TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 12, No. 10, Octobe
r 20
14, pp. 7395
~ 740
3
DOI: 10.115
9
1
/telkomni
ka.
v
12i8.535
2
7395
Re
cei
v
ed
De
cem
ber 1
3
, 2013; Re
vi
sed
Jun
e
24, 201
4; Acce
pted July 20, 20
14
Networked
Automatic T
est System Based on
Message-
Oriented Middleware
Xiong Fengg
uang*, Ha
n Xie
North Un
iversit
y
of Ch
ina C
o
mputer an
d Co
ntrol Eng
i
n
eeri
ng, T
a
i
y
ua
n 03
005
1, Chi
n
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: xfgnc
it98@s
i
na.com
A
b
st
r
a
ct
Aim
at the c
u
rrent status
of automatic test system
that foc
u
ses only on
a single or the s
a
me type
of w
eapons'
perfor
m
a
n
ce t
e
sting, a
net
w
o
rked auto
m
atic test system b
a
se
d on
mess
ag
e-ori
e
n
t
e
d
mi
ddl
ew
are is
prop
osed
in th
i
s
pap
er, an
d in
ner arc
h
it
ectur
e
an
d pr
ocess
of pus/sub
in
mess
ag
e-ori
e
n
t
e
d
mi
ddl
ew
are ar
e describ
ed. B
a
sin
g
on the mess
ag
e-or
i
e
n
t
ed mi
ddl
ew
ar
e, the net
w
o
rked auto
m
atic test
system
c
an s
h
ield the c
o
m
p
lexity
and diver
s
ity of the bottom
t
e
st
system
and t
e
st equipment, and
can
reali
z
e
cross-p
l
atform c
o
mmunic
a
tion
of vario
u
s te
st data abo
ut w
eap
ons. Exper
i
m
e
n
ts show
that the
netw
o
rked
aut
omatic test sy
stem
bas
ed
o
n
messa
ge-
ori
ented
mid
d
lew
a
re is
an
int
e
g
r
ated test
platf
o
rm
w
h
ich can prov
ide co
mpreh
e
n
s
ive perfor
m
an
ce test fo
r coordin
a
ted e
nga
g
e
ment
on a v
a
r
i
ety of w
eapon
s.
Ke
y
w
ords
:
n
e
tw
orked aut
o
m
atic test sy
stem, test
da
ta, mess
age-
o
r
iente
d
mid
d
l
e
w
a
re, mess
ag
e
pub
lish
i
ng,
me
ssage rec
e
ivi
n
g, mess
age su
bscripti
on
Co
p
y
rig
h
t
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Ra
pid deve
l
opment a
n
d
application
of key
inform
ation technol
ogy prom
ote
s
a ne
w
revolution i
n
military affairs an
d develo
p
ment of
info
rmation warfa
r
e, and p
r
omo
t
es devel
opm
ent
and revol
u
tio
n
of modern
military theory and ope
rational mod
e
l
, such a
s
a network-ce
n
t
ric
warfa
r
e. The
prerequi
site
of winning
in the
future
wars dep
en
ds on
wheth
e
r co
ope
rativ
e
enga
geme
n
t
betwe
en
wea
pon
platform
s ca
n b
e
impl
e
m
ented. T
herefore, b
u
ildin
g an
integ
r
at
ed
test platform
for co
mpre
hen
sive pe
rforma
nce test of a varie
t
y of weapo
ns'
coo
p
e
r
ati
v
e
enga
geme
n
t is a cru
c
ial task.
Gene
ral
spe
a
kin
g
, the in
tegrated
me
asu
r
em
ent a
nd device which a
r
e
as core of
comp
uter
an
d ca
n autom
atically
pe
rform ce
rtain te
st task u
nde
r
the co
ntrol of
pro
c
ed
ure a
r
e
calle
d autom
atic test syste
m
[1, 2] (ATS). In or
der to i
m
prove the m
odernization l
e
vel of airborne
equipm
ent m
a
intena
nce, the re
se
arch
of automati
c
test
syst
e
m
has be
co
me the fo
cu
s of
developm
ent
in the world
military equipment [3
]. Current ATS has bee
n wid
e
ly applied to a
variety of we
apon
and
eq
uipment te
st
s an
d ha
s
b
e
com
e
n
e
ce
ssary a
s
sura
nce
of weap
on'
reliabl
e ope
ration. Ho
wev
e
r, the cu
rre
nt ATS is
different from
the military, and is la
ck o
f
interop
e
rability betwee
n
d
i
fferent sy
ste
m
s. So, ATS can't a
dap
t to the mo
dern j
o
intne
s
s'
s
deman
d on m
u
lti-we
apo
n systems, multi
-
level mainte
nan
ce.
This
pap
er ai
ms to b
u
ild a
netwo
rked A
T
S based
on
messa
ge-ori
ented mi
ddle
w
are [4,
5] (MOM
). T
he me
ssag
e-based mid
d
le
ware i
s
bet
ween the
botto
m equi
pment
and a
ppli
c
ati
o
n
system
s, and
shiel
d
s
com
p
lexity and di
versity of
the
bottom test
system, an
d
achi
eves
dat
a's
effective tran
smissio
n
an
d
wide
sha
r
in
g on the
int
e
rnet th
roug
h MOM'
s st
rong fun
c
tion
of
sen
d
ing an
d
receivin
g d
a
ta, and enh
ances inte
ra
ction of net
worke
d
ATS and relia
bility o
f
wea
pon
s' coo
r
dinatio
n test.
2. Architectu
r
e of Au
tom
a
tic Tes
t
Sy
s
t
em
In general, ATS is mad
e
of hard
w
a
r
e and
software. Ha
rdware is m
ade
of host
comp
uter
，
ge
neral b
oard
，
test instru
me
nt and mea
s
ured o
b
je
ct. Board a
nd test instru
ment
are
respon
sibl
e to control equ
ipment or for si
gnal a
c
qui
sition, and can be ba
sed
on various
data
bus,
su
ch
as PCI, PXI, GPIB. The me
asu
r
ed
obj
ect
refe
rs to
an
y test equi
p
m
ent, syste
m
s o
r
sub
s
ystem
s
,
prod
uctio
n
lin
es, etc. In th
e proc
ess of
putting up te
st syst
em, user u
s
u
a
lly pa
ys
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 739
5
– 7403
7396
attention to th
e ope
ration
of test boa
rd
an
d inst
ru
me
nt. Software i
s
m
ade of I/O int
e
rface, drive
r
s
of board an
d
instrume
nt, executa
b
le a
pplication.
I/O interface provide
s
the
interface of
the
bottom h
a
rd
ware
drive
r
a
n
d
is respon
si
ble fo
r
comm
unication
s b
e
t
ween
the
ho
st comp
uter
and
the physi
cal instru
ment. Instrum
ent driv
er is
a software a
s
sembly
betwee
n
the upper
software
and
a
spe
c
ifi
c
in
stru
ment,
and i
s
re
spo
n
s
ible
for t
r
an
sferrin
g
in
stru
ction bet
wee
n
appli
c
ation
a
n
d
instru
ment,
and i
s
a
mi
ddle
comm
u
n
icatio
n laye
r bet
wee
n
I/O interfa
c
e
and a
ppli
c
ati
o
n
.
Executable a
pplication pro
m
otes u
s
e
r-friendly intera
ction interface
to execute test ope
ration
by
instru
ment d
r
i
v
er, and p
r
o
m
otes
data a
nalysi
s
, dat
a
pro
c
e
ssi
ng, a
nd data
sto
r
a
ge. I/O interfa
c
e,
board and in
strum
ent driv
er, executa
b
l
e
appli
c
ati
on
con
s
titutes th
e entire software a
r
chitectu
re
of automatic t
e
st syste
m
. The arch
ite
c
ture is sh
own as Figure 1.
Figure 1. Architecture of Automatic Te
st
System
3. Arc
h
itec
ture o
f
Netw
o
r
ke
d
Automatic T
est Sy
stem Based
on M
essag
e-o
r
ie
nte
d
Mid
d
l
e
w
ar
e
Network
e
d automatic tes
t
s
y
s
t
em
[6] (NATS) i
s
a
ki
n
d
of
automati
c
te
st
system
built
on
netwo
rk. T
h
rough m
e
ssa
ge-o
r
ie
nted
middle
w
a
r
e
and n
e
two
r
k,
NATS ba
se
d on me
ssa
ge-
oriente
d
mid
d
leware
can
build a net
worked te
st
system. Each
node in the
netwo
rk i
s
a
host
comp
uter
whi
c
h obtai
ns te
st informatio
n
throug
h the test bus
co
nn
ecting te
st eq
uipment. Each
host
comp
uter can
re
cei
v
e additional
informatio
n
sent by the
other h
o
st
compute
r
thro
ugh
messag
e-orie
nted mi
ddle
w
are, a
nd
al
so
publi
s
h
e
s
inform
ation
to othe
r h
o
st
co
mpute
r
th
at
sub
s
crib
es th
e informatio
n. Its archite
c
tu
re is
sho
w
n a
s
Figu
re 2.
Figure 2. Architecture of Networke
d Automa
tic Te
st System Base
d on Messa
g
e
-
oriente
d
Mid
d
l
ew
ar
e
Ho
st comput
er, conn
ecte
d to the
sp
ecific
test in
strum
ent in
orde
r to
obt
ain te
st
informatio
n of
variou
s test i
n
stru
ment
s, i
s
a
no
de i
n
t
he
NATS. After
obtainin
g
t
e
st info
rmatio
n,
host
co
mpute
r
al
so
ne
ed t
o
an
alyze,
sh
ow,
stor
e a
n
d
se
nd th
ose i
n
formatio
n. O
t
her, e
a
ch h
o
s
t
comp
uter
ca
n sub
s
cri
be test inform
atio
n of other
ho
st com
puter.
That is to sa
y, different host
comp
uter
ca
n reali
z
e inte
ractio
n with real in
form
ation and
can
compl
e
te coo
r
dinatio
n test
of
wea
pon. M
e
ssage
-o
riente
d
middl
ew
are is a
bri
d
g
e
for test i
n
formation
ex
cha
nge,
and
is
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Networked A
u
tom
a
tic Test
System
Bas
ed on Me
ssa
ge-O
r
ie
nted
… (Xiong Fe
nggu
ang
)
7397
respon
sibl
e for re
ceivin
g data publi
s
h
ed by hos
t
comp
uter a
n
d
distri
buting
information
to
sub
s
crib
ers which in
clu
de
host comp
ute
r
s o
r
integr
ated test platform. Integrated
test platform
is
a set
of appli
c
ation
syste
m
in the
serve
r
.
Thro
ugh
me
ssage
-o
riente
d
middle
w
a
r
e
,
integrate
d
te
st
platform
ca
n
re
ceive te
st
inform
ation
of ea
ch
ho
st com
pute
r
in
the n
e
two
r
k acco
rdin
g t
o
deman
d, and
also
ca
n sen
d
co
mman
d
to ea
ch h
o
st
co
mpute
r
to m
a
ke it te
st sp
ecific i
n
form
a
t
ion
for integrate
d
test platform. Throug
h amount of
test data [7] from different
host comput
er,
integrate
d
te
st platform
can
reali
z
e
coord
i
nation
te
st of
we
apon
s
whi
c
h i
n
clu
d
e
s
a
nalysi
s
of te
st
data, simulati
on of test pro
c
e
ss, real-tim
e displ
a
y and
so on.
4. De
sign a
nd Impleme
n
ta
tion o
f
Automa
tic
Te
st Sy
stem
Bas
e
d o
n
M
essag
e-o
r
ie
nted
Mid
d
l
e
w
ar
e
Messagi
ng
m
i
ddle
w
are i
s
resp
on
sible
fo
r
sen
d
ing
an
d receiving m
e
ssag
es on
d
i
fferent
appli
c
ation
systems, an
d
can
reali
z
e
platform-i
nd
epen
dent dat
a excha
nge.
Comm
unica
ting
partie
s
sen
d
and re
ceive messag
e
in
synchrono
us
o
r
a
s
yn
chrono
us mode,
so actual
phy
sical
acce
ss i
s
un
n
e
ce
ssary. Th
erefo
r
e, the
messag
e-
orie
nted middl
eware i
s
ada
pte
d
to cou
p
le lo
ose
host compute
r
up in order t
o
intera
ctive comm
uni
cati
on and
coo
r
di
nated test.
4.1.
Design a
nd Implementa
tion of Mes
s
a
g
e-orie
nte
d
Middle
w
a
r
e
'
s Archi
t
ec
tu
re
The key technolo
g
y of messag
e-orie
nted
middl
e
w
are is m
e
ssag
e tran
sfer, and the
model of me
ssag
e tran
sfer
inclu
d
e
s
point
to point and publi
s
h-su
bscribe.
a) Point-to
-po
i
nt model
In this mod
e
l, the pro
d
u
c
er of messag
e is
called
sen
d
e
r, and th
e consume
r
of messag
e
is calle
d re
cei
v
er. The sen
der send
s a messag
e to a queue an
d is stored in the
queue, and t
he
receiver
obtai
ns me
ssag
e from the
que
u
e
. The di
sadv
antage
s of thi
s
mod
e
l is t
h
at the me
ssa
g
e
sent by
sen
d
e
r
can
only b
e
re
ceived
by only a re
ceiver, and
it sev
e
rely limits th
e use range
and
make
s info
rm
ation sh
arin
g poor.
b) Pub/sub m
odel [8]
In this model,
con
s
um
er of
messag
e is
called a
s
publi
s
he
r an
d co
n
s
ume
r
of me
ssag
e is
calle
d as d
e
scrib
e
r. Th
e di
fference with
point-t
o
-
poi
nt is that a me
ssage p
ubli
s
hed to a subj
ect
can
be
re
ceiv
ed by mo
re
d
e
scrib
e
rs. Pu
b/sub
mod
e
l i
s
a m
odel
ba
sed
on p
u
sh, and m
e
ssa
g
e
can
autom
atically b
r
oad
ca
st to con
s
um
ers.
Th
is mo
del can e
n
su
re that m
e
ssage i
s
sha
r
e
d
widely an
d is able to ada
pt for transfe
r of te
st data betwe
en di
fferent
ho
st comp
uters. T
he
architectu
re o
f
message
-ori
ented middl
e
w
are ba
sed o
n
pub/sub mo
del is sho
w
n
as Figu
re 3.
Figure 3. Architecture of Messag
e-o
r
ient
ed Middle
w
a
r
e
c) Me
ssag
e q
ueue
Messag
e qu
eue is
a co
mmuni
cation
model of a
s
y
n
ch
ron
o
u
s
transfe
r mod
e
betwee
n
different appli
c
ation
s
, and i
n
clu
d
e
s
two conception
s
: messag
e and
queue.
Messag
e is a
carrier of test data, and
mainly con
s
ist
s
of messa
ge
head an
d me
ssage
body. The a
r
chite
c
tu
re of
messag
e is sho
w
n
as
Figure 4. M
e
ssag
e he
ad
is metad
a
ta
of
messag
e, an
d is
ne
ce
ssa
r
y inform
atio
n wh
en m
e
ssag
e delive
r
s, and in
clu
d
e
s
: ID, flag, type,
publi
s
hing
da
te, expire
dat
e, persi
sten
ce, prio
rity
, pu
blish
e
r, recei
v
er, length,
a
nd
so o
n
. The
cod
e
is sho
w
n as follo
wing
:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 739
5
– 7403
7398
publi
c
cla
s
s Messag
e {
publi
c
long ID;
//Message'
s ID
publi
c
int
flag;
//Message'
s
Flag
publi
c
int
type;
//Message'
s
type
publi
c
Date p
ublishing_
dat
e;
//Message'
s p
ublishing dat
e
publi
c
Date e
x
pire_d
ate;
//Message'
s
e
x
pire
data
publi
c
boole
a
n
persiste
n
ce
;
//Message'
s
state of persi
st
ence, the value is
true/fals
e
publi
c
byte
priority;
//Message'
s
p
r
iority
publi
c
Socket
Cha
nnel pu
bli
s
he
r;
//Publishe
r of Messag
e
publi
c
Socket
Cha
nnel rece
iver
;
//Subscribe
r
of Message
publi
c
int
leng
th;
//Message'
s
l
ength
publi
c
byte[]
data;
//Message'
s
content
}
Messag
e
ca
n be
divide
d
into
event
messag
e
an
d no
rmal
m
e
ssag
e, an
d
we
can
disting
u
ish th
em by the val
ue of "flag", in the
me
ssag
e hea
d, wh
ose value i
s
0 t
hat den
otes t
he
messag
e i
s
event me
ssa
ge a
nd i
s
1
that den
ot
es the me
ssag
e is no
rmal
messag
e. Event
messag
e is di
vided into reg
i
ster an
d un
re
gister of ap
pli
c
ation, sub
s
cription an
d un
sub
s
criptio
n
of
distrib
u
tion st
rategy, and we can di
sting
u
ish them
by
the value of "type", in the
messag
e hea
d,
who
s
e
value
is 0 th
at d
enote
s
regi
st
er
of ap
p
lica
t
ion, and
i
s
1 that
den
otes un
regi
ster
of
appli
c
ation, and
i
s
2 de
notes su
bscription
of
di
stribution
strategy, and i
s
3 that de
not
es
unsub
scriptio
n of di
strib
u
tion
st
rategy.
Normal m
e
ssage
en
cap
s
ul
ates te
st d
a
ta of info
rmati
o
n
excha
nge, an
d ha
s 256
kin
d
s of type
s, a
nd we ca
n
al
so di
sting
u
ish
them by the
value of "type"
in the me
ssa
ge he
ad. In g
eneral, we
ca
n distin
gui
sh
any kin
d
s of
messag
e a
c
cordin
g to "fla
g"
and "type" in
the me
ssage
head. In
addit
i
on, "data" i
s
messag
e b
o
d
y
and its cont
ent is te
st d
a
ta
whi
c
h can be
text information, voice,
image an
d othe
r bina
ry information.
Queu
e is al
so a p
ubli
c
st
orag
e
spa
c
e
and it
exist
s
in
memo
ry
or
physi
cal
file, but
gene
rally it exists in mem
o
ry in orde
r to ens
u
r
e the
spe
ed of message tra
n
sfe
r
. Queue
can
be
divided into sende
r que
ue,
event queue
and waiti
ng
queu
e. Send
er que
ue is
use
d
to stori
n
g
norm
a
l message. On the one han
d, messag
e-o
r
ient
ed middle
w
a
r
e pushe
s the norm
a
l message
received f
r
o
m
the p
ubli
s
h
e
r of
normal
messag
e in
to
the
send
er q
ueue,
but on
the othe
r h
a
n
d
,
messag
e-orie
nted middle
w
are al
so obtai
ns me
ssage
f
r
om the send
er que
ue an
d
send
s it to the
sub
s
crib
ers i
n
ord
e
r to
rea
lize me
ssag
e
transfe
r.
Eve
n
t queu
e is
u
s
ed to
stori
n
g
event messa
ge.
On the
on
e h
and, me
ssag
e-o
r
iente
d
mi
ddle
w
are
p
u
shes the
even
t messag
e
re
ceived
from
the
publi
s
he
r of event messa
ge into the e
v
ent
queue,
but on the ot
her h
and, m
e
ssag
e-o
r
ient
ed
middle
w
a
r
e a
l
so obtai
ns
e
v
ent messag
e and ma
ke
it
handle
d
by e
v
ent manag
er. Waiting qu
e
u
e
is u
s
ed to
st
oring
no
rmal
messag
e u
n
succe
ssfully
sent to sub
s
cribers. The
m
a
in rea
s
on th
at
norm
a
l message can't be sent to sub
s
crib
ers
is tha
t
communi
cat
i
on betwe
en
host compute
r
s
discon
ne
cts.
So, on
ce
co
n
nectio
n
re
su
mes,
normal
messag
e in
waiting
qu
eu
e can
be
sen
t
to
the subscribers in order to ensure
the
rel
i
ability of message transfer.
d) Que
ue ma
nage
r
Queu
e mana
ger is respon
sible for ma
n
aging all eve
n
t queue
s which in
clud
es pushin
g
the messa
g
e
into que
ue, poppi
ng me
ssag
e from
q
ueue. In a
d
d
i
tion, in ord
e
r to en
su
re t
h
e
efficien
cy of
pushing
and
poppi
ng m
e
ssage, m
u
lt
i-thread
is
used to ma
nag
ing qu
eue; t
he
method
of p
u
shi
ng a
nd
poppi
ng me
ssag
e is
syn
c
hroni
ze
d in
orde
r to e
n
sure
pu
shing
and
poppi
ng me
ssag
e in ord
e
r
at the same ti
me.
e) Di
strib
u
tio
n
strate
gy manage
r
Ho
w to distri
bute normal
messag
e to subscri
b
e
r
s i
s
determi
ned b
y
distribution
strategy.
So, according to the diff
erence
of
requirement for
message, different
NA
TS will
subscribe
different di
stribution
strategy whi
c
h
form
s a
table of distrib
u
tion
strategy n
a
m
ed
"T_Di
s
tributio
nStrategy". Class of distri
bution s
t
rategy is
s
h
own as
following.
publi
c
cla
s
s Distri
butionSt
r
ategy {
S
o
cket
C
ha
n
nel so
c
k
et
;
int[]
types
;
}
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Networked A
u
tom
a
tic Test
System
Bas
ed on Me
ssa
ge-O
r
ie
nted
… (Xiong Fe
nggu
ang
)
7399
In this
cla
ss,
"so
c
ket" obj
ect i
s
a
so
cket
con
necti
ng NA
TS wit
h
me
ssage
-o
riented
middle
w
a
r
e, and is u
s
ed t
o
descri
b
ing IP and port of
a sub
s
crib
er; "types" is an array in which all
sub
s
crib
ed
messag
e type of a su
bscribe
r
are in
cl
uded. Th
e fu
nction
of dist
ribution
strat
egy
manag
er i
s
t
o
insert a
di
stributio
n
strategy in
to "T
_Di
s
tributio
n
S
trategy", an
d upd
ate dat
e in
"T_Di
s
tributio
nStrategy" throug
h a
dist
ributio
n strategy,
and del
ete data from
"T_Di
s
tributio
nStrategy" thr
ough a di
strib
u
tion strate
gy.
f) Event manager
Event mana
g
e
r i
s
m
a
inly u
s
ed
to p
r
o
c
e
s
sing
the
appli
c
ation'
s event
re
que
sts, i
n
cluding
event re
ceiv
er, se
cu
rity detecto
r, eve
n
t handle
r
,
a
nd re
sp
on
se
notifier. The
pro
c
e
ss flo
w
is
gene
rally divided into the followin
g
step
s:
Step 1: event receive
r
obt
ains a
n
event
message fro
m
event que
ue, and then
se
curity
detecto
r
ch
ecks whethe
r t
he eve
n
t me
ssage
me
et the
system'
s
se
curity; if th
e dete
c
tion
d
oes
not pass, re
spon
se notifier notifie
s the result to the a
pplication,
an
d if the detection passe
s, the
pro
c
e
ss fo
rward
s
step 2;
Step 2: Pa
rsing the
value
of "type" fro
m
the
event
me
ssa
ge; if
the valu
e i
s
0 that
denote
s
the
event is re
gister a
nd process forwards st
ep 3; if the value is 1
that denote
s
the
event is
unre
g
ister an
d p
r
oce
s
s forwa
r
ds
step
4;
if the value i
s
2 that de
not
es the
event
is
sub
s
cript
i
o
n
of
dist
ribut
io
n
st
r
a
t
egy a
n
d
process fo
rward
s
step
5;
i
f
the valu
e i
s
3 th
at de
not
es
the event is u
n
su
bscri
p
tion
of distributio
n strate
gy and pro
c
e
ss fo
rwards
step 6;
Step 3: Gene
rating a
n
obje
c
t of distrib
u
tion st
rategy that contain
s
a
"socket" obj
e
c
t, and
a "types" va
ri
able
who
s
e
value i
s
n
u
ll; a
nd then,
calli
ng di
strib
u
tio
n
strategy ma
nage
r to
add
the
obje
c
t into "T_Di
s
tributio
n
S
trategy", and pro
c
e
ss fo
rwards
step 7;
Step 4: Co
mpari
ng on
e
by one "socket"
obje
c
t in distrib
u
t
ion strate
gy from
"T_Di
s
tributio
nStrategy" with "socket"
obje
c
t
in the event message; if the comp
ari
s
o
n
is
successful, the mat
c
hed
di
stribution strategy
fr
om
"T_Distri
butionSt
r
ategy"
will be del
eted f
r
om
the table a
ccordin
g to
call
ing del
ete fu
nction
of dist
ribution
strategy man
age
r, and p
r
o
c
e
s
s
fo
r
w
ar
ds
s
t
ep 7
)
;
Step 5: Co
mpari
ng on
e
by one "socket"
obje
c
t in distrib
u
t
ion strate
gy from
"T_Di
s
tributio
nStrategy" with "socket"
obje
c
t
in the event message; if the comp
ari
s
o
n
is
su
ccessful, t
he mat
c
he
d
distrib
u
tion
st
rategy from "
T
_Di
s
trib
utio
nStrategy" wi
ll be u
pdate
d
b
y
the event m
e
ssage obj
ect
accordin
g to
calling updat
e function
of
distribution st
rategy
m
anager
to add
the "d
ata" in the
ev
ent me
ssage
to the "t
ype"
variable
in th
e matc
hed
di
stributio
n
stra
tegy
obje
c
t, and proce
s
s forwards ste
p
7);
Step 6: Co
mpari
ng on
e
by one "socket"
obje
c
t in distrib
u
t
ion strate
gy from
"T_Di
s
tributio
nStrategy" with "socket"
obje
c
t
in the event message; if the comp
ari
s
o
n
is
su
ccessful, t
he mat
c
he
d
distrib
u
tion
st
rategy from "
T
_Di
s
trib
utio
nStrategy" wi
ll be u
pdate
d
b
y
the event m
e
ssage obj
ect
accordin
g to
calling updat
e function
of
distribution st
rategy
m
anager
to delete som
e
values
of "type" variable
in the
matche
d distri
bution
strat
egy o
b
je
ct acco
rdin
g to
the "data" in event messa
ge,
and proce
ss fo
rward
s
step 7);
Step 7: Resp
onse notifier
notif
ies
the res
u
lt to NATS.
g) Th
read m
a
nage
r
In messa
g
e
-
oriente
d
mi
d
d
leware, i
n
orde
r to
en
sure the
efficien
cy of m
e
ssag
e
pro
c
e
ssi
ng, t
he e
nqu
eue
(pushing
me
ssag
e into
qu
eue) an
d
deq
ueue
(p
oppi
n
g
me
ssag
e from
queu
e) of ev
ent messa
g
e
and no
rmal
messag
e ar
e
operating at
the re
spe
c
tive threa
d
. Among
them, the logic p
r
ocess
of enque
ue
and de
que
ue
of normal
messag
e is
not the sam
e
and
executio
n e
n
v
ironme
n
t is
not the
sa
me,
so
for the
se
nding
qu
eue
i
n
pe
r
unit tim
e
, the
amou
n
t
o
f
enqu
eue
and
dequ
eue
is
certainly in
co
n
s
iste
nt. We
u
s
e to th
e ratio of the
num
ber
of enq
ue
ue's
messag
e a
n
d
deq
ueu
e's m
e
ssag
e to
de
scribe
the
th
rough
put effici
ency
of sendi
ng q
ueu
e, an
d it
is sh
own as f
o
rmul
a 1.
Thro
ugh
put = numbe
r of normal me
ssa
ge pu
she
d
into query pe
r seco
nd /
numbe
r of no
rmal me
ssag
e poppi
ng fro
m
query pe
r seco
nd
(1)
Mean
while, consi
deri
ng th
e numbe
r of
normal m
e
ssag
e is en
ormous, if thro
ughp
ut
efficien
cy co
ntinue
s lo
w,
an a
m
ount
of
no
rmal
me
ssag
e
will b
e
piled
up i
n
th
e sendi
ng
qu
eue
over time. O
n
the one h
and, it will a
ffect t
he efficiency of the
impleme
n
tation of messa
ge-
oriente
d
mid
d
leware; o
n
the othe
r ha
n
d
it c
an al
so
make
appli
c
ation re
ceive
a messa
ge
in
delay. Thu
s
, by adding th
read ma
nag
e
m
ent into
me
ssage
-o
riente
d
middle
w
a
r
e
and in
cre
a
si
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 739
5
– 7403
7400
or redu
cing t
he num
ber
o
f
dequeu
e' th
read
acco
rdi
ng to thro
ug
hput in o
r
de
r to balan
ce t
h
e
efficien
cy of
enqu
eue
and
dequ
eue
of
norm
a
l me
ssage. Th
e n
u
m
ber of de
q
ueue'
s th
rea
d
is
sho
w
n in formula 2.
4.2.
Pub/sub of
Messa
ge-oriente
d Middlew
a
r
e in NAT
S
a) Co
nstitue
n
t
element of messag
e'
s pu
b/sub
Pub/sub
of
messag
e in
NATS con
s
ists of
pu
blish
e
r, sub
s
cribe
r
, me
ssage
center
and
table of di
stri
bution
strate
g
y
, so
we
can
descri
be the
pub/sub of m
e
ssag
e with
a
four-tu
p
le
(P, S,
C, T). P is a set of the mess
age's publisher
{p1,
p2, ..., pn}, and S is a set of message's
s
u
bsc
r
iber
{s1, s
2
, ..., s
n
}.
P may intersec
t
with S,
so that is
to say that
the mess
age's publisher
and me
ssag
e
'
s sub
s
crib
er
are n
o
t ab
sol
u
te, and they
can
be a u
n
i
ty (both a p
ublisher
and
a
descri
b
e
r). C
is the messa
ge's
cente
r
, and co
nsi
s
ts o
f
messa
ge qu
eue, queu
e manag
er, event
manag
er, thread ma
nag
e
r
, distri
bution
strategy
m
anag
er, an
d
is re
sp
on
sible for
cont
act
messag
e'
s p
ublisher
and
sub
s
crib
er. Not only message-ori
ented
middle
w
a
r
e receive
s
me
ssage
publi
s
hed by
messag
e'
s pu
blish
e
r, but also send
s me
ssage to me
ssag
e's
sub
s
criber. Wh
en C
is
equal
to 1,
all
of the
me
ssa
ge's pu
blishin
g
an
d
sub
s
cri
p
tion u
s
e
a
m
e
ssag
e
cente
r
, an
d it i
s
th
e
simple
st
m
e
ssag
e's
pu
b/sub system. Usually, with
in NATS
the
r
e are multiple
messag
e cen
t
ers,
so C i
s
gre
a
ter than 1. T is a table of d
i
stribut
io
n strategy whi
c
h
ensure
s
correct distri
butio
n of
messag
e bet
wee
n
publi
s
h
e
rs
with su
bscrib
ers.
b) Pro
c
e
ss of
message
's p
ub/su
b
Step 1: NA
TS start
s
th
e functio
n
o
f
messag
e'
publi
s
hing,
a
nd then
ge
n
e
rate
s a
messag
e: "msg", and then
sets the value of "flag"
in
the messa
ge to 1 (It sets the messag
e as a
norm
a
l message in fact),
and then sets the value o
f
other variab
le in the messag
e, at last, set
test data to "msg" in the m
e
ssag
e;
Step 2: NATS send
s the "m
sg" to messa
ge ce
nter;
Step 3: After
messag
e'
s ce
nter re
ceiving
the
message
, acco
rdin
g that the flag's value is
one, the me
ssage
can
b
e
determined
as a n
o
rm
a
l
messag
e, a
nd then p
a
rsing the type
o
f
messag
e jud
ges
wheth
e
r
a ''types'
'
in some di
stri
buti
on strategy contain
s
the type of messa
g
e
; if
a "types" in some di
stri
bution strate
gy cont
ain
s
the type of
messag
e, then there i
s
ATS
sub
s
cribi
ng t
he me
ssag
e, and
then
a
c
cording
to q
u
eue m
ana
ger ,the me
ssag
e is pu
sh
ed i
n
to
sen
d
ing
que
u
e
; if not a "types" i
n
some
distri
but
ion
strategy contai
ns the
type o
f
messag
e, so
there is n
o
su
bscrib
er to su
bscrib
e the messag
e and there i
s
unn
ecessary to pu
sh the messag
e
into sen
d
ing
queu
e;
Step 4: Message'
s ce
nter
parse
s
the value of ''pe
rsisten
c
e'' in m
e
ssag
e; if the value is
one th
en th
e
messag
e i
s
stored
in th
e d
a
taba
se, oth
e
r
wi
se t
he m
e
ssage
do
es n
o
t sto
r
e i
n
th
e
databa
se;
Step 5: Message'
s ce
nter resp
on
se
s to ATS t
hat the
messag
e is receive
d
su
ccessfully.
c) Mai
n
pro
c
e
ss of me
ssag
e' sub
s
cri
p
tio
n
Step 1: ATS
start
s
the fun
c
tion of me
ssage'
s su
bscri
p
tion;
Step 2: Pop
p
ing
a me
ssage from th
e
se
ndi
n
g
q
u
eue, a
nd
set
t
ing the
messag
e a
s
"msg", an
d p
a
rsi
ng th
e "type" in th
e "m
sg", an
d final
ly getting a
d
i
stributio
n
strategy from t
h
e
"T_Di
s
tributio
nStrategy" an
d then namin
g it as "ds ";
Step 3: If the
"types" in "d
s" contai
ns the
ty
pe of me
ssage, the
n
a
c
cordin
g to "
s
o
c
ket" in
"ds", sendin
g
the me
ssag
e to sub
s
crib
ing ATS;
if sendin
g
the
m
e
ssag
e un
su
ccessfully, th
en
assigni
ng the
socket to the
"receive
r" in
messag
e and
pushi
ng it to waiting q
ueu
e;
Step 4: If existing unvi
s
ite
d
distrib
u
tion
strat
egy in "T
_Di
s
tributio
n
S
trategy", then gettin
g
a next distrib
u
tion strate
g
y
and namin
g it as "ds"
, and jumpi
ng
step 3; if all the distrib
u
tio
n
strategi
es a
r
e
visited, then jumping
step
5;
Step 5: If th
ere is me
ssa
ge in sen
d
in
g queue,
the
n
jumping st
ep 2; if there is not
messag
e in sendin
g
queu
e
,
then waiting
until receivin
g messag
e a
nd jumpin
g st
ep 2.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Networked A
u
tom
a
tic Test
System
Bas
ed on Me
ssa
ge-O
r
ie
nted
… (Xiong Fe
nggu
ang
)
7401
d) Main p
r
o
c
e
ss of sub
s
crib
ing and di
stri
buting
Step 1: ATS
start
s
the fun
c
tion of su
bscription an
d di
stributio
n;
Step 2: ATS
gene
rate
s
a
messag
e
na
med
as "msg
", and
set
s
th
e flag
of "m
sg" to
ze
ro,
and
set
s
the
type of "msg" to two,
a
nd
conve
r
t the
content
o
f
messag
e a
s
byte a
r
ray
and
assigni
ng it to data of "msg
";
Step 3: ATS
sen
d
s "m
sg" to messag
e'
s cente
r
;
Step 4: After
messag
e'
s ce
nter re
ceiving
the
message
, acco
rdin
g that the flag's value is
zero, the me
ssag
e can b
e
determi
ned
a
s
an
event
m
e
ssag
e, and t
hen
calling
e
v
ent manag
er to
compl
e
te su
b
s
cription of di
stributio
n stra
tegy.
e) Main p
r
o
c
e
ss of un
su
bscribing di
stri
bu
tion strategy
Step 1: ATS
start
s
the fun
c
tion of
un
su
bscribi
ng di
stribution
strate
gy;
Step 2: ATS
gene
rate
s
a
messag
e
na
med
as "msg
", and
set
s
th
e flag
of "m
sg" to
ze
ro,
and
set
s
the
type of "m
sg" to three,
and
co
nver
t t
he
conte
n
t of
messa
ge
as byte array
a
n
d
assigni
ng it to data of "msg
";
Step 3: ATS
sen
d
s "m
sg" to messag
e'
s cente
r
;
Step 4: After
messag
e'
s ce
nter re
ceiving
the
message
, acco
rdin
g that the flag's value is
zero, the me
ssag
e can b
e
determi
ned
a
s
an
event
m
e
ssag
e, and t
hen
calling
e
v
ent manag
er to
compl
e
te un
subscri
p
tion of
distributio
n strategy.
5. Applicatio
n of ATS Ba
sed on Mes
s
a
ge-oriented
Middle
w
a
r
e
For
example,
in ATS of
co
ordin
a
ted
en
gagem
ent of
air-gro
und
weapo
n which
is
sho
w
as
Figu
re
6, first a
e
ri
al reconnai
ssan
ce
investigate
s
and l
o
cates e
nemy's targ
ets in
air,
se
a a
n
d
land; and the
n
enemy'
s inf
o
rmatio
n is p
ublished
via
messag
e-orie
nted middle
w
are, an
d gro
u
nd
control
cente
r
sub
s
cribe
s
the ene
my's i
n
formatio
n
while
g
r
oun
d control ce
nter also su
bscrib
es
longitud
e
, altitude, height,
spe
ed, a
z
i
m
uth
and o
t
her info
rmat
ion which
are publi
s
he
d
by
unman
ned a
e
r
ial vehicl
es a
nd fighter, as
well a
s
lo
cati
on, transmit status whi
c
h a
r
e publi
s
h
ed
by
pre
c
isi
o
n
-
gui
ded mi
ssil
e
s;
after analy
z
i
ng the info
rm
ation pu
blish
ed by ae
rial
reconn
aissa
n
ce,
unman
ned
a
e
rial ve
hicl
e
s
, fighter a
nd p
r
e
c
is
i
o
n
-
guid
ed
missile
s, groun
d control
center
publi
s
he
s inf
o
rmatio
n of d
a
mage
d ta
rg
et to me
ssag
e-o
r
iente
d
mi
ddle
w
are, an
d ba
se
d on
the
distrib
u
tion strategy, messag
e-o
r
ie
nted
middlew
are
distribute
s
the inform
atio
n to preci
s
io
n-
guide
d mi
ssil
e
, fighter and
unma
nne
d a
e
rial ve
hicl
es; finally, unm
anne
d a
e
rial
vehicle
s
, figh
ter
and
pre
c
i
s
io
n-gui
ded
mi
ssile
s atta
ck
the target a
nd p
ubli
s
h t
he d
a
mag
e
d
inform
ation
to
messag
e-orie
nted mid
d
le
ware, an
d
evaluati
on
control
ce
nte
r
sub
s
cribe
s
the dam
a
g
e
informatio
n, and co
ndu
cts
a comp
re
hen
sive analy
s
is.
Figure 6. ATS of Coordi
na
ted Engage
m
ent of Air-gro
und Weap
on
6. Test and
Analy
s
is
6.1. Test Tar
g
et
Messag
e-orie
nted middle
w
are mai
n
ly provi
des a
co
mmon info
rm
ation co
mmu
nicatio
n
platform for
ATS, s
o
tes
t
mainly c
o
ns
iders
tw
o a
s
p
e
cts: first, the time that
messag
e-orie
nted
middle
w
a
r
e d
i
stribute
s
m
e
ssage
s
to su
bscrib
ers wh
ether ca
n
me
et the re
qui
re
ments;
se
con
d
,
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 739
5
– 7403
7402
the middle
w
are
system
can
wh
eth
e
r di
strib
u
te
accurately messag
es to the sp
ecified
sub
s
crib
er
s.
6.2. Test Env
i
ronment an
d Test
Data
Software E
n
vironme
n
t: Wind
ows XP
ope
rating
system
and
ATS of co
ordin
a
ted
enga
geme
n
t of air-g
rou
nd
wea
pon.
Hardware Environme
n
t: There a
r
e 7 co
mputers, and
a compute
r
is used as a
messag
e-
oriente
d
mid
d
l
ewa
r
e
se
rver, and oth
e
r si
x comp
ut
ers are used
to simulating
th
e host com
pute
r
of unman
ned
aerial vehi
cl
es, fighter, p
r
eci
s
io
n-
guid
ed missile, a
e
rial reconn
a
i
ssan
ce, grou
nd
control cente
r
an
d
eval
uat
ion control center.
T
h
o
s
e
com
pute
r
s'
config
uratio
n
s
a
r
e Intel
Core2
Quard CP
U 2
.
40 GHZ a
nd
RAM 2G.
The ho
st co
mputer
of ae
rial reconnai
ssan
ce
publi
s
hes i
n
form
ation of ene
my, the ho
st
comp
uter of
unmann
ed aerial vehi
cle
s
, aircraft an
d pre
c
isi
on-g
u
ided mi
ssil
e
s publi
s
h th
eir
informatio
n,
whe
n
the
len
g
th of m
e
ssa
ge in
sendi
ng
que
ue i
s
eq
ual to
200
0,
messag
e-orie
nted
middle
w
a
r
e
starts to
di
stri
bute m
e
ssag
e, whil
e the
host
co
mput
er
of ae
rial
reco
nnai
ssan
ce,
unman
ned
a
e
rial vehi
cle
s
an
d preci
s
ion
-
gui
ded
missil
e
s
stop
publishing i
n
formatio
n, and
distrib
u
tion time will be re
cord in Tabl
e 1. In the
sam
e
way, distri
b
u
tion time of distrib
u
ting 4
000,
5000 a
nd 10
0
00 messa
g
e
s
will be re
co
rd in Table 1.
Table 1. Di
stribution Time
of Message
-o
riented Mi
ddl
ewa
r
e
2000 M
e
ssage
4000 M
e
ssage
5000
M
e
ssage
10000
M
e
ssage
total distribution tim
e
（
M
illisecond
）
120159
232877
311362
654186
Distribution tim
e
per reco
rd
（
M
illisecond
）
60 58
62
65
The ho
st co
mputer of un
manne
d aeri
a
l vehi
cle
s
, aircraft and preci
s
ion
-
gui
de
d missil
e
s
publi
s
h da
ma
ged info
rmati
on 10
00 time
s. The
stat
e
m
ent of dist
ri
bution me
ssa
ge is
sho
w
n
in
Table 2.
Table 2. Statement of Dist
ributing M
e
ssage
Sender
host com
pute
r
of
unmanned
aerial vehicles,
Host com
put
er of
aircraft
Host com
put
er of
precision-
guided m
i
ssiles
Receiving tim
e
s
1000
1000
1000
6.3. Result
Analy
s
is
By compari
n
g
these two
se
ts of above test, we ca
n draw the con
c
lu
sion
s:
(1) T
he ba
sic di
stri
buti
on efficien
cy
of
messag
e-o
r
iente
d
middle
w
a
r
e
is abo
ut
60s/record, a
nd di
strib
u
tin
g
40
00
message
wa
ste
s
a
t
the lea
s
t, which
sho
w
s th
at the n
u
mb
e
r
of
thread
s of distributing me
ssag
e is optim
al.
(2) Th
e Accu
racy rate of messag
e-orie
nt
ed middle
w
are di
strib
u
ting messag
e is 100%,
and this in
dicates that the distrib
u
tion a
c
cura
cy
of the messag
e-oriented middl
e
w
are is reliabl
e.
7. Conclusio
n
Aim at the cu
rre
nt status
o
f
the ATS that fo
cuse
s o
n
only on a sin
g
le or the
sa
me type
of weap
on
s' perfo
rman
ce
testing, this p
aper int
r
od
uces me
ssage
-oriente
d
midd
lewa
re into A
T
S
in order to no
t only shi
e
ld
complexity of the bottom
test equipm
ent,
but also impl
ement te
st da
ta
transfe
r bet
ween differe
nt test equipm
ent throug
h the messag
e
'
s publi
s
h a
n
d sub
s
criptio
n
.
Acco
rdi
ng to test on me
ssage-ori
ented
middle
w
a
r
e a
bout publi
s
hi
ng and
sub
scribing m
e
ssa
g
e
we can draw
the con
c
lu
sio
n
that
the message
-o
riente
d
middle
w
a
r
e
can create a
n
integrate
d
test
platform for a
variety of weapon
s an
d im
plement
com
p
reh
e
n
s
ive p
e
rform
a
n
c
e t
e
st for weap
o
n's
coo
r
din
a
ted e
ngag
ement.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Networked A
u
tom
a
tic Test
System
Bas
ed on Me
ssa
ge-O
r
ie
nted
… (Xiong Fe
nggu
ang
)
7403
Referen
ces
[1]
Ying
lia
ng F
e
n
g
,
Shujua
nW
an
g, Bin F
eng, RuiW
an
g, Yuan
He, T
ong Z
hang. Devel
opme
n
t
of an auto
test sy
st
em for humidit
y
s
ensors.
Sensors an
d Actuators A: Physica
l
. 200
9; 152(1): 10
4-1
09.
[2]
Lai Qi
ng,
Hu
a
Jian-W
e
i,
Lü
Y
un, C
hen
Yu
e-
F
e
i, Xu
Ji
an.
Rese
arch on g
ener
al
r
e
la
y pr
otection
a
u
to-
test s
y
stem
soft
w
a
re.
Pow
e
r System
Protec
tion an
d
C
ontr
o
l
. D
i
an
li
Xiton
g
Ba
oh
u
yu
K
ongz
hi.
201
0
;
38(3): 90-
94.
[3]
Yuan Qi
ngfe
n
g
,
Lu Hui, Sh
en
Shitua
n. Auto t
e
st
s
y
stem r
e
s
ource
descri
p
ti
on meth
od
bas
ed o
n
XML.
Journ
a
l of Bei
jing
Univ
ersity
of Aerona
utic
s and Astron
a
u
tics
.
Beiji
ng
Han
g
kon
g
Ha
ngtia
n Da
xue
Xu
eb
ao. 20
10;
36(1): 11
4-1
1
7
.
[4]
Z
hai L
i
fan
g
, Li
Chu
n
y
u
an, S
un L
i
p
i
ng.
Res
earch
on
the
messag
e
-ori
en
ted mid
d
l
e
w
a
r
e
for
w
i
rel
e
s
s
sensor n
e
t
w
ork
s
.
Journal of C
o
mputers
. 20
1
1
; 6(5): 104
0-1
046.
[5]
Miao C
hun-Y
u
,
Shi Mei-Li
n,
Jian
g Jin-L
e
i.
MOM-S: Message-
orie
nted
middl
e
w
a
r
e ba
sed on W
e
b
service.
Jour
na
l on Co
mmunic
a
tions
. T
ongxin
Xu
eba
o. 200
6;
27(11): 96-
10
0.
[6]
Jia Z
h
i
y
a
n
, H
a
n
Xie.
Net
w
ork
autom
atic test
s
y
stem
bas
ed
on d
a
tab
a
se m
i
ddl
e
w
a
r
e.
Lect
u
re N
o
tes i
n
Electrical E
ngi
neer
ing. 2
013;
218(
3): 603-
60
9.
[7]
Jian
Ni, Yu
nl
o
ng Z
h
ang.
Stu
d
y
of test
dat
a g
ener
atio
n method
bas
ed
on evo
l
utio
na
r
y
alg
o
rithm.
Te
lkom
n
i
ka
. 20
14; 12(1): 8
18-
822.
[8]
Carril
ho S
e
rg
i
o
, Esaki
Hiros
hi.
A pu
b/sub
messag
e
d
i
stri
butio
n arc
h
itec
ture for d
i
sru
p
t
ion to
lera
nt
net
w
o
rks.
IEICE T
r
ansactions
on Informatio
n
and Syste
m
s
. 2009; E9
2-D(1
0
): 1888-
18
96.
Evaluation Warning : The document was created with Spire.PDF for Python.