TELKOM
NIKA
, Vol.12, No
.3, Septembe
r 2014, pp. 7
03~710
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v12i3.99
703
Re
cei
v
ed Ap
ril 24, 2014; Revi
sed
Jul
y
7, 2014; Accept
ed Jul
y
25, 2
014
Chord-based Resource Ident
i
fier-to-Locator Mapping
and Searching for the Future Internet
Huanlin Liu, Hongy
ue Dai, Shuaiy
ong
Wu, Sheng
Huan
g
Schoo
l of Com
m
unic
a
io
n an
d Information En
gin
eeri
ng, Ch
o
ngi
qng U
n
iv
ers
i
t
y
of Posts an
d
T
e
lecommunic
a
tions
Cho
n
g
w
e
n
R
o
ad 2#, Na
n’
an
District, Chon
g
q
in
g Cit
y
,
4
0
0
0
65, Chi
n
a
e-mail: li
uh
l2@
s
ina.com
A
b
st
r
a
ct
A great
ma
ny
prob
le
ms, suc
h
as sca
la
bil
i
ty,
ma
ppi
ng data
search
ing, hig
h
frequ
ency up
date
o
f
ma
pp
ing
d
a
ta, arise
i
n
th
e f
u
ture
netw
o
rk
resourc
e
ma
p
p
in
g syste
m
fo
r its vast
data
proc
essin
g
n
eed.
F
u
ture N
e
tw
ork Ch
ord (F
N
C
hord),
an
al
gori
t
hm
bas
ed
on
Chor
d a
n
d
ai
ms at so
lvin
g th
e
reso
urces
id
e
n
tity
ma
pp
ing
a
nd s
earch
ing
pr
obl
em,
is p
u
t for
w
ard by tak
i
ng
adv
anta
ge of the
q
ual
it
ies
of
scal
abi
lity, ra
pi
d
search
ing
spe
ed, hi
gh s
earc
h
in
g effici
ency
and
flexi
b
le
n
a
min
g
of ch
or
d in
ord
e
r to s
o
lve th
is pr
obl
em.
W
hat
’
s
more,
an extra
inter
e
st nod
e in
dex t
abl
e for F
N
C
h
ord is
des
ign
e
d
to rec
o
rd th
e
hotsp
ot reso
u
r
ce
ma
pp
ing l
o
cati
on in the p
a
p
e
r. So, the resource se
ar
chi
n
g strategy, w
h
ich is na
me
d a
s
Interest Index
T
able F
u
ture N
e
tw
ork Chord (
IIT
-F
N Chord) is propos
ed
to
search the r
e
s
ource i
n
the pa
per. T
he entro
py
w
e
ight metho
d
is used to c
a
l
c
ulate th
e no
d
e
inter
e
st
leve
l
accordi
ng th
e
interest n
ode
s
’
r
e
sourc
e
ite
m
onli
ne ti
me an
d visite
d times
and to r
enew
the inter
e
st i
ndex ta
ble. M
o
reov
er, prob
a
b
ility re
pl
ace
m
ent
meth
od is
prop
osed to re
plac
e the outd
a
ted
item o
n
in
ter
e
st index tab
l
e
w
i
th new
item. Simulati
on res
u
lt
s
show
that the alg
o
rith
m can
decre
as
e the
avera
ge se
arc
h
in
g late
ncy,
averag
e searc
h
i
ng ho
ps an
d thu
s
incre
a
ses the s
earch
ing effici
e
n
cy for the resource searching.
Ke
y
w
ords
: future inter
net, ch
ord, in
terest in
dex tabl
e, entr
opy
w
e
ight, av
erag
e searc
h
in
g hops
1. Introduc
tion
With the
rapi
d develo
p
me
nt of Internet
and th
e qui
ck
eme
r
ge
nce of mobil
e
Internet
,
clou
ding
co
m
puting [2], Int
e
rnet
of thing
s
[3], the
pro
b
lem of
Future Intern
et co
nstantly o
c
cu
rs
[4], such as
the problems of Internet
mobility,
routi
ng
scalability
,
mu
lti-homi
n
g and
securit
y
.
Many solutio
n
s are put forward
to sa
tisfy the need of future Internet net
work, for exa
m
ple,
virtualizatio
n
netwo
rk to
pol
ogy [5], rede
sign
cont
e
n
t centri
c n
e
two
r
k a
r
chite
c
ture [6], and sp
lit
resou
r
ce id
en
tifier and
lo
ca
tor [7], etc.
What ide
n
tifier
and l
o
cator
split mean
s i
s
to rep
r
e
s
e
n
t the
doubl
e sem
a
ntic of IP with identifier a
nd locato
r re
spe
c
tively. At present, the identifier an
d
locato
r split a
r
chite
c
tu
re, for whi
c
h the realizati
on of reso
urce ide
n
tifier mappin
g
servi
c
e sy
ste
m
[8] is of great importan
c
e, is wid
e
ly studi
ed by
worl
dwide re
sea
r
che
r
s. The
core issue to reali
z
e
the re
so
urce i
dentifiers m
a
pping
syste
m
is it
s resource map
p
in
g
an
d searchi
ng
a
l
gorithm,
whi
c
h
sho
u
ld be a
b
l
e
to satisfy scalability, mass data map
p
i
ng, update a
n
d
inquiry, etc.
In the P2P system, the DHT-ba
s
ed re
sou
r
ce se
arching metho
d
inclu
d
e
s
Cho
r
d [10],
CAN [11], P
a
stry [12] a
n
d
Tape
stry [
13] in the P
2
P stru
ct
u
r
e
d
network. T
he Chord ba
sed
method has t
he adv
antage of
self-organization, scal
ability,
robust
ness
and rapi
d inquiry for
its
distrib
u
ted h
a
sh tabl
e ba
sed id
entifier mappin
g
sy
stem. So, the techn
o
logy
of Chord m
a
y
become o
n
e
of the pro
m
ising
ca
ndi
date technol
ogi
e
s
for the
resource
se
arching
of future
Internet n
e
twork. B
u
t in th
e future Int
e
rnet netwo
rk, t
he issu
es of
scalability, qu
ick i
nqui
ry, hi
gh
freque
ncy ma
pping a
nd da
ta update
will
rise in th
e re
sou
r
ce map
p
i
ng se
rvice sy
stem for the
r
e
will be m
a
ss data an
d ap
plicatio
n in it. It is found t
hat one
of three o
r
so the
indexe
s
in t
he
Cho
r
d routin
g table a
r
e re
petitive, which red
u
ce
s th
e inqui
ry efficiency an
d in
crea
se
s the h
ops
of inqui
ry [14
]. Moreove
r
,
a sm
all n
u
m
ber
of no
de
s with h
o
tsp
o
t inform
ation
are vi
sited m
o
st
often than the
vast majority
of informatio
n acce
ss
nod
es, whi
c
h l
e
a
d
s to lon
g
-tail
effect of inqu
iry
data [15] and
incre
a
ses th
e inquiry average ho
ps
an
d
delay. Takin
g
full advanta
ge of Cho
r
d, the
Future
Netwo
r
k
Cho
r
d (F
N Chord), an i
m
prove
d
al
g
o
rithm of re
source ide
n
tifier map
p
ing a
nd
sea
r
ching i
n
identifier a
nd
locato
r split future n
e
two
r
k, is put fo
rward to d
e
crea
se the ave
r
a
g
e
inquiry d
e
lay
and the
ave
r
age i
nqui
ry
hop
s in
the
pape
r. The i
m
prove
d
alg
o
rithm b
a
sed
on
intere
st ind
e
x table in
the f
u
ture
network Ch
ord
is pro
posed to
solv
e the lo
ng
-tail
effect p
r
obl
e
m
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 3, September 20
14: 70
3 – 710
704
for ma
ss dat
a inqui
ry. By con
s
id
erin
g b
o
th interest
n
ode
s onli
ne ti
me an
d a
c
ce
ssed time
s,
we
take
advantage of both entropy
weight algorit
hm
and probabilit
y replacem
ent algorithm
to
cal
c
ulate the
node’
s intere
st level and to decid
e whi
c
h Cho
r
d entry
to be repla
c
e
d
in the intere
st
node in
dex table re
spe
c
tively.
2. IT-FN Ch
o
r
d Identifier
Mapping Me
thod
To decre
ase the averag
e sea
r
ching h
o
p
s an
d
avera
ge se
archin
g
delays in the future
Internet net
work, an i
m
proved IIT-F
N Chord re
sou
r
ce ide
n
tifier map
p
ing
and
sea
r
ch
ing
algorith
m
whi
c
h is b
a
sed o
n
intere
st ind
e
x table is propo
sed in the
sectio
n.
An intere
st node ind
e
x table for IIT-FN
Cho
r
d is d
e
si
gned in tabl
e 1 firstly.
Table 1. Stru
cture of inte
re
st node in
dex
table for one
resou
r
ce item
NID
Node IP and p
o
rt
NID precusor
no
de
NID successor node
VT
VC
Each nod
e
j
o
ins
th
e
IIT-FN Cho
r
d sh
ould kee
p
inf
o
rmatio
n of i
t
s precusor
node, its
su
ce
ssor n
o
d
e
and its fing
er tabl
e. And
what’
s
more, an interest no
de index tabl
e for one ite
m
i
s
formed. In th
e interest n
o
de ind
e
x tabl
e, there
mu
st be info
rmati
on ab
out the
serve
r
s visite
d by
the current
se
rver i
n
thi
s
ta
ble. It shoul
d
inclu
de ite
m
s
like th
e IP
ad
dre
s
s a
nd
po
rt numb
e
r of t
h
e
NID (Node’
s hash
Identif
ication), the
IP and p
o
rt
num
ber
of t
he pre
decesso
r
of
t
h
is NID and
th
e
IP and port numbe
r of the succe
s
so
r of this
NID in IIT-FN Ch
ord, experi
e
n
c
ed nu
mbe
r
of
perio
ds d
u
rin
g
the item idle times (VT
)
and item visited times (V
C).
Four
con
d
itio
ns mu
st be g
uara
n
teed in
t
he de
sign of i
n
tere
st index table:
(1)
Rea
s
o
nab
le table scale;
(2)
Hotter no
de is, more likely nod
e sh
ould be in the
table 1;
(3) Ea
ch ne
w item should
be able to joi
n
in the table;
(4) Ea
ch outd
a
ted inform
ation sh
ould b
e
elim
inated fro
m
the table as so
on a
s
po
ssi
ble.
Firstly, sin
c
e
the SHA-1 i
s
use
d
as the
DHT
ha
sh fu
nction in IIT-FN Chord, there
sho
u
ld
be 160 items in the finger table. According to t
he 20/80 prin
cipl
e
,
studies sho
w
that there are
only 20 ite
m
s whi
c
h
are fre
quently inq
u
ired in
ever
y 1
00 item
s. So, we
assu
me t
hat the n
u
mb
e
r
of items in the intere
st nod
e index table
is 160*
20%
=32.
Secon
d
ly, ca
lculate inte
re
sting level of
each n
ode.
The impo
rtan
ce of one re
sou
r
ce
item, namely
its inte
re
sting
level, is obtai
ned
by
ap
plying the
ent
rop
y
weig
ht alg
o
r
ithm [16]
whi
c
h
take
s both idl
e
time and accesse
d
times into con
s
id
eration. The
i
-th item’s interesting level
of
node i
s
cal
c
ul
ated in equ
ation (1
).
12
ii
i
R
VT
VC
(1)
Whe
r
e, VT
i
and
VC
i
re
p
r
esent
the n
ode’
s
item
i
idle time
s
and visite
d times,
respec
tively. Parameters
β
1
and
β
2
a
r
e
the wei
ght
co
efficient of V
T
and
VC, re
spe
c
tively. VT
and VC
are i
n
itialize
d
wh
e
n
a ne
w item
is add
ed to th
e node
as VT
0 and 1, resp
ectively. Here,
VT0 is a big
value acco
rdi
ng to the network’
s
re
so
ur
ce a
c
c
e
s
s
lev
e
l,
suc
h
as
w
e
set
V
T
0=
32
,
VC=
1
,
i
=1,
2,
…, VT0
at th
e alg
o
rithm’
s
begin
n
ing to
evaluate
nod
e’s i
n
tere
st fo
r resource
ite
m
i
. If the node
is n
o
t visite
d by othe
r n
ode in
on
e ti
me pe
riod
fo
r the item
i
,
the item’s VT0
decrea
s
e
s
a
b
out 1. If the item’s
VT redu
ce
s to 0, the item
i
is ki
cked out of the
intere
st nod
e
index table. If the nod
e i
s
visited by ot
h
e
r
no
de in
one
time peri
od f
o
r the item
i
, the item’s V
C
increa
se
s a
b
out 1. If we
calcul
ate all
ite
m
s’ inte
re
st l
e
vel of n
ode,
we
so
rt all ite
m
s
acco
rdin
g
to
each item’s in
terest level from high to lo
w ord
e
r.
1
12
1
12
,1
w
ww
(2)
Whe
r
e,
w
1
is t
he entro
py weight of VT, and
w
2
is the e
n
tropy wei
ght
of VC.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Chord-Based Resource Identifier to Locator
Mapping
and Searching for .... (Huanlin Liu)
705
2
11
1
(1
)
/
(
1
)
j
j
we
e
(3)
The
e
1
and
e
2
are the entro
py of VT and VC re
spe
c
tively.
0
11
1
2
2
2
11
11
ln
;
l
n
0l
n
VT
VC
ii
i
i
ii
ep
p
e
p
p
ln
V
T
V
C
(4)
The
p
i
1
an
d
p
i
2
are
i
-th
item’s p
r
op
ortion of VT
paramete
r
and VC pa
rameter,
r
e
spec
tively.
0
12
11
/;
/
VT
VC
ii
i
i
i
i
ii
p
VT
VT
p
V
C
V
C
(5)
Thirdly, repla
c
e
one
item
in the
no
d
e
’s
i
n
tere
st i
ndex tabl
e
usin
g the
probability
repla
c
e
m
ent algorith
m
[17] which take
s all t
he items intere
stin
g level as proba
bility judge
criterion. The
higher the intere
st level the item has, the less likely it will be repl
aced.
Set
P
is the
alternative probability of each item
in set R by a new item. Here,
set R is
the each item’s intere
st level of node, an
d
P
= {
P
1
,
P
2
, … ,
P
VT
0
},
R
={
R
1
,
R
2
, … ,
R
VT
0
}. When
a
new item is
adde
d to the interestin
g node
s index
table, the item’s altern
ative prob
ability
P
i
is
initialized
as
0.5. The alternative probability of
i
-th item in the int
e
re
st index table is
de
sig
ned
as
follow.
2
0
2
1
1
0.
5
*
1
i
i
VT
i
i
R
P
R
(6)
Acco
rdi
ng
to the
equ
ation (6), we kno
w
that
the
large
r
R
i
val
ue of
a item, the smaller
P
i
.
Lastly,
we de
cide wh
ether the
nod
e
is o
u
tdated.
An i
n
tere
st nod
e
is taken a
s
a
n
outdate
d
n
ode
and will b
e
el
iminated from
the interest
node in
dex t
able as lo
ng a
s
it is not visited for a certa
i
n
numbe
r of pe
riod
s. In othe
r wo
rd
s, if on
e item’s VT
o
f
the node is
0, the item is
kicke
d
out of
the
intere
st node
index table.
3. IIT-FN Ch
ord resou
r
ce
searching
Before
we
de
scribe
the
propo
sed
chord-ba
se
d re
so
urce sea
r
chin
g
st
rategy, we
defin
e
some
symbol
as follows.
NID re
prese
n
t
s the ID of a node. VC is visi
ted times
of the target node. Usually
, VC=1
whe
n
a new it
em is add
ed to the intere
sting node
in
de
x table. And
VC is add
ed 1 as targ
et node
is visite
d o
n
ce. VT in
dicat
e
s th
e
node’
s rem
a
inin
g av
ailable
idle
times.
Usually,
we
a
s
sume
that
the initial value of VT is
32, namely VT0=3
2
at the beginni
ng. If a
target no
de is not visited
durin
g on
e p
e
riod, VT i
s
redu
ced to
1. If VT=0, the
n
this n
ode i
s
elimin
ated f
r
om the i
n
terest
index table i
mmediately. IIT is intere
st node in
dex table.
Whe
n
no
de
n
in the IIT-F
N Cho
r
d
se
nd
s a re
so
urce i
nquiry K, the
pro
c
e
s
s of re
sou
r
ce
sea
r
ching i
s
desi
gne
d as f
o
llow.
Step 1
: Che
c
k IIT to see i
f
there is a
n
y invalid or o
u
t
dated nod
e i
n
it. If
there i
s
any,
delete it from
the IIT. Rene
w su
cce
s
sor
node n
u
mbe
r
and precursor no
de num
ber of ea
ch n
ode
on IIT. Updat
e finger table
of each n
ode
on IIT-FN
Ch
ord.
Step 2
: Inq
u
i
r
y the resource K o
n
no
d
e
n
. If re
sou
r
ce K i
s
on
n
ode
n
, the res
o
urce
sea
r
ching
su
ccess an
d the sea
r
ching p
r
ocess go
es t
o
Step 3
. Else if resou
r
ce K is not on node
n
, the resource sea
r
ching g
oes to
Step 4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 3, September 20
14: 70
3 – 710
706
Step 3
: Establish a ne
w item to IIT of n
ode
n
with VT=32, VC=1. If IIT of node
n
is full,
callin
g the p
r
obability re
pl
acem
ent alg
o
r
ithm to re
pla
c
e
some item
of IIT with this ne
w item. Else
if IIT is not full, the new item is adde
d di
rectly to the IIT.
Step 4
: In
qui
ry the IIT an
d
judge
wheth
e
r
the
re
so
urce
K need
s o
n
ly
one hop. If
n
<K
<
su
cc
es
so
r (
n
), then resource K is on the one ho
p nod
e
of succe
s
sor (
n
); Els
e
goes
to
Step 5
.
Step 5
: Re
ad the
IIT i
t
em by item
to find
out
if there i
s
any item
i
n
which
p
r
ec
us
or
(
n
’)<K<
n
’. If s
o
, we s
e
t the item with VT=
32,
VC=
V
C+1. Els
e
, goes
to
Step 6
.
Step 6
: Ch
eck IIT-FN
Cho
r
d’s fing
er ta
ble of node
n
to find the node
x
whi
c
h
is both
clo
s
e
s
t to and smalle
r tha
n
K. If
the node
x
exis
t, then s
e
t
n
=
x
a
nd re
so
urce searchin
g su
cce
ss
goe
s to
Step 2
. Else, the reso
urce K inq
u
iry fails and
the pro
c
e
ss o
f
resou
r
ce K sea
r
ching e
n
d
s.
Figure 1. The
flow cha
r
t of resou
r
ce K searchin
g
4. Results a
nd Analy
s
is
To eval
uate
the
perfo
rmance
of IIT-F
N
Ch
o
r
d
re
so
urce
m
appin
g
a
nd
sea
r
ching
strategy, by compa
r
ing
with the FN Cho
r
d sy
st
em wit
hout the IIT, the 100 time
s
experim
ents
are
carrie
d out in
five different node
s scal
abi
lity env
ironm
ents, whi
c
h
a
r
e 12
8, 256, 512, 102
4, 20
48,
respec
tively in P2PSim s
o
ftware.
In Figure
2, we compares
probabilit
y density function
(P
DF) of query hops under
different F
N
Cho
r
d n
ode
numbe
r N=1
28, 256, 51
2
,
1024, 204
8
,
resp
ectively
. As sho
w
n i
n
Figure 2, und
er differe
nt node num
be
r in FN Chor
d system, the IIT-FN
Cho
r
d
se
arching
strate
gy
need
s ave
r
a
ge fewer q
u
e
r
y hop
s com
parin
g with
F
N
Chord sea
r
chi
ng
strate
gy. The re
ason is
that we d
e
sig
n
an IIT for t
he IIT-F
N Ch
ord
syst
em, t
he re
so
urce
query
req
uest gets the target
node fa
ster
according to
node’
s inte
rest ind
e
x tab
l
e. For exa
m
ple, in Figu
re 2(c), the
n
ode
numbe
r of F
N
Cho
r
d
system is
N=51
2, the num
b
e
r of 1,
2, 3
,
4, 5, 6, 7, 8 hop
s
su
ccess
sea
r
ching
is
9, 17, 24,
22,
12, 9,5, 2 tim
e
s, resp
ec
tively in IIT-FN
Chord
s
y
s
t
em. But, the number
of 1, 2, 3,
4,
5, 6, 7, 8
hop
s
su
ccess
se
arching
is 2,
8, 18, 2
7
, 26,
13, 4,
2 time
s, re
sp
ectivel
y
in
FN Ch
ord. It is quite cle
a
r t
hat IIT-FN Chord i
s
more likely to end up with su
ccessful se
arch
ing
of less
hop
s.
In Figure 3
(e), N=20
48, the num
ber
of
1, 2, 3, 4, 5, 6, 7, 8, 9, 10 hop
s succe
s
s
sea
r
ching
is
4, 5, 12, 2
4
,
22, 16,
9,
4,
3,1, re
spe
c
tively in IIT-F
N
Cho
r
d. But, t
he n
u
mbe
r
of
1, 2,
3, 4, 5, 6, 7,
8, 9,10 hops
su
cces
s searchin
g is 1, 4, 8, 18, 25, 24,
13, 5, 2, 0
times, re
sp
ecti
vely
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Chord-Based Resource Identifier to Locator
Mapping
and Searching for .... (Huanlin Liu)
707
in FN Chord
.
It is quite
clea
r that IIT-FN
Cho
r
d i
s
more likely
to end up
with su
cce
s
sful
sea
r
ching of l
e
ss hop
s.
01
234
567
0.
00
0.
05
0.
10
0.
15
0.
20
0.
25
0.
30
0.
35
PD
F
Q
uer
y
H
ops
IT-F
N
C
hord
FN Ch
o
r
d
(
a
)
n
o
d
e
nu
mb
er
N=
1
28
01234
5
6789
0.
0
0
0.
0
5
0.
1
0
0.
1
5
0.
2
0
0.
2
5
0.
3
0
0.
3
5
PD
F
Q
u
er
y
H
o
ps
I
T
-
F
N
Ch
o
r
d
F
N
C
hor
d
(b) n
ode n
u
m
ber N=2
5
6
0
1
2
3456
78
9
1
0
0.
00
0.
05
0.
10
0.
15
0.
20
0.
25
0.
30
0.
35
PDF
Q
u
er
y H
ops
IT
-
F
N
C
h
o
r
d
F
N
C
hor
d
(c) nod
e num
ber N=5
1
2
0123
4
5
67
8
9
1
0
0.
0
0
0.
0
5
0.
1
0
0.
1
5
0.
2
0
0.
2
5
0.
3
0
0.
3
5
PD
F
H
o
p
C
o
un
t
(
N
=
10
24)
IT
-F
N
C
h
o
r
d
FN
C
h
o
r
d
(d) n
ode n
u
m
ber N=1
024
01
2
3
45
6
7
89
1
0
1
1
0.
00
0.
05
0.
10
0.
15
0.
20
0.
25
0.
30
0.
35
PDF
Que
r
y
Ho
ps
I
I
T
-
F
N
Ch
ord
F
N
Ch
ord
(e) n
ode n
u
m
ber N=1
024
Figure 2. Pro
bability den
sity function (P
DF ) of
que
ry hop
s with different nu
mbe
r
of nodes
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 3, September 20
14: 70
3 – 710
708
It is exactly th
e same ten
d
e
n
cy for N=12
8, 256, 1024.
So we ca
n co
nclu
de that:
(1)
Searchin
g proce
s
s in IIT-FN Chord ha
s a
hig
her
probability to g
e
t the need
e
d
informatio
n
than in FN
Chord, na
mely
it is more likely
to get the requi
red info
rmation fo
r IIT-F
N Ch
ord
with less
hops
.
(2)
As the nu
mb
er of ide
n
tifier map
p
ing
se
rv
ers in
cre
a
ses, the in
quiry hops
of IIT-FN
Ch
ord
tend more likely to be less than 0.5
logN
0.5, he
re
N is the nod
e numbe
r of FN Cho
r
d
sy
st
em.
1
2
8
2
56
51
2
1
024
204
8
2
3
4
5
6
A
verage suce
ss sea
r
c
hing
hop
N
u
m
b
e
r
of
node
s
F
N
C
h
o
r
d
IIT
-
F
N
C
h
o
rd
Figure 3. Average
su
ccess searchi
ng ho
ps versu
s
nu
mber of no
de
s
In Figure 3,
we compa
r
e
the reso
urce averag
e succe
ss
sea
r
ching ho
ps ve
rsu
s
the
numbe
r
of no
des. By a
naly
z
ing
the F
N
Cho
r
d
re
so
ur
ce st
r
a
t
egy
p
r
oce
s
s,
w
e
kn
ow a sea
r
c
h
can
locate th
e target nod
e wit
h
in the limite
d
hop
s, nam
ely algorith
m
’
s
complexity about 0.5
lo
gN
in
FN Cho
r
d. O
n
the othe
r h
and, the
co
m
p
lexity of IIT-FN Cho
r
d is
about 0.5
logN
0.5. A
s
th
e
numbe
r of
n
ode in
crea
se
s, the ave
r
a
ge succe
s
s
sea
r
ching
ho
ps of IIT-FN Cho
r
d i
n
cre
a
se
s
slo
w
er than
F
N
Cho
r
d. Thi
s
i
s
si
mply be
cau
s
e
of the i
n
crea
se
of su
cc
ess
pro
bab
ility of one h
o
p
sea
r
ching
aft
e
r
addin
g
th
e
nod
e inte
re
st node
ind
e
x
table. With
th
e in
crease
of the
numb
e
r
o
f
node
s, the su
ccess proba
b
ility of one ho
p sea
r
ching
i
s
even high
er, thus the differen
ce valu
e of
the averag
e succe
ss
sea
r
ching ho
ps b
e
twee
n FN Cho
r
d and IIT-F
N Chord will be
even large
r
.
As sho
w
n in
Figure 4, we
can find the a
v
er
age
sea
r
ching laten
c
y of FN Cho
r
d i
n
crea
se
s
with the incre
a
se of the number of ident
ifier m
appin
g
nodes in the
Internet. It is the same with
the IIT
-FN Chord. What is more
, the average
sea
r
ching laten
c
y of IIT
-FN Ch
ord is mu
ch l
e
ss
than FN
Ch
ord and g
r
o
w
s slo
w
er
as th
e
numbe
r of
th
e mappi
ng n
o
des i
n
crea
se
s.
This i
s
sim
p
ly
becau
se the con
s
tantly up
dated In
tere
st
node index table increa
se
s
the resource search spe
ed.
In the IIT
-FN
Cho
r
d, the n
ode
s on IIT
keep hig
h
re
source inte
re
st level by usin
g entro
py wei
ght
algorith
m
to
cal
c
ulate ite
m
’
s
onli
ne ti
me and
vi
sited times co
mpre
hen
sivel
y
.
And the outdated
node
s a
r
e
re
placed by
ne
w no
de
s
with proba
b
ility repla
c
e
m
ent
method.
T
h
e
two me
asures
guarantee th
e node
s on IIT
high interest level and re
duce its re
so
urce se
arch
e
d
time.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Chord-Based Resource Identifier to Locator
Mapping
and Searching for .... (Huanlin Liu)
709
1
2
8
256
51
2
1
024
204
8
25
0
30
0
35
0
40
0
45
0
50
0
55
0
A
v
e
r
age s
earch
ing late
ncy
(ms)
Nu
m
b
e
r
o
f
no
de
s
F
N
C
h
o
r
d
IIT
-F
N
C
h
o
r
d
Figure 4. Average
su
ccess searchi
ng lat
ency versu
s
the numb
e
r of
node
s
5. Conclusio
n
In this
pape
r, an IIT-F
N
Cho
r
d i
s
pro
posed
to i
m
prove th
e
Chord
sy
stem
identifier-
locato
r ma
pp
ing efficie
n
cy
and
spe
ed
up the
re
sou
r
ce
searchi
n
g for the fut
u
re Inte
rnet.
By
desi
gning
a
n
extra inte
re
st
index tabl
e fo
r the
no
de, th
e hot
spot
re
source
can
be
locate
d q
u
ickl
y.
To imp
r
ove t
he efficie
n
cy
of intere
st in
dex table,
th
e entropy
we
ight algo
rithm
is int
r
od
uce
d
to
cal
c
ulate the
resou
r
ce item
’s interest lev
e
l. T
he outda
ted re
sou
r
ce is re
pla
c
ed b
y
new item wi
th
the proba
bility repla
c
e
m
e
n
t method.
As the
appli
c
ation
s
and
node
s i
n
crea
se i
n
the
future
Internet,
the
IIT-FN Cho
r
d can ke
ep
Cho
r
d of
P2P
advanta
ges and
i
m
prove
s
reso
urce
sea
r
ching
efficien
cy an
d
spe
ed. IIT-F
N
Cho
r
d i
s
expecte
d to
solve th
e Internet i
nhe
re
nt
scalability and performance probl
em for the future Internet.
Ackn
o
w
l
e
dg
ements
This resea
r
ch wa
s fund
e
d
by the national natu
r
e
scien
c
e fou
n
d
a
tion of Chi
n
a (NSF
C
6127
5077,
6
1371
096, 5
1
1755
35),
by the 97
3 n
a
tion
al
program o
n
key ba
sic rese
arch
proj
e
c
t of
Chin
a (2
012
CB315
803
),
and by the b
a
si
c an
d fron
tier
re
se
arch prog
ram of
Chong
qing (CSTC
2013j
cyjA400
52).
Referen
ces
[1]
Inoue
A, N
a
g
a
hata
R, Ishi
i Y,
et a
l
.
Mo
bil
e
I
n
ternet-acc
ess
be
havi
o
r
ana
l
ysis
. Proceedings
of IEEE
confere
n
ce o
n
SNPD. K
y
oto. 201
2: 766-
770.
[2]
Xu
X. F
r
om
clou
d c
o
mp
uting
to cl
ou
d man
u
facturi
ng.
R
obotics
an
d co
mput
er-inte
g
rate
d
ma
nufactur
i
ng
.
2012; 2
8
(1): 7
5
-86.
[3]
Xi
a F
,
Yang L
T
, W
ang L, et
al. Internet of T
h
ings.
Interna
t
iona
l Jour
nal
of Co
mmun
icat
ion Syste
m
s
.
201
2; 25(9): 11
01-1
102.
[4]
Menth M, Hartmann M, Klei
n
D.
Global loc
a
tor, local l
o
ca
tor,
and ide
n
tif
i
er split (GLI-split).
Fu
tu
re
Internet
. 201
3; 5(1): 67-9
4
.
[5]
Liu Y. D
e
ve
lop
m
ent of n
e
t
w
or
k conver
ge
nce
and
future
int
e
rnet.
Jo
urna
l of
Cho
n
g
q
in
g Univers
i
ty
of
Posts and T
e
l
e
communic
a
tio
n
s
(Natural Sci
e
nce Editi
on)
. 2
010; 22(
6): 693
-697.
[6]
Led
erer S, Mu
eller
C, Ra
in
er B, et al.
Ada
p
tive strea
m
in
g over c
onte
n
t centric n
e
tw
orks in
mo
bi
l
e
n
e
t
wo
rks u
s
ing
m
u
l
t
ip
le
li
n
k
s
. Procee
din
g
s
of IEEE C
o
nferenc
e o
n
C
o
mmunic
a
tio
n
s
Workshops
(ICC). Budap
e
s
t. 2013: 677-
6
81.
[7]
Martinez-J
uli
a
P, Skarmeta A
F
. Be
y
o
nd th
e
separ
at
ion
of i
dentifi
e
r a
nd l
o
cator: bui
ldi
ng
an i
d
e
n
tit
y
-
base
d
ov
erla
y
net
w
o
rk
arch
itecture for th
e f
u
ture Inter
net.
Co
mp
uter N
e
tw
orks
. 2013;
5
7
(10): 2
2
8
0
-
230
0.
[8]
Rui T
.
T
he stu
d
y
of the archi
t
ecture an
d ke
y me
ch
an
ism i
n
the ide
n
tifier
and loc
a
tor s
p
lit net
w
o
rk.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 3, September 20
14: 70
3 – 710
710
PhD thesis
. Ch
angs
ha: Nati
on
al Univ
ersit
y
of
Defense T
e
chnol
og
y. 20
09.
[9]
Shen
H, Li
Z
,
Li J. A D
H
T
-
Aided
Ch
unk-Dr
i
ven Ov
erl
a
y fo
r Scala
b
l
e
a
n
d
Efficient Pe
er-
t
o-Peer
Live
Streamin
g.
IEEE Transactions
on Parallel
and Distributed S
ystem
s
. 20
13; 24(1
1
): 212
5-2
137.
[10]
Lan M. Structured P2P a
l
go
rithm
w
i
t
h
fuzzy qu
er
y
.
Jo
urn
a
l of Cho
ngq
in
g Univers
i
ty of Posts an
d
T
e
leco
mmunic
a
tions (N
atural
Scienc
e Editio
n)
. 2013; 2
5
(5)
:
680-68
5.
[11]
Ratnas
am
y
S,
F
r
ancis P,
Hand
le
y M, et al. A scalabl
e content-a
ddr
ess
abl
e net
w
o
rk.
ACM
. 2001;
31(4): 16
1-1
7
2
.
[12]
Ro
w
s
tro
n
A, Drusche
l
P.
Pastry:
Scalable, Dece
ntrali
z
e
d Obje
ct Locati
o
n, and R
outin
g for Larg
e
-
Scale
Peer-to
-
Peer Syste
m
s
. Procee
din
g
s
of IF
IP/ACM Conf
erenc
e
on
Distrib
ut
ed S
y
stems
Platforms. Hei
del
berg. 2
001:
329-
350.
[13]
Lin
g
H, String
J, Rhea S
C
, e
t
al. T
apestr
y
:
a
resil
i
e
n
t glo
b
a
l-scal
e
ov
erla
y for serv
ice d
epl
o
y
ment
.
IEEE Journal
on Selected Areas in Communications
. 20
04;
22(1): 41-
53.
[14]
Che
n
G, W
u
G. G-Chord: a
n
impr
ove
d
ro
uting
al
gorithm
for Ch
ord.
J
o
urna
l of S
outh
east U
n
ivers
i
ty
(Natural Sc
ien
c
e Editio
n)
. 20
07;
37(1): 9-1
2
.
[15]
W
u
LL, L
ues
ukpras
ert L,
Lee
L.
Res
e
a
r
ch an
d the
l
ong ta
il: A l
a
rge-scal
e
citati
on a
n
a
l
ysis
.
Procee
din
g
s of
IEEE Confere
n
ce on S
y
stem
Science (HIC
SS). Big Islang
. 2009: 1-1
0
.
[16]
F
u
Y, W
u
X, Y
e
Q, et a
l
. An
appr
oach
for
in
fo
rmation
s
y
stems sec
u
rit
y
ri
sk assessm
ent
on
fuzz
y set
and e
n
trop
y-
w
e
ig
ht. Acta E-l
ectronica Si
nica
.
2010; 3
8
(7): 1
489-
149
4.
[17]
Gao G. T
he stud
y
of Distri
b
u
t
ed C
a
che
me
chan
ism i
n
P2
P net
w
o
rk.
Ph
D thesis
. W
u
h
an: H
uazh
o
n
g
Univers
i
t
y
of Scienc
e an
d T
e
chno
log
y
. 2
011.
Evaluation Warning : The document was created with Spire.PDF for Python.