TELKOM
NIKA
, Vol. 11, No. 10, Octobe
r 2013, pp. 6
187 ~ 6
192
ISSN: 2302-4
046
6187
Re
cei
v
ed Ap
ril 19, 2013; Revi
sed
Jul
y
1
5
, 2013; Acce
pted Jul
y
25,
2013
A New P2P Identity Authentication Method Based on
Zero-Knowledge Under Hybrid P2P Network
Xi
y
u
Pang*
1
, Cheng
Wang
1
, Yuhong Zhang
2
1
Department o
f
Information T
e
chn
o
lo
g
y
an
d Electric
Eng
i
ne
erin
g, Shan
don
g Jiao T
ong Un
iversit
y
, Ji
na
n,
250
35
7, Chin
a, Ph./F
ax: +
86-531—
80
687
92
0
Email: xi
yu
pan
g@1
26.com
2
Department o
f
Computer T
e
chno
log
y
, W
u
Xu
n Hi
gh Sch
o
o
l Li
ao Ch
en
g, Chin
a, +
86-63
5—73
21
046
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: xi
yu
pa
ng@
1
26.com
A
b
st
r
a
ct
On the basis o
f
analy
z
i
n
g the
shortcomin
gs of tr
adition
al a
u
thentic
atio
n
mec
h
a
n
is
m synthetica
lly
,
the p
aper
pres
ents a
new
ki
n
d
of P2
P i
denti
t
y authe
nticati
o
n
mod
e
l. In
t
h
e
new
P2P I
den
tity authe
nticati
o
n
mo
de
l, it authe
nticat
es the i
d
entity of nod
es
by using a n
e
w
Z
e
ro-Know
ledg
e pro
o
f ide
n
tificatio
n
sche
m
e
w
h
ich co
mbin
es the
adv
ant
age
of RSA.
Besid
e
s, dur
in
g the
proc
ess
of val
i
dati
n
g
nod
es
’
ide
n
tity, CA
(Center Aut
h
e
n
ticatio
n
) do
es
n
’
t
nee
d to par
ticipate i
n
.
At last, Simu
latio
n
results show
that the new
P
2
P
ide
n
tity authe
nticatio
n metho
d
can i
m
pr
ov
e the safety of network effectively.
Ke
y
w
ords
:
Z
e
ro-Know
le
dg
e,Identity Auth
ent
icatio
n, CA, RSA
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
P2P has
bee
n wid
e
ly appl
ied in the fiel
ds
su
ch
a
s
d
i
stribute
d
co
mputing, file
sha
r
ing
and
swappi
n
g
, insta
n
t me
ssagin
g
, net
work-attache
d
sto
r
ag
e, web
sea
r
ching,
network
ca
ching,
etc.. By far, the mo
st successful ap
plica
t
ion of P2P h
a
s b
een th
e f
ile sh
arin
g an
d swap
ping
a
s
eviden
ced
by
the P2P
tool
s li
ke
Na
pste
r, Gnut
ella, Bit
T
orrent
and
e
D
on
key
whi
c
h have
be
co
me
the main
sou
r
ce
s
of inte
rnet traffic.
Howeve
r,
a
c
co
mpanie
d
with
the g
r
eat
su
ccess
of the
s
e
to
o
l
s
,
ma
n
y
n
e
w
pr
o
b
l
ems
ha
ve
ar
is
en
, w
h
ic
h ar
e
mo
s
t
ly
c
a
us
ed
b
y
th
e lack o
f
tr
us
t,
s
u
ch
as
the malicio
us
node
s that de
liberately spread the
un
aut
hori
z
ed, tamp
ered o
r
even
virus files.
In P2P system, each entit
y gets involved in
the internet ra
ndoml
y
and voluntarily and
each entity varie
s
in its
ca
pacity an
d rel
i
ability. T
he relation
ship a
m
ong e
n
tities is mo
re
simil
a
r to
the compli
cat
ed so
cial rela
tionshi
p [1]-[3]. The
peer-to
-
pee
r net
work lacks centrali
zed mo
nitorin
g
mech
ani
sm a
nd the re
cog
n
ize
d
credibl
e third par
ty authority, so
the c
onventio
nal cent
rali
ze
d
se
curity infra
s
tru
c
ture (for example, PKI and
Kerb
ero
s
) i
s
no l
onge
r suita
b
l
e
for use, which
make
s it difficult to realize t
he trus
t ma
na
gement in the
peer-to-p
e
e
r
netwo
rk.
The re
putatio
n system i
s
a
solution to th
e pro
b
lem of
the lack of tru
s
t and it can
help to
achi
eve the followin
g
two
obje
c
tives:
1. With the
re
ward o
r
p
enal
ty of reputatio
n
a
s
the i
n
ce
ntive, most n
ode
s which in
tend to
seri
ou
sly tra
n
sa
ct a
r
e
co
mpelled
to
e
x
amine th
eir own
behavi
o
rs a
nd th
ey are in
clin
ed
to
cho
o
se a fixed identity, which will redu
ce
the difficulty i
n
identity managem
ent;
2. Wh
en th
e
nod
es in th
e sy
stem
are
in n
eed
for
help, they
ca
n resort to
the trust
system i
n
ch
oosi
ng a
cred
ible strang
e n
ode a
s
a t
r
an
sa
ction o
b
je
ct, which
will e
x
pand the t
r
u
s
t
relations
h
ip network
.
Ho
wever, the
trust syste
m
can o
n
ly help
use
r
s
sele
ct more
credible
node
s, but it fails to
provide
a m
e
ch
ani
sm
wh
ich
ca
n en
a
b
le u
s
e
r
s to
verify that their t
r
an
sa
cti
on o
b
je
ct is
the
sele
cted no
d
e
. Therefo
r
e,
in addition to the trust
system, it is nece
s
sary to provide a suita
b
le
identity verification mechan
ism for the se
curi
ty man
a
g
e
ment in the peer-to-pee
r
netwo
rk.
The fundam
e
n
tal con
c
ept
of zero
kno
w
l
edge is
that
one pa
rty (the prover) atte
mpts to
convin
ce the
other p
a
rty (the verifi
er) th
at a stateme
n
t is true, wi
t
hout providin
g the verifier
any
useful info
rm
ation.
Feige
-
Fiat-Sh
a
mir al
gorith
m
is the first
one
that i
s
b
a
se
d on th
e
zero-kn
o
wl
ed
ge proof
[7], [8] and its fund
ame
n
tal co
ncept i
s
that P (the
prove
r) i
n
itially send
s
a nu
mber X to
V
(the
verifier), V se
nds b
a
ck P a
rand
om num
ber, then P send
s V a nu
mber Y en
crypted by the P’s
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 10, Octobe
r 2013 : 618
7 –
6192
6188
private
key a
nd the
ran
d
o
m
numb
e
r,
a
nd the ve
rifi
er identifies the
prove
r
by ch
ecking wh
eth
e
r
X
and y satisfy
a certai
n formula. Its se
curity is
in p
r
opo
rtion to
2-kt ,but it has the follo
wing
defect
s
:
1. If a malicio
us no
de tamp
ers th
e information
of P’s
identity certifi
c
ate an
d re
pl
ace
s
V’s
publi
c
key wi
th its public
key, its private key
will be
use
d
in com
puting Y and
y can pass the
verification of
V. In
this wa
y, the malici
ous no
de can
comp
ass its il
legal pu
rpo
s
e
.
2. If a mali
cio
u
s
nod
e inte
rcept
s the
nu
mber X
sent
by P to V, it
may de
rive y
accordin
g
to the final formula, so that
the malicio
us
node
can imp
e
rsonate P to
gain som
e
in
terest
s.
Becau
s
e
the
external
inf
o
rmatio
n ex
chang
e is tim
e
-con
sumi
ng,
this
algo
rith
m is
no
t
ideal for the
applicatio
ns like sma
r
t card
. Da
eHun
Nyang and
JooSe
c
k Son
g
prop
osed an
identificatio
n
sch
eme b
a
s
ed o
n
the zero-kn
o
wl
ed
ge pro
o
f an
d the sche
me req
u
ire
s
less
comm
uni
cati
on traffic a
n
d
comp
utation,
whi
c
h i
s
suitable for
sm
art card
syst
e
m
. Becau
s
e t
h
e
algorith
m
is si
mple, the se
curity of the scheme ha
s yet to be improv
ed.
Therefore, thi
s
pa
pe
r p
r
o
pose a
ne
w i
n
te
ra
ctive ide
n
tification
scheme
und
er t
he P2P
netwo
rk b
a
se
d on the existing ze
ro-kn
o
w
led
ge ide
n
tification, whi
c
h security is base
d
on th
e
difficulty in the factori
z
atio
n of
large nu
mbers an
d cracking
RSA.
2. The Frame of Ne
w
Id
e
n
tit
y
Authen
tication
Bas
e
d On Zero
-Kno
w
l
e
dge
The
cu
rrent
P2P frame
w
ork fall
s into
thre
e
categ
o
rie
s
: the
ce
ntralized
stru
cture,
the
compl
e
te dist
ributed p
a
ttern and the p
r
ese
n
tly popul
ar semi-cent
ralize
d
and
semi-di
s
trib
uted
pattern. The
semi
-centrali
zed a
nd se
mi-di
s
tribute
d
pattern inte
grate
s
the a
d
vantage
s of
the
centralized P
2
P qui
ck search
and th
e
co
mplete di
stri
b
u
tion of p
u
re
P2P. Therefo
r
e, the
existin
g
P2P appli
c
ati
ons m
o
stly a
dopt semi-ce
n
tralized a
n
d
semi
-di
s
tribu
t
ed pattern,
su
ch a
s
Ka
zaa
model, Skype
,
etc. Besides the regist
ry manag
em
e
n
t center, the
r
e
is no ce
ntrali
zed
serve
r
. The
se
curity m
o
d
e
l propo
se
d i
n
this pa
per i
s
u
nde
r the
semi-centralized a
nd
semi
-distrib
uted P
2
P
netwo
rk.
The ne
w in
teractive ide
n
tity authenticat
ion
sche
me inclu
d
e
s
two pha
se
s: the
prep
ro
ce
ssin
g before a
u
th
enticatio
n an
d
the pha
se o
f
identity authenticatio
n:
1. The pha
se of prep
ro
cessing b
e
fore
authenti
c
at
ion. Du
ring
the early st
age of
establi
s
hi
ng t
he sy
stem, e
a
ch
nod
e g
o
e
s to th
e ma
nagem
ent
ce
nter fo
r regi
stration a
nd t
h
e
manag
eme
n
t cente
r
will all
o
cate to e
a
ch
node a
uniq
ue identity, public
key an
d
private key a
nd
so on.;
2. The pha
se
of identity authent
ication. It is the process of
identity authentication
among
node
s by u
s
i
ng the ne
w
zero
-kno
wledg
e identity aut
hentication m
e
thod. Durin
g
this process,
the
manag
eme
n
t cente
r
is not required to be
involved.
2.1. Preproc
essing be
for
e
Authen
tica
tion
Duri
ng th
e e
a
rly
stage
of
system
op
era
t
ion, ea
ch
no
de firstly go
e
s
to th
e m
a
n
ageme
n
t
cente
r
for reg
i
stration d
u
ri
n
g
whi
c
h the
manag
em
e
n
t cente
r
entitle
s ea
ch no
de
with a uniq
u
e
ID
use
d
to solely
represent ea
ch u
s
er.
Then, the
ma
nagem
ent
ce
nter u
s
e
s
a o
ne-to
-on
e
fun
c
tion
h(x) to
cal
c
ulate
the
node’
s
publi
c
key, e:
h(ID) = e
(1)
For ea
ch u
n
iq
ue ID, there i
s
only one
co
rre
sp
ondi
ng p
ublic
key,e.
After computi
ng the publi
c
key of the
node,
acco
rd
ing to the De
mytko deterministic
method of
big
prime
gen
eration, the ma
nagem
ent ce
nt
er first gen
e
r
ates two
big
prime
s
: p a
n
d
q,
then com
pute
s
their p
r
od
uct acco
rdin
g to the formula.
n=p*
q,
φ
(n)
=
(p-
1
)*
(q
-1)
(2)
Then, it cal
c
u
l
ates the priv
ate key,
s, according to the
following formula:
s
2
=e
- 1
m
od
φ
(n)
(3)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A New P2P Identity Authe
n
tication Met
hod Based o
n
Zero
-Kno
wl
edge
Und
e
r
… (Xiyu Pan
g
)
6189
Figure 1. System Establi
s
hment Pro
c
e
s
s
Finally, the server
sen
d
s t
he ID, the pu
blic
key (e), t
he private
ke
y (s) a
nd the
prod
uct
(n) of the two
big prime
s
to
the node, th
en publi
c
i
z
e
s
the one-to
-o
ne functio
n
h
(
x). Beca
use th
e
two prim
es
(p and q
)
an
d
()
n
are p
r
ohi
bited from di
scl
o
sin
g
, it is re
comm
end
ed
to destroy
them, then th
e mana
gem
e
n
t cente
r
d
o
e
s
n’t involv
e th
e next proce
s
s. Du
ring t
he
next pro
c
e
s
s
o
f
authenti
c
atio
n, the manag
ement cent
er is not req
u
ired to be invo
lved. The sp
ecific p
r
e
p
ro
cess
before a
u
the
n
tication i
s
as sho
w
n in the
Figure 1.
2.2. Identit
y
Authen
ticati
on Proces
s
Assu
ming
tha
t
P and
V is p
e
rformi
ng i
d
e
n
tity authenti
c
ation
with
V
as th
e verifie
r
and
P
as the reque
ster, the pro
c
e
ss i
s
as
sho
w
n in the Figure 2
：
(1) P send
s V its ID, the public key (e) a
nd the pro
d
u
c
t (n).
(2) V
che
c
ks the ID and
the publi
c
ke
y (e)
sent b
y
P acco
rdin
g to the one
-to-o
n
e
function. If th
ey are
on
e-to
-one
corre
s
p
onde
nce,
the
n
save the
ID, the
publi
c
key
(e
) a
nd
the
prod
uct (n). If the identity conform
s
,
then
continu
e
, otherwi
se g
o
to 9.
(3) P rando
ml
y cho
o
ses
an
intege
r (r),1
<
r<n
,then
cal
c
ulates X a
c
co
rding
to form
ula(4
)
,
and send
s th
e result to V.
X=
r
e
m
od n
(4)
(4)V
saves X
and ra
ndo
mly choo
se
s an i
n
teger,
n
b
b
,
,
2
,
1
,
n
and se
cretly se
nd
s
it to P.
(5)P
cal
c
ul
ates y a
c
cordi
ng to fo
rmul
a (5
) a
nd V’
s rand
om n
u
m
ber
se
que
n
c
e, then
sen
d
s y
t
o
V
.
2
m
od
s
yr
b
n
(5)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 10, Octobe
r 2013 : 618
7 –
6192
6190
Figure 2. Identity Au
thentication Process
(6) V che
c
ks whethe
r X
and y mean
s the fo
rmul
a (6). If it mean
s, then continue,
otherwise go
to 9.
n
b
y
X
e
mod
1
(6)
(7) Subtract
1 fro
m
the
numb
e
r of
iteration
s
a
nd if it b
e
comes zero,
then the
authenti
c
atio
n is su
cce
ssf
ul, otherwi
se
contin
ue.
(8)
Rep
eat st
ep 1 to 9.
(9)The auth
e
n
tication fail
s and the a
c
ce
ss i
s
deni
ed.
3. Simulation
To verify th
e i
m
prove
d
ze
ro
-kn
o
wl
edg
e i
dentit
y authe
ntication
alg
o
r
ithm p
r
o
p
o
s
e
d
in thi
s
pape
r, seve
ra
l simulation e
n
vironm
ents
are e
s
tabli
s
h
ed.
In the
simul
a
tion, the
nod
es
not
only i
n
clu
de th
e
common
no
de
s b
u
t al
so
a
ce
rtain
numbe
r of m
a
licio
us n
ode
s whi
c
h
play evil role
s
in identity authe
ntication. At the begi
nnin
g
of
the test, th
ere a
r
e
100
no
des in th
e n
e
twork incl
udin
g
ab
out 1
0
%
malicio
us no
d
e
s. In th
e te
st
ing
simulatio
n
e
n
v
ironme
n
t, the di
stributio
n
of sh
are
d
d
a
ta conforms with Zipf la
w a
nd the
Zipf in
dex
is valued bet
ween [0.62
，
1.25]. In the test, the shared data
ar
e
d
e
fined in th
e
form of files
and
all files fall into 50 cate
g
o
rie
s
with a
total numbe
r of 60. The 50 differe
nt kind
s of files are
labele
d
1, 2… …, 50. Among them, t
he file with l
abel 1 i
s
the
most pop
ula
r
one in th
e P2P
system, the l
abel 2 is le
ss pop
ular
an
d so on.
Th
e
r
efore, in the P2P system
, there are m
o
st
copi
es of the
file with label
1, the label 2
has
le
ss copi
es an
d so on.
Rand
omly pl
ace the
60 files
on 10 no
de
s and ma
ke sure each nod
e at least ha
s o
ne file.
In the
simul
a
tion, ea
ch
no
d
e
ave
r
agely
complete
s
50 t
r
an
sa
ction
s
a
nd e
a
ch tran
saction
make
s a no
d
e
download a
file locally. During the tr
an
sa
cting proce
ss, a no
de firstly perform
s
the
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A New P2P Identity Authe
n
tication Met
hod Based o
n
Zero
-Kno
wl
edge
Und
e
r
… (Xiyu Pan
g
)
6191
identity authe
ntication
of the nod
e to
tra
n
sa
ct an
d verifies that if the nod
e is the
one it wa
nts
to
transact. T
h
e
nod
e rand
o
m
ly sen
d
s re
que
st for th
e
sha
r
ed
file a
n
d
when fini
shi
ng d
o
wnloadi
ng,
the copy of the file is save
d locally. In the fi
rst test, the nod
es p
e
rform identity authenti
c
atio
n by
usin
g the Fei
ge-Fi
at-Sham
ir algo
rithm a
nd in
the se
cond test, the
node
s ado
pt the improve
d
zero-kn
o
wl
ed
ge ide
n
tity authentication
algorith
m
p
r
o
posed in
this pape
r. After
the compa
r
ison
betwe
en the two test
s, the result is as
sh
own in the Fi
gure 3.
Time
The success of the download rate
Figure 3. The
Result of Simulation
In the figure
3, the hori
z
o
n
tal axis is ti
me
and the
vertical o
ne i
s
su
cce
ssful
rate of
downloadi
ng.
As
sh
own in
the fig
u
re
3,
the
se
cu
ri
ty perfo
rman
ce
of
the algo
rithm
p
r
op
osed
in
this pape
r is better than that of Feige-Fiat-S
h
a
mir
algorith
m
. The rea
s
on i
s
that the mod
e
l
prop
osed in t
h
is p
ape
r ad
opts the
auth
enticatio
n scheme b
a
sed
on zero
kn
o
w
led
ge, so if
a
malicio
us
no
de impe
rson
ates the p
r
ov
er, it can
not
kno
w
the p
r
iv
ate key (s)
of the prove
r
a
nd
there i
s
not e
noug
h time f
o
r hi
m to figu
re o
u
t t
he
pri
v
ate key
(s)
by cal
c
ulatin
g the p
r
o
d
u
c
t (n
)
and the publi
c
key (e
), so it cannot wo
rk out y in
the
fifth
step an
d finally it ca
nnot satisfy the
con
d
ition in the sixth step
and fails to p
a
ss the verification.
In the
simul
a
tion, some
mali
ciou
s
node
s i
n
ven
t
a meth
od
agai
nst th
e ab
ove
circum
stan
ce
s. They
sen
d
the verifier th
e ID of
the
prover a
nd thei
r own
publi
c
keys. In this
way,
they can u
s
e
their private
keys to cal
c
ula
t
e y in
the fifth step an
d pa
ss the ve
rifica
tion in the sixth
step. Howev
e
r,
the proposed
algo
rithm will verify the informati
on
sent by t
he prover to the
verifier a
c
cording to the on
e-to-one fun
c
tion and tho
s
e who
do not
corre
s
p
ond
will be denie
d
, so
the mali
ciou
s nod
es will fa
il. To cheat
succe
ssf
ully, t
he mali
cio
u
s
node
s m
u
st fi
nd the
X an
d
y
whi
c
h mea
n
the co
ndition
s in the sixth
step. Acco
rdi
ng to the algorithm, if it w
ants to find the
prop
er X a
n
d
y, it must kn
ow the
ran
d
o
m
numb
e
r
(b
) sent by the
verifier
(V) t
o
the p
r
ov
er
(P)
becau
se the
final formula
requi
re
s the random
n
u
mb
er (b
). Suppo
sing the m
a
liciou
s
no
de h
a
s
intercepted th
e ran
dom n
u
m
ber
(b
), it need
s to wo
rk out y accord
ing to formul
a (6
), whi
c
h
can
be reg
a
rd
ed
as a RSA cra
cki
ng proble
m
, of which Xb
can be see
n
as the ciph
ertext obtaine
d by
RSA en
crypti
ng y via u
s
in
g the p
ubli
c
key (n, e
)
, so
that the p
r
o
b
lem tu
rn
s in
to the
solutio
n
of
the plaintext corre
s
p
ondin
g
to the cip
h
e
rtext
unde
r t
he RSA en
crypting syste
m
without
kn
owin
g
the private ke
y (s), it is con
t
radicto
r
y to the su
ppo
sitio
n
.
Assu
ming
the
mali
ciou
s
no
de
can
p
r
edi
ct the r
and
om
numbe
r
(b
), it
can
firstly ra
ndomly
cho
o
se a n
u
m
ber y, com
putes X
acco
rding to f
o
rm
ula(6
)
, then
send X to the
verifier a
nd t
hus it
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 10, Octobe
r 2013 : 618
7 –
6192
6192
can p
a
ss the
authenti
c
atio
n, but the
pro
bability of pre
d
icting b i
s
1/
m (
m
)and after
t times
of loops, the
probability wil
l
becom
e1 /
m
t. Therefore, the ne
w
scheme
has hi
gh reliability and
s
a
fety.
4. Conclusio
n
s
The pa
pe
r prese
n
t a ne
w
P2P identity authent
i
c
atio
n ba
sed
on Z
e
ro
-Kno
wled
ge. In the
new sch
e
me,
the
cal
c
ulati
on formula
o
f
y in
the id
entity authen
tication i
s
si
mpler than
the
formula
use
d
in the Feig
e-Fiat-S
hami
r
algorithm.
T
he calcul
atio
n formul
a of X is not o
n
ly
simple
r than t
he verifying formul
a in the Feige
-
Fi
at-Shamir al
gori
t
hm but also
safer th
an th
at in
Feige
-
Fiat-Sh
a
mir al
gorith
m
. Becau
s
e i
n
the ne
w
scheme, even
i
f
when X
and
b are kno
w
n
,
y
can
not be cal
c
ulate
d
.
The
cal
c
ulati
on formula
of
private
key i
n
the
ne
w
scheme i
s
more compli
cate
d so its
safety is bett
e
r. In additio
n
, the final verifyi
ng form
u
l
a of the ne
w scheme i
s
simpler th
an ot
her
algorith
m
.
Ackn
o
w
l
e
dg
ements
This pa
pe
r is sup
port
ed by JiNa
n Technolo
g
y
Developm
ent Planning
Project
(201
221
140,
2012
2114
1),
Shan Do
ng Ji
ao Tong
Univ
ersity Proje
c
t
(Z20
121
5).
Referen
ces
[1]
Che
n
K, H
w
a
n
g
K, Chen G.Heuristic d
i
sco
ver
y
of
rol
e
-b
a
s
ed trust chain
s
in peer-to-
pe
er net
w
o
rks
.
IEEE Transactions on Par
a
l
l
el
and Distri
bute
d
Systems.
20
09; 41(2): 1
600
-160
4.
[2]
W
Y
Lai, CM Che
n
, B Jeng.
Informati
on
e
xchan
ge
mec
h
anis
m
b
a
se
d
on re
putatio
n i
n
mobi
le P
2
P
netw
o
rks
. Proceed
ings
of the
T
h
ird Internat
ion
a
l Co
nfere
n
c
e on Intel
lig
e
n
t Information
Hidi
ng a
n
d
Multimed
ia Si
g
nal Proc
essin
g
.
2007; 2: 200-
204.
[3]
Li
Xi
ong,
Lin
g
Li
u. PeerT
r
ust: Supp
ortin
g
Re
puta
i
on-
Based T
r
ust for Peer-to-Pe
er Electro
n
i
c
Communities.
IEEE Transacti
ons of Know
le
dge a
nd D
a
ta Engi
neer
in
g
. 2004; 16(
7): 843
-857.
[4]
Resnick
P, Z
e
ckhaus
er R, F
r
i
edm
an E,
etal.
Rep
u
tatio
n
S
ystems.
Journ
a
l
:
Co
mmun
icati
ons of A
C
M
.
200
0; 43(1
2
): 45-48
[5]
R Z
hou, K H
w
ang, M Cai.Go
ssipT
rust for
fast r
eputation a
ggre
gatio
n
in
p
eer-to-p
eer net
w
o
rks.
IEEE
T
r
ansactio
n
s o
n
Know
led
ge a
nd Data En
gi
n
eeri
n
g
. 20
08; 3
8
(2): 894-
89
9.
[6]
Jianmi
ng F
u
e
t
al. "PerformT
rust:
T
r
ust mo
del i
n
tegr
ated
past an
d curre
nt performanc
e in P2P fil
e
shari
ng s
y
ste
m
s". on Com
puter S
y
stem
s and Ap
plic
ations, 2
008.
AICCSA 20
08.
IEEE/ACS
Internatio
na
l C
onfere
n
ce
o
n
. 200
8; 1: 36-42.
[7] Z
heng
Y.
A Conc
eptu
a
l Architecture of
a T
r
usted Mobile Env
i
ro
n
m
ent
. Proc.of t
he Second
Internatio
na
l W
o
rksho
p
on S
e
curit
y
, Priv
ac
y
and T
r
us
t in P
e
rvasiv
e an
d U
b
iq
uitous
Com
putin
g. 20
06;
1: 100-1
06.
[8]
Esther Pal
o
mr, Juan ME.
De
alin
g w
i
th Sp
or
adic Stra
ng
ers, or t
he (U
n) S
u
itab
ility of Tru
s
t for Mobil
e
P2P Sec
u
rity,
Tapiador.
1
8
th
Internati
o
n
a
l
W
o
rkshop
on
Data
b
a
se an
d Exp
e
rt
S
y
stem
s
App
licati
ons.
200
7; 1: 45-50.
[9] W
ang
Che
ng.
T
he Stu
d
y
of P2P S
e
cur
i
ty Mod
e
l B
a
s
ed
on
Identity
Authe
n
ticati
o
n
a
n
d
T
r
ust
mec
h
a
n
is
m.
N
an Jin
g
Post a
nd T
e
lecommu
nicati
ons. 20
07
.
[10]
Z
hao Z
,
W
e
i B, Dong
X.
Detecting wor
m
hole attacks
in wireless sensor netwo
rks with statistical
analysis.
Proce
edi
ngs of the Internat
i
o
n
a
l Co
nferenc
e on Inf
o
rmatio
n
Engi
n
eeri
ng. 20
10: 2
51-2
54.
Evaluation Warning : The document was created with Spire.PDF for Python.