
d
n
ji
i
rt
x
x
r
s
gp
Bec
a
us
e
{}
j
i
x
x
, s
o
1
()
n
n
ji
i
rx
x
r
s
r
s
. We have that
1
()
mo
d
m
o
d
n
ji
i
rt
x
x
r
s
rs
gp
g
p
(4)
Member
j
u
decrypts the cixp
hertext as foll
ows:
(1
)
0
/m
o
d
in
i
ij
n
rp
x
rs
i
mg
g
p
1
()
/m
o
d
n
ji
i
rt
x
x
rs
rs
mg
g
p
(5)
m
So the pro
c
e
ss of de
cryption is re
prese
n
ted as follo
ws:
{}
,
{
}
,
,
{
}
((
)
)
rp
p
ii
ji
xxg
r
s
g
DE
m
m
{
1
,
2
,
...,
}
in
(6)
3.3 Rek
e
y
Phase
A r
e
k
e
ying
is
tr
ig
g
e
r
e
d
w
h
en
me
mb
er
sh
ip
c
h
an
g
e
s
in
s
p
ac
e
ne
tw
ork
s
.
3.3.1 Membe
r
Joining
Whe
n
a ne
w membe
r
1
n
u
join
s in
sp
ace ne
twork,
1
n
u
sele
cts a
ne
w de
cryption key to
encryption ke
y
{}
i
P
{
1
,
2
,
...,
}
in
and tells the
C
its perio
di
c time.
Step 1:
Th
e
member
1
n
u
select
s a ran
dom value
1
n
x
as de
cryptio
n
key from
[1
,
2
]
p
, and it sends the secret value
1
n
x
to the
C
via a secu
re
cha
nnel (e.g., face-to
-
fa
ce
manne
r).
Step 2:
The
C
receives th
e
value
1
n
x
and ke
ep it se
cretly, afterwa
r
d
s
th
e
C
doe
s the
following:
(1)
The
C
sel
e
cts a
ran
d
o
m
value
'
t
from
[1
,
2
]
p
to
c
o
mpute
'
'
0
mo
d
t
Pg
p
,
update
s
1'
((
)
)
tt
ii
PP
{
1
,
2
,
...,
}
in
an
d
1'
00
()
jj
t
t
PP
{
1
,
2
,
...,
}
jn
with
'
t
and
{}
i
x
{
1
,
2
,
...,
}
in
, s
o
'1
'
ii
pp
t
t
;
(2)
Th
e
C
upd
ates th
e p
ubl
ic e
n
cryption
key
by com
puting
'
1
'
11
mo
d
n
tx
PP
g
p
,
1
'
1
()
n
x
nn
PP
and
1
'
11
()
m
o
d
in
p
x
ii
PP
g
p
,
{
1
,
2
,
...,
}
in
. The up-to-date pu
blic encryptio
n
key is ch
an
ged from
{}
i
P
to
'
{}
i
P
{
1
,
2
,
...,
1
}
in
, the
C
also compute
s
the value
'
1
1
0
mo
d
n
xt
n
Pg
p
;
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Key Agre
em
ent Protocol in
DSN (Jia
n Zh
ou)
813
(3) T
he
C
writes the en
crypti
on key
'
{}
i
P
and
0
{}
i
P
{
1
,
2
,
...,
1
}
in
in the
B
;
Step 3:
The
B
issue
s
the p
ublic
encrypti
on key
'
{}
i
P
and
0
{}
i
P
. The issue ti
me of
encryption ke
y is also give
n. The form o
f
'
{}
i
P
is sh
own as follows;
'
'
12
3
1
1
'
'
1
2
1
3
112
3
1
2
''
12
2
3
1
1
2
2
1
1
2
1
1
(.
.
.
)
'
1
(.
.
.
.
.
.
)
'
2
(
...
...
...
...
...
.
..
..
.
.
..
)
1
mo
d
m
o
d
mo
d
m
o
d
...
mo
d
m
o
d
nn
nn
n
ni
n
i
n
i
i
n
n
n
tx
x
x
x
x
p
t
x
x
x
xx
x
x
xx
x
p
p
t
x
x
xx
x
x
xx
x
x
x
x
x
x
x
x
x
n
n
Pg
p
g
p
Pg
p
g
p
Pg
p
g
p
P
''
11
2
1
...
...
'
mo
d
m
o
d
ni
n
n
pt
x
x
x
x
x
gp
g
p
(7)
Step4:
The member
1
n
u
gets the
update
d
publi
c
en
cry
p
tion key
'
{}
i
P
{
1
,
2
,
...,
1
}
in
and
1
0
n
P
from the
B
.
3.3.2 Membe
r
Leav
ing
Whe
n
a
me
mber
n
u
leaves sp
ace net
wo
rks,
the
C
up
dates the p
u
b
lic e
n
cryptio
n
key.
The othe
r me
mbers keep d
e
cryptio
n
key
unch
ang
ed. The procedu
re is sh
own as follows.
Step1 :
The member
n
u
sen
d
s the l
eavin
g messa
g
e
s
to the
C
wh
en
n
u
come
s int
o
the rang
e of the
C
’s co
ntrol
area;
Step2 :
If
C
accept
s the req
u
isition of lea
v
ing, the
C
does the followi
n
g
:
(1) T
he
C
sel
e
ct
s
ran
dom
v
a
lue
'
t
from
[1
,
2
]
p
, compute
s
1'
''
((
)
)
tt
ii
PP
and
'
0
mo
d
t
Pg
p
,
{
1
,
2
,
...,
1
}
in
, s
o
''
1
'
ii
pp
t
t
;
(2) Th
e
C
co
mputes
'
''
0
mo
d
t
Pg
p
,
'
''
''
11
(/
)
m
o
d
n
tx
PP
g
p
,
1
''
'
'
1
()
n
x
nn
PP
and
''
1
''
''
(/
(
)
)
m
o
d
in
px
ii
PP
g
p
,
{
2
,
3
,
...,
1
}
in
;
(3) T
he
C
computes
'
0
mo
d
i
xt
i
Pg
p
with
0
P
,
{
1
,
2
,
...,
1
}
in
;
(4) Th
e
C
writes th
e encryptio
n key
''
{}
i
P
{
1
,
2
,
...,
1
}
in
and
0
{}
i
P
{
1
,
2
,
3
,
...,
1
}
in
in the
B
;
Step 3:
T
he
B
issue
s
th
e p
ublic en
crypti
on
key
''
{}
i
P
and
0
{}
i
P
{
1
,
2
,
...,
1
}
in
.
The issue tim
e
of encryption key is al
so
given.
''
{}
i
P
is
s
h
own as
follows;
'
''
12
1
1
'
''
1
2
1
1
2
3
21
21
2
''
'
22
1
1
3
1
1
2
2
''
1
(
.
.
...
.
)
''
1
(
.
..
..
.
.
.
.
)
''
2
(
.
..
..
.
...
..
.
)
''
2
''
1
mo
d
m
o
d
mo
d
m
o
d
......
mo
d
m
o
d
mo
d
n
nn
n
n
nn
n
n
n
tx
x
x
p
tx
x
x
x
x
x
x
x
x
x
p
pt
x
x
x
x
x
x
x
x
n
pt
n
Pg
p
g
p
Pg
p
g
p
Pg
p
g
p
Pg
p
g
'
12
1
..
.
mo
d
n
xx
x
p
(8)
3.3.3 Ke
y
Replacing
W
h
en
a memb
e
r
n
u
need
s to u
pdate
th
e de
cryption
key from
n
x
to
'
n
x
, the
'
n
x
is
a
legitimate element in s
e
t
{}
i
x
{
1
,
2
,
...,
}
in
. The procedu
re is sh
own as follows.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No. 2, Februa
ry 2013 : 809 – 818
814
Step1:
The member
n
u
sele
cts a secret value
'
n
x
from
[1
,
2
]
p
at
rand
om an
d send
s
it to the
C
via a secure
cha
n
nel, it tells the
C
to replace
n
x
with
'
n
x
;
Step2:
If the
requ
est is all
o
we
d, the
C
does the follo
wi
ng:
(1) T
he
C
computes
'
{}
i
P
{
1
,
2
,
...,
}
in
with
{}
i
P
{
1
,
2
,
...,
}
in
and
'
nn
x
xx
;
'
1
1
2
32
1
12
2
1
'
1
()
'
2
((
)
)
'
3
(
(
(
.
..
(
(
)
)
..
.
)
)
'
......
nn
n
nn
ii
i
n
n
n
n
pt
x
t
x
pt
x
t
x
p
p
p
p
t
xt
xt
x
p
p
p
p
p
t
xt
x
t
xt
xt
x
i
Pg
Pg
g
Pg
g
Pg
g
(9)
(2) T
he
C
sele
cts a
ran
dom
value
'
t
from
[1
,
2
]
p
, comp
utes
'1
'
'
mo
d
i
pt
t
i
Pg
p
,
'
mo
d
t
gp
and
'
0
mo
d
i
xt
i
Pg
p
with
{}
i
x
;
(3)
The
C
write
s
u
pdate
d
d
e
c
ryption
key
'
{}
i
P
{
1
,
2
,
...,
}
in
and
0
{}
i
P
{
1
,
2
,
...,
}
in
in the
B
.
Step 3:
T
he
B
issue
s
the pu
blic en
cryptio
n
key
'
{}
i
P
and
0
{}
i
P
. T
he issue time
of
encryption ke
y is also give
n again.
'
{
1
,
2
,
...,
}
i
Pi
n
is shown as follo
wing;
''
12
3
1
''
'
12
1
3
1
2
3
1
2
''
12
(
...
)
1
(
...
..
.
)
2
(
.
..
..
.
)
mo
d
m
o
d
mo
d
m
o
d
......
mo
d
m
o
d
n
nn
n
ni
n
tx
x
x
x
p
tx
x
x
x
x
x
x
x
x
x
p
pt
x
x
x
x
n
Pg
p
g
p
Pg
p
g
p
Pg
p
g
p
(10)
3.3.4 Mainte
nance Pha
s
e
W
h
en
a me
mb
e
r
i
u
{
1
,
2
,
...,
}
in
come
s i
n
to the
ran
g
e
of the
C
after
a time period,
the
i
u
’s d
e
cryp
tion key rem
a
ins un
ch
ang
ed
sin
c
e
the
i
u
doe
s not im
plement
the
pro
c
e
s
s o
f
rekeying. Me
anwhile, the
i
u
gets the up
-t
o-date p
ubli
c
encryption
{}
i
P
and
0
i
P
from the
B
in
time. But if an existing
me
mber’
s
p
e
rio
d
i
c time in
nex
t meeting
with the
C
is mo
re than th
e p
r
e-
config
uratio
n
time, the
C
would implem
e
n
t rekeyin
g
with memb
er leaving’s re
key method t
o
revoke the overtime mem
b
er’s le
gitimat
e
identity.
4. Securit
y
Proof
Theorem 1
T
he AKMSN key manage
m
ent scheme
meets
key indepe
nden
ce.
Proof
Key in
depe
nden
ce
i
s
note
d
that
an PPT atta
cker who
com
p
romi
sin
g
so
me key
can not com
putes othe
r keys.
Fi
rstly
any
elem
ent
i
x
belo
nging
to set
{}
i
x
is sel
e
cted as
decryption
ke
y at rand
om
by membe
r
i
u
,
so it
i
s
dif
f
i
cult
ly
t
o
gue
ss
su
c
c
e
ssf
ul
ly
t
he se
cr
et
value
i
x
for member
()
j
ui
j
witho
u
t any inform
ation. Seco
n
d
ly, whethe
r
comp
romi
se
d
the
j
u
has
capabilit
y to crack the value
i
x
acco
rd
ing to
{}
i
P
in publ
ic ch
ann
el or
not. We supp
ose
d
that
j
u
ha
s
kn
own
any el
e
m
ent value
i
n
set
{}
i
x
exce
pt
i
x
,
so
ea
ch
i
P
re
presents as
mo
d
ii
i
ax
b
i
Pg
p
. Thu
s
the
qu
estion
of
re
so
lving value
i
x
with
k
n
ow
n va
lu
e
{}
i
P
,
{}
i
a
an
d
{}
i
b
is red
u
ce to the que
stion o
f
resolving val
ue
i
x
with kno
w
n value
mo
d
i
x
gp
, obviously this
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Key Agre
em
ent Protocol in
DSN (Jia
n Zh
ou)
815
is a DH p
r
obl
em [25], whi
c
h is
negli
g
ib
le to crack fo
r the
j
u
. Above all It is neg
ligible that a
disclo
sure of a decryptio
n key co
mpromise
oth
e
r
decryption keys, so the
AKMSN sch
e
me
meets key
indepe
nden
ce.
Theorem 2
T
he AKMSN schem
e gua
ra
ntees b
a
ckwa
r
d se
cu
rity ag
ainst any PPT attacker.
Proof
It is su
re that a ne
w joining m
e
mber
1
n
u
has capability of decryptin
g a secret
whi
c
h i
s
gen
erated
with
t
he u
pdate
d
e
n
cryptio
n
key
'
{}
i
P
, b
e
c
a
us
e it’s
de
cr
yp
tio
n
k
e
y
!
n
x
is
the root of e
q
uation
1
''
1
()
(
)
n
i
i
f
xt
x
x
{
1
,
2
,
...,
1
}
in
. However, sin
c
e th
e value
!
n
x
is not th
e roo
t
of equatio
n
1
()
(
)
n
i
i
f
xt
x
x
, a se
cret wit
h
onb
eforeup
date en
cryption
key
{}
i
P
{
1
,
2
,
...,
}
in
is
not be
de
crypted
by th
e
!
n
u
. At time, th
e value
0
P
is
enc
r
ypted with the
se
cret
set
{}
i
x
{
1
,
2
,
...,
}
in
, it is negligibly
that member
1
n
u
cra
ck t
h
e
0
{}
i
P
d
ue to
!
n
x
is
not belon
g to
set
{}
i
x
{
1
,
2
,
...,
}
in
. In a word, the AKMSN sche
m
e
gua
rante
e
s backward
s
e
c
u
rity when a new member joins
in net
work
.
Theorem 3
T
he AKMSN schem
e gua
ra
ntees fo
rwa
r
d
security agai
nst any PPT attacker.
Proof
Withou
t loss
of gen
e
r
ality, the exited mem
ber i
s
n
u
. it has revo
ked the
cap
abi
lity
of de
crypt
se
cret,
whi
c
h
in
clud
e two a
s
pect
s
, on
e i
s
it’s secret
ke
y
n
x
is
not
de
cryption key for
encryption ke
y
{}
i
P
; the othe
r
one i
s
th
e
n
u
ca
n not
revise t
he e
n
cryption
key
{}
i
P
to res
e
t the
value
n
x
as de
cryption key.
Point to
the first on
e, after membe
r
n
u
leaves sp
ace ne
twork, the
value
n
x
is
not belo
ng
to the set
{}
i
x
, s
o
the
n
x
is not th
e
root of e
q
u
a
tion
1
''
1
()
(
)
n
i
i
f
xt
x
x
. If
the member
n
u
uses
n
x
to decrypt a
ciph
ertext, which is sh
own a
s
follows
:
1
0
1
1
1
(1
)
0
(1
)
()
mo
d
mo
d
mo
d
in
i
in
n
in
i
in
i
n
ni
i
n
rp
x
r
s
i
rp
x
rt
x
x
r
s
gp
gp
gp
(11)
As the
eq
uat
ion
1
1
()
mo
d
m
o
d
n
ni
i
rt
x
x
r
s
rs
gp
g
p
is not t
r
ue, so
n
u
can
not get
plaintext
m
with
n
x
. Point
to t
he s
e
c
o
nd one, after the
n
u
leaves sp
ace netwo
rk, the
C
re-
sele
cts the
ra
ndom
value
'
t
, and
comp
utes
0
{}
i
P
, so the
n
u
can not reset
n
x
as a decrypti
o
n
key ag
ain be
cau
s
e th
e probability of g
e
tting the value
0
P
is negligi
b
le ba
sed
on
DH p
r
o
b
lem.
Above all, the AKMSN sch
eme gua
ra
ntees forwa
r
d
secu
rity when
a membe
r
lea
v
es.
5. Performan
ce Analy
ses
5.1 Storag
e Cos
t
Each me
mbe
r
ha
s a de
cry
p
tion key an
d
a encry
ption
key, their st
orag
e cost are a unit
and
1
n
units as
encryption ke
y is comp
ose
d
of
1
n
parts. So storage
co
st of decryptio
n key
is
(1
)
O
and
st
ora
g
e
cost
of e
n
cryption key is
()
On
,
s
o
the
r
e
la
tion
s
h
ip
be
tw
een
s
t
or
ag
e
c
o
st o
f
our p
r
op
osed
encryptio
n key and sp
ace
netwo
rk
scal
e
is a linea
r correlation.
5.2 Round
Complexit
y
The spa
c
e n
e
twork
time delay
is sh
orten
effi
ciently
if roun
d co
mplexity is re
duced. In
agre
e
ing the
sha
r
ed e
n
cry
p
tion key, the
round n
u
mb
er between
membe
r
and
C
is two, in first
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No. 2, Februa
ry 2013 : 809 – 818
816
roun
d memb
er su
bmits a
rand
om
se
cret as a decryption key to
C
, the seco
nd round, the
C
sen
d
s the
en
cryption
key t
o
mem
b
e
r
s a
nd b
u
lletin
bo
ard. In
rekeying p
h
a
s
e, if
a ne
w memb
er
joins n
e
two
r
k, firstly the ne
w mem
ber
se
nds th
e de
cry
p
tion key to
C
,
se
con
d
ly
t
he
C
update
s
the publi
c
e
n
c
ryption
key on bull
e
tin bo
ard, finally th
e ne
w me
mb
er
save
s the
update
d
sha
r
ed
encryption
ke
y from bulleti
n boa
rd. If a membe
r
l
eav
es the n
e
two
r
k, the exitin
g membe
r
o
n
ly
sent a l
eavin
g me
ssage to
the
C
, and th
e
C
update
s
the
encryption
key in bulletin
board. So
the roun
d co
mplexity of rekeying i
s
c
o
n
s
tant value when mem
bership ch
ang
es.
5.3 Compu
t
a
t
ion Comple
xit
y
In con
s
ulting
sha
r
ed e
n
cryption key p
hase,
each membe
r
sel
e
cts rand
om value as
decryption
key and th
e
C
implem
ents
1
1
n
i
n
i
C
expon
entia
l modul
ar
co
mputation
s
f
o
r
encryption ke
y.
In en
cryptio
n
pha
se,
en
crypter
sel
e
ct
s two
ran
d
o
m
value
an
d impl
ement
s
2
n
expone
ntial modula
r
com
putation
s
,
de
crypte
r
imple
m
ents
(1
)
/
2
nn
expo
nential mo
du
lar
comp
utation
s
for decrypti
ng. In order to
reduce t
he burden o
f
computatio
n, a few sa
me
expone
ntial modula
r
co
mputation co
uld be imple
m
ented by the
C
in the initial phase for
decrypter, th
e numb
e
r of
expone
ntial m
odula
r
com
putation
s
is redu
ce
to
n
for decrypter, b
u
t
the numbe
r o
f
exponential
modula
r
com
putation
s
is in
cre
a
sed to
2
1
(3
)
/
2
1
n
i
n
i
Cn
n
.
In memb
er j
o
ining
ph
ase
,
the
C
sel
e
ct
s a
rand
om
value
and
i
m
pleme
n
t
47
n
expone
ntial modula
r
com
putation
s
, joining mem
ber
sele
ct a ran
d
o
m value.
In member
exiting pha
se, the
C
sele
cts a ra
ndo
m value and
implement
s
41
n
expone
ntial
modula
r
com
putation
s
, if membe
r
repl
ac
e
s
its
de
cry
p
tion key with
new
se
cret value,
it needs to se
lect a ran
dom
value,
C
implements
43
n
exponential modul
a
r
com
putation
s
.
From a
bove
analysi
s
, the relation
shi
p
b
e
twee
n the membe
r
’s
co
mputation
co
mplexity
and spa
c
e ne
twork scale i
s
a linear corre
l
ation.
5.4 Scalability
Due to the
ra
nge of rekeyi
ng is limited t
o
singl
e mem
ber, so it is e
a
sily to ne
w
membe
r
join in
net
wo
rk
without
oth
e
r m
e
mb
ers
agre
e
, in
oth
e
r
wo
rd, th
e
perfo
rman
ce
of the
pro
p
o
s
ed
scheme i
s
no
t deteriorated
significantly whe
n
t
he sca
l
e of network becom
es m
o
re big
g
e
r
. And
the procedu
re of exten
d
in
g net
work
si
ze is si
m
p
le, n
e
w m
e
mb
ers
sele
ct o
ne
ra
ndom
num
be
r a
s
decryption ke
y and agre
e
p
ublic e
n
cryption key with th
e serve
r
C
.
5.5 Commun
i
cation Ov
erload
As the 1-affe
cts
-n
pro
b
le
m is solved i
n
our propo
sed AKMSN schem
e, the schem
e’s
comm
uni
cati
on overl
oad i
s
minimal, o
n
ly the
updat
ed memb
er
partici
pate
s
i
n
re
key process.
The legitimat
e
membe
r
s’
comm
uni
cati
on overl
oad i
n
clu
d
e
s
se
nd
ing a secret
value to
C
an
d
getting en
cry
p
tion key th
at has
1
n
parts. So the legitimate me
mbers’ com
m
unication
overloa
d
is
2
n
.
5.6 Compari
s
on
Curre
n
tly spa
c
e net
work
key manage
m
ent scheme
s
are the
comb
ine of several
groun
d
key man
age
ment sche
me
s, so th
e com
pari
s
on
re
sult
is only bet
we
en ou
r propo
sed
schem
e
and
exiting key m
anag
ement
schem
es. In T
able 1,
we
co
mpare some
of the cu
rrent
ly protocols
with
the propo
se
d
schem
e ag
a
i
nst the foll
o
w
ing
criteria,
inclu
d
ing
key
indep
end
en
ce, com
putatio
n
,
roun
d, sto
r
ag
e and
seve
r cap
ability. In comp
ari
s
on
to other
sche
mes, the rou
nd cost of th
e
AKMSN sche
me is
con
s
ta
nt value, so it
is un
re
late
d
to spa
c
e
net
work
scale. T
he complexit
y
o
f
comp
utation
and
storag
e is h
a
s a
linear correlation
with
netwo
rk scale. Ho
weve
r, the
comp
utation
compl
e
xity is more tha
n
the schem
es b
a
se
d on
som
e
topology
structu
r
e, such
as
Oc
topus
and
DH-LK
H
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Key Agre
em
ent Protocol in
DSN (Jia
n Zh
ou)
817
Table 1. Co
m
pari
s
on of So
me Gro
und K
e
y Manage
m
ent Protocols
in Agreei
ng Key Phase
Scheme
Ke
y
Indep
enden
ce
Computation
round
storage
Ke
y
Manag
er
Server
capability
AKMSN Y
e
s
n
2
2
n
Limited
Ingemarson et
al.
Y
e
s
n
1
n
1 NO
GDH
Yes
1
n
n
n
NO
Octopus
Y
e
s
34
n
4
22
4
n
1 NO
STR Y
e
s
n
n
n
NO
DH-LKH
Yes
2
lo
g
n
2
lo
g
n
2
lo
g
1
n
NO
BD Y
e
s
3
2
1
NO
GKMP
Y
e
s
null
1
2
Powerful
Secure Lock
Y
e
s
Chinese
Remainder
Theore
m
2 1
Powerful
LKH Y
e
s
hash
2
lo
g
n
2
lo
g
n
Powerful
SAKM NULL
0
1
1
Limited
Kronos
NULL
hash
0
1
No online
Probabilistic key
sharing
Y
e
s
0
0
Part of ke
y
s
Off-line and ke
y
pool
6. Conclusio
n
s and Fu
tur
e
Work
In this pape
r, the AKMSN schem
e is p
r
o
posed, ea
ch membe
r
ha
s a different de
cryption
key and ha
s a common
publi
c
encryp
t
ion key for se
cure co
m
m
unication. Whe
n
a me
mber
leaves or join
s the
net
work, the range
of
re
keyi
ng
is li
mited to
singl
e mem
ber. A
nd n
on-upd
ated
membe
r
s’
de
cryption
key still ke
ep
s its validit
y to public
en
crypti
on key. As t
he 1
-
affect
s
-n
probl
em i
s
de
al with
succe
ssfully, the
p
e
rform
a
n
c
e
o
f
the pu
rpo
s
e
d
sch
e
me i
s
not rel
a
ted to
the
mobility, topology stru
cture
and
netwo
rk scale. Me
an
while the g
o
o
d
scalability of the propo
sed
scheme i
s
a
d
vantage of
extending th
e netwo
rk
scale
. In the future, we pl
an
to study more
efficient key
manag
eme
n
t scheme
s
for
spa
c
e n
e
two
r
ks.
Referen
ces
[1]
Posner E
C
, Stevens
R. Spa
c
e commu
nica
tion-Past, Pres
ent, an
d F
u
tur
e
.
IEEE Comm
unic
ations
Maga
z
i
ne
. 1
9
8
4
; 22(5), 8-21.
[2]
Z
eng F
e
ng, Y
ao
La
n, Ch
en
Z
h
iga
ng. Imp
a
c
t of to
p
o
lo
g
y
and
traffic o
n
i
n
terferenc
e
an
d ro
uting
i
n
IEEE 802.11
w
i
reless mes
h
n
e
t
w
o
r
k.
Te
lkomn
i
ka
. 20
12; 10(
4): 823-8
30.
[3]
Mantoro
T
edd
y, Z
a
k
a
ri
ya
A
ndri. S
e
curi
ng
e-ma
il c
o
mm
unic
a
tion
us
in
g h
y
b
r
id
cr
ypt
o
s
y
st
em o
n
andr
oid-
bas
ed mobil
e
dev
ices
.
Telkom
nika.
2
012; 10(
4): 827
-834.
[4]
Wen JP, Zhang YJ, Zhao B.
L-ba
nd SA
R-pr
ocessor
for the
Chi
nes
e SAR
satellit
e
. ICCE
A 20
04-2
0
0
4
3rd Internati
o
n
a
l Co
nfere
n
ce
on Co
m
putati
o
nal El
ectroma
g
netics an
d
its Appl
icatio
ns. BRSI. 2004; 3:
399-
402.
[5]
Bell DJ, Ces
a
r
one R, El
y T
,
Ed
w
a
r
d
s C, T
o
w
n
es S.
Mars netw
o
rk: a Mars orbitin
g
co
mmu
n
ic
ations
and
nav
igati
o
n
satell
ite co
nstellati
on
.
IEEE Aerospace Conferenc
e
Proc
eedings. P
a
sadena. 2000;
7:
75-8
8
.
[6]
Harn
e
y
H, Mucken
h
irn C.
Group Key Manag
ement
Protoc
ol (GKMP) Architecture
. Internet
Engi
neer
in
g T
a
sk F
o
rce. RF
C: 2093. 1
997.
[7]
Chio
u GH, C
hen
W
T
. Secure Bro
adc
ast usin
g S
e
cur
e
L
o
ck.
IEEE Transactions
on S
o
ftwar
e
Engi
neer
in
g
. 1989; 15(
8): 929
-934.
[8]
Ingemars
on I,
T
ang D. W
ong C. A
Conf
erenc
e Ke
y
Di
stributio
n S
y
st
em.
IEEE Tra
n
sactions on
Information Theory
.19
82; 28(
5), 714–
72
0.
[9]
Steiner M, T
s
u
d
ik G, W
a
idne
r M.
Diffie-Hell
ma
n key distri
butio
n exten
d
e
d
to grou
p co
mmu
n
icati
o
n
.
Procee
din
g
C
C
S '
96 Procee
din
g
s of the 3rd AC
M confe
r
ence o
n
Com
puter an
d com
m
unic
a
tion
s
securit
y
. N
e
w
York.199
6; 3: 31-37.
[10]
Becker
C, Will
e U.
Co
mmu
n
i
c
ation
co
mp
lex
i
ty of gr
ou
p ke
y distri
butio
n
.
CCS '
9
8 Proc
e
edi
ngs
of t
h
e
5th ACM confe
r
ence o
n
Com
puter an
d com
m
uni
c
a
tions se
curit
y
. Ne
w
Y
o
r
k
.1998; 5: 1-6.
[11]
Burmester M, Desme
d
t Y.
A
secure and efficient conf
er
ence key distribution system
. Lecture Not
e
s
in Com
puter S
c
ienc
e. 199
4; 950(1
995): 2
75–
286.
[12]
Yacin
e
C, Ham
i
da S.
Group K
e
y
M
ana
gem
e
n
t Protocols: A Novel T
a
xon
o
m
y
.
Internati
o
n
a
l. Journ
a
l of
Information T
e
chno
logy
. 2
005
; 2(2): 105-11
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No. 2, Februa
ry 2013 : 809 – 818
818
[13]
Chu
ng KW
, Goud
a M, L
a
m
SS.
Secure
grou
p
co
mmu
nicati
ons usin
g
key grap
hs
.
Net
w
ork
i
ng.
IEEE/ACM
T
r
a
n
sactio
ns on.2
000; 8(1): 1
6
-3
0.
[14]
Kim Y, Perrig A,
T
s
udik G.
Tree-b
a
sed gr
o
up key agre
e
m
ent
.
ACM T
r
ansactions
o
n
Inf
o
rmatio
n
a
n
d
S
y
stem Sec
u
rit
y
. 20
04: 7(1): 6
0
-96.
[15]
Lin
Y, Ko
ng
XW
, W
u
GW
, Lin C. T
r
ee-bas
ed M
u
lticast K
e
y Ma
nag
eme
n
t in
ub
iq
uitou
s
comp
ut
i
n
g
envir
onme
n
t.
Internati
o
n
a
l Jo
urna
l of Ad Ho
c and Ub
iq
uito
us Co
mp
uting
.
201
1; 8(1): 27-
35.
[16]
Omar C, Anis K.
A logical n
e
i
gh
bor tree se
cure gro
up co
mmu
n
icati
on s
c
he
me for w
i
reless se
nsor
netw
o
rks
.Ad Hoc Net
w
orks. 2
012; 10(
7): 141
9-14
44.
[17]
Steer D, Stra
w
c
z
y
nsk
i
LL, Diffie W
,
W
e
iner M.
A Secure Audi
o T
e
leco
nfe
r
ence Syste
m
. CRYPT
O
'88
Procee
din
g
s o
n
Advanc
es in
cr
y
p
to
lo
g
y
. N
e
w
York. 1
988;
520-
528.
[18] A
Perrig.
Efficient C
o
ll
abor
ative key M
a
nag
e
m
ent
pro
t
ocols for Se
cure Auto
no
mous Grou
p
Co
mmun
icati
o
n.
Internati
o
n
a
l
W
o
rkshop
o
n
Cr
yptogr
aph
ic
techni
qu
es a
n
d
E-commerc
e: Hon
g
K
ong.
199
9; 1-11.
[19]
W
ong C, Gou
da KM, Lam
SS.
Secure G
r
oup C
o
mmu
n
i
catio
n
s Usi
n
g
Key Graphs.
IEEE/ACM
T
r
ansactions o
n
Net
w
ork
i
ng.
200
0; 8(1): 16–
30.
[20]
Klao
udat
ou E,
Konsta
ntino
u
E.
A Survey
on C
l
uster-B
a
s
ed Grou
p Ke
y Agree
m
ent
Protocols fo
r
WSNs
. Communic
a
tions S
u
rve
y
s & T
u
torial
s. 2011; 13(
3): 429-
442.
[21]
Setia S,
Kous
sih S, J
a
j
odi
a
Har
der E.
Kr
onos: A
sca
la
ble
gro
u
p
re-k
eyin
g a
ppr
oac
h for s
e
cur
e
mu
lticast.
IEEE S
y
mp
osium o
n
Securit
y
an
d
Privac
y
.
N
e
w
Y
o
rk. 2000; 1: 2
15-2
2
8
[22]
Eschen
au
er L,
Gligor VD.
A
key-man
age
ment sche
m
e fo
r distribut
ed se
nsor n
e
tw
orks
. Procee
din
g
CCS '
02 Pr
oce
edi
ngs
of the
9th ACM c
onf
erenc
e
o
n
Co
mputer a
nd c
o
mmunicati
ons
securit
y
.
N
e
w
York. 2002; 1:
41-4
7
.
[23]
Ho
w
a
rt
h MP, Iy
engar S, Sun ZL, Cruickshank H.
D
y
nam
ic
s of Key
M
anagement in Sec
u
re Satellit
e
Multicast.
IEEE Journal on Selected
Areas in
Comm
unications
. 2004; 2
2
(2)
:
308-31
9.
[24]
Arslan MG, Al
agoz F
.
Sec
u
ri
ty issues a
nd
perfor
m
a
n
ce st
udy of key
ma
nag
e
m
ent tec
h
niq
ues ov
er
satellit
e links
.
2006
11th Internati
o
n
a
l W
o
rksho
p
on C
o
mputer Ai
de
d Mode
lin
g a
nd Des
i
gn
o
f
Commun
i
cati
o
n
Links a
nd N
e
t
w
ork.T
r
ento. 2006: 12
2-1
28.
[25]
Eng B, Ro
ber
t HD, Z
hu HF
.
Variatio
ns of
Diffie-He
ll
ma
n Prob
le
m
. Lecture Notes
in Computer
Scienc
e. 200
3; 836: 30
1-31
2.
Evaluation Warning : The document was created with Spire.PDF for Python.