OLIMPO, AN AD-HOC WIRELESS SENSOR NETWORK
SIMULATOR FOR OPTIMAL SCADA-APPLICATIONS
J. Ba bancho, F.J. Molina, C. Le´on, J. Rope o and A. Ba bancho
Depa amen o de Tecnolog´ıa Elec ´onica, Uni e si y o Se ille,
C/ Vi gen de ´
A ica, 7, Se ille 41011, Spain
Phone.: (+034) 954 55 71 92, Fax: (+034) 954 55 28 33.
e-mail: [email p o ec ed]
ABSTRACT
This pape in oduces OLIMPO, an use ul simula ion ool
o esea che s who a e de eloping wi eless senso com-
munica ion p o ocols. OLIMPO is a disc e e-e en simula-
o design o be easily econ igu ed by he use , p o iding a
way o design, de elop and es communica ion p o ocols.
In pa icula , we ha e designed a sel -o ganizing
wi eless senso ne wo k o low da a a e. Ou p emise is
ha , due o hei inhe en sp ead loca ion o e la ge a eas,
wi eless senso ne wo ks a e well-sui ed o SCADA appli-
ca ions, which equi e ela i ely simple con ol and moni-
o ing.
To show he acili ies o ou simula o , we ha e s ud-
ied ou ne wo k p o ocol wi h OLIMPO, de eloping se -
e al simula ions. The pu pose o hese simula ions is o
demons a e, quan i a i ely, he capabili y o ou ne wo k
o suppo his kind o applica ions.
KEY WORDS
Ad-hoc wi eless senso ne wo k, mul i-hop Packe Radio
Ne wo k (PRN), E en Disc e e Schedule (EDS), ne wo k
simula o , SCADA applica ions.
1 In oduc ion
Wi eless da a ne wo ks ha e ecei ed esea ch in e es in
ecen yea s due o he desi e o connec i i y. New mul-
imedia se ices, such as Wo ld Wide Web, e-mail, and
da a ile ans e s ha e been p o ided by di e en wo king
g oups. Two e y well known examples a e he 802.11 and
he 802.15 Ins i u e o Elec ical and Elec onics Enginee s
(IEEE) wo king g oups. 802.11 has designed se e al p o-
ocols de ining he Wi eless Local A ea Ne wo k (WLAN)
s anda d [1]. 802.15 wo ks in he design o se e al s an-
da ds based on ne wo ks ha employ no ixed in as uc u e
and ha e communica ion links less han 10 me e s (known
as Pe sonal Ope a ing Space, POS) in leng h (Wi eless Pe -
sonal A ea Ne wo ks, WPAN) [2]. Bo h o ganiza ions a e
de eloping he de ini ion o p o ocols wi h da a h ough-
pu s g ea e han 100 Mb/s.
Howe e he e a e o he po en ial wi eless ne wo k
applica ions, wi h elaxed h oughpu equi emen s. These
applica ions a e indus ial con ol and moni o ing, home
au oma ion and consume elec onics, secu i y and mili-
a y sensing; asse and in en o y acking; supply chain
managemen ; in elligen ag icul u e; and heal h moni o -
ing [3]. These low-da a- a e applica ions in ol e sensing
o one o m o ano he . This is he eason why hese ne -
wo ks a e usually called wi eless senso ne wo ks, o Low-
Ra e WPANs (LR-WPANs), because hey equi e sho -
ange links wi hou a p eexis ing in as uc u e. Some o
he cha ac e is ics o an LR-WPAN a e desc ibed by he
s anda d 802.15.4 [4] and a e ela ed below:

O e - he-ai da a a es o 250 kb/s, 40kb/s, and 20 kb/s

S a o pee - o-pee ope a ion

Alloca ed 16 bi sho o 64 bi add esses

Alloca ion o gua an eed ime slo s (GTs)

Ca ie sense mul iple access wi h collision a oidance
(CSMA-CA) channel access

Fully acknowledge p o ocol o ans e eliabili y

Low powe consump ion

Ene gy de ec ion (ED)

Link quali y indica ion (LQI)

16 channels in he 2450 MHz band, 10 channels in he
915 MHz band, and 1 channel in he 868 MHz band.
We ha e ocused ou esea ch in ne wo ks ha sup-
po his kind o applica ions. In pa icula , we a e e y
in e es ed in echnologies o Au oma ic Me e Reading
unc ions (AMR) and Dis ibu ion Au oma ion (DA) o
u ili ies applica ions (wa e , gas and elec ici y) [5][6].
These applica ions a e o en dis ibu ed h oughou many
compu e s. Howe e , cus ome da a polled om senso
ne wo ks, que ies, emo e con ol o de s and ne wo k man-
agemen messages mus be p ocessed by a unique compu e
named U ili y Con olle (UC). This en i y wo ks as a mas-
e in a Supe iso y Con ol and Da a Acquisi ion (SCADA)
sys em. UC con ols and collec s da a om many emo e
uni s (senso nodes). The e a e many s udies which show
ha solu ions based on low powe adio ne wo ks a e i-
able and ha hey o e he bes cos /pe m o mance a io
[7][8][9]. Since he ea ly 90’s, adio and mic ocon olle
echnologies ha e allowed he de elopmen o sma sen-
so ne wo ks. Many esea ch g oups ha e ocused hei in-
es iga ions in he ne wo k op imiza ion. Howe e nei he
o hem ha e de elop a ne wo k o op imizing his kind o
applica ions.
The en i onmen in which we a e de eloping ou
s udy has he ollowing ea u es:

Communica ion is usually be ween nodes and he UC.

Nodes a e densely deployed wi hin he ange a ea.

Topological changes a e possible (a node can be added
o elimina ed i adio ange changes).

Nodes mus implemen mul i-hop mechanisms o le
some kind o packe s each he des iny node.

Nodes can be loca ed wi hou p e-planning.

Nodes a e p one o ailu e.

Nodes ha e no mobili y.
Ad-hoc ne wo ks a e bes sui ed o suppo hese ea-
u es. We p opose (in his pape ) a simula ion ool o s udy
a mul i-hop sel -o ganizing wi eless senso ne wo k. We
ha e called his simula o OLIMPO.
This pape is o ganized as ollows. In sec ion 2, a
b ie summa y o senso ne wo k pe o mances objec i es
is gi en. In sec ion 3, ou ne wo k p o ocol is p esen ed.
In sec ion 4, we desc ibe he OLIMPO design. In sec ion
5, he esul s o ou ne wo k simula ion o e OLIMPO a e
discussed. Concluding ema ks a e gi en in sec ion 6 and
u u e wo ks a e lis ed in sec ion 7.
The goal o his pape is o ou line he design o ou
sel -o ganizing mul i-hop ne wo k p o ocol, showing e-
sul s o se e al expe imen s made wi h ou ne wo k sim-
ula o , OLIMPO.
2 Senso Ne wo k Pe o mance Objec i es
Wi eless senso s design mus be de eloped assuming he
ollowing de ices’ cha ac e is ics:

Low powe consump ion and low supply ol age.

Low cos .

Flexibili y.

Sho ansmission ange.

Embedded p ocesso implemen ing a dis ibu ed in el-
ligence p ocess [10].

Mic oelec omechanical sys ems (MEMS), ac ua o s
and adio echnology in eg a ed in a educed space
[11].
App App App App App
Ne Machine
Rou e Ne wo k
Manage
Llc BROADCAST
MacCon ol
BROADCAST
Llc
MacCon ol
MAC
Rou ing
Table
P oximi y
Table
ST
ACK In To Base
ACK
ExR
ACK/NoF a NoACK
NoF a
Link ExR
F agmen
Collec Ala m Telecon ol Teleme y Poll a me e
LDa a.Ind
LDa a.Con
LDa a.Req
LinkQueue
LDa a.Con
LE o .Ind
LDa a.Ind
SAP 0 SAP 1 SAP 2 SAP 3
Figu e 1. OLIMPO p o ocol s ack
In ha sense, he e is a end o de elop and comme -
cialize his kind o embedded sys ems. This is he case o

an in e na ional associa ion o semi-
conduc o companies, so wa e p o ide s, sys em in eg a-
o s, and o iginal equipmen manu ac u e s (OEMs) such as
ZMD

, ATMEL

CompXs

, Mic ochip

, be onic

and chipcon

[12]. Ano he example o his end is
neuRFon

senso , om Mo o ola

Labs, a ela i e sim-
ple de ice ha has been inspi ed in senso y ecep o s o he
ne ous sys em, which a e capable o de ec ing changes in
he en i onmen due o s imuli [13].
Gi en he na u e o hese nodes, ou ne wo k mus as-
sume some limi a ions in he ne wo k pe o mance. Due
o he low adio powe (less han 100 mW), he low ba -
e y li e and he noisy RF band (Indus ial Scien i ic Band,
ISM) he da a h ough pu will be qui e low. Simula ions
made by Mo o ola

wi h his neuRFon

ne wo k show
an a e age da a h oughpu o 40 kbps wi h a aw bi a e o
250 kps [14][15]. Fu he mo e, he packe s’ mul i-hop ne-
cessi y o eaching he end, inc ease he messages la ency.
Expe imen s made by Mo o ola

wi h a neuRFon

ne -
wo k, o med by 64 nodes wi h en me e s o adio ange
and a maximal numbe o ou hops in o de o each he
end, show a messages la ency o 2.6 seconds [14][15].
Ano he e y impo an issue we ha e o conside is
he du y channel ime. This ime is closely ela ed wi h he
channel con en ion. Tha ’s why we mus conside he in lu-
ence o he design o he Medium Access Con ol (MAC)
laye on he collision e ec . Slo ed and unslo ed chan-
nel access a e p o ided in se e al MAC designs. We ha e
measu e his i em in ou simula ion.
3 Sel -O ganizing Senso Ne wo k P o ocol
and SCADA Applica ion
Many ne wo k p o ocols ha e been p oposed ecen ly
[16][17][18][19][20]. Howe e , we ha e designed a ne -
wo k p o ocol o op imize SCADA applica ions [21] [22].
We ha e ocused ou esea ch on a modi ied unslo ed
CSMA-CA channel access mechanism. Each ime a node
wishes o ansmi da a ames, i will wai o a andom
pe iod. I he channel is idle, he de ice will ansmi he
da a. I he channel is ound o be busy, he node wai s o
ano he andom pe iod be o e ying o access he channel
again.
The design has been inspi ed ollowing he ISO model
o a chi ec u e o Open Sys ems In e connec ion (OSI)
[23]. The p o ocol s ack de ined uses wo link laye s p o-
ocol as depic ed in igu e 1. Wi h his s uc u e e e y node
is capable o p ocess - ansmi and ecei e- b oadcas pack-
e s while main aining a da a link dedica ed o o he node.
I is possible o selec he ype o da a link -wi h o wi h-
ou agmen a ion- and he simul aneous numbe o hem.
Acknowledgemen se ice is a ailable.
Ou mul i-hop ne wo k p o ocol c ea es a spanning
ee logical s uc u e o e he physical opology. All nodes
de elop an algo i hm o main ain his logical backbone
whe e he oo o he ee -we e e i as base s a ion- will
ha e access o all ne wo k nodes. In addi ion, e e y node
will be able o ou e i s packe s o he base s a ion.
Nodes always know he pa h o oo h ough he pa -
en node, bu any selec ed algo i hm mus allow he oo
node (UC) o ha e an app oxima e image o cu en ee
opology. This way, packe s om nodes o he oo do
no con ain in o ma ion abou ou ing. Con e sely, mes-
sages om oo o nodes gene ally use an explici ou -
ing scheme, so packe s mus con ain in o ma ion abou he
pa h. Packe heade size mus be op imized because adio
ames should be as sho as possible o educe he ans-
mission e o a io, e ec i e bandwid h, e c.
The collec ing algo i hm is shown in igu e 2. An
applica ion laye in all he nodes ecei es he o de . This
node passes he oken o one o i s child en ( igu e 2(2)),
and wai s o a esponse. When i is ecei ed, ( igu e 2(3))
he node sends da a o he oo , and i passes he collec o -
de o he nex child ( igu e 2(4)). Only when he e a e no
mo e child en o be polled, will he node answe wi h i s
own da a ( igu e 2(6)). Only one oken is being passed be-
ween pa en and child en, so no pa h heade in o ma ion
is necessa y, and he answe oken is simul aneously low-
ing o he oo . To p e en a los oken, a imeou pe iod
gua an ees he oken passes o he nex child.
1
2
34
5
6
0
100
120 125
Token sen
Token o pa en
Da a o base s a ion
Figu e 2. Collec ing scheme
Node s uc u e
c ea ion
Global
ini ia ion
Channel
access
MAC
execu ion
Link
Manage
execu ion
Enabled
ime s
e esh
E en ?
Uppe laye
execu ion
NODE
EXECUTION
Yes
No
Figu e 3. Execu ion model o a node in an OLIMPO sim-
ula ion
Figu e 4. 127 nodes ne wo k simula ion, wi h OLIMPO
4OLIMPO, simula o design
The mo i a ion o de elop a new ne wo k simula ion ool
is he need o capabili y o ack speci ic a ic -SCADA
packe s- om one node o base s a ion h ough mul i-hops.
The e a e o he ne wo k simula o s such as OPNET [24],
ns-2 [25], WinneuRFon [17], Senso SimII [26]. Howe e
we ha e designed a new simula o based on he E en
Disc e e Schedule (EDS), o suppo ou ne wo ks appli-
ca ions o e a high numbe o senso nodes. OLIMPO
has been c ea e wi h C++ Bo land

Builde 6, c ea ing
a iendly G aphical Use In e ace (GUI).
OLIMPO is open sou ce, p o iding an use ul ool o
simula e new communica ion o applica ion p o ocols o
o he esea che s.
The cu en e sion o OLIMPO p o ide a g aph-
ical expe imen con igu a ion. Use can c ea e se e al
expe imen s changing easily he simula ion pa ame e s,
e.g.: ime s, andom p ocess, adio ange, nodes deploy-
men . . . OLIMPO also p o ide se e al simula ion modes -
MAC s ep, Logical Link Con ol (LLC) s ep, ne wo k s ep
and applica ion s ep- p o iding a way o debug p o ocols.
While execu ing he simula ion many epo s a e gene a ed
in eal ime illing a ex ile, ha is closed and sa ed when
he simula ion ends.
Figu e 3 shows he execu ion model o a node, and
igu e 4 shows he GUI c ea ed.
5OLIMPO Simula ion and Resul s
The simula ed ne wo k model includes a maximal numbe
o 127 senso nodes popula ed o e 600 me e s x 600 me-
e s a ea. The deploymen o nodes has been chosen in an
hexagonal dis ibu ion, o ming a dense ne wo k, whe e e -
e y node has a adio ange o 60 me e s and consequen ly
he e is a maximal numbe o neighbo s o 6. Figu e 4
shows he spanning ee gene a ed. Each node es ablish a
0 5 10 15 20 25 30 35 40
0
1
2
3
4
5
6
7
Numbe o nodes
A e age end- o-end delay (sec)
Figu e 5. A e age End o End Delay
logical pa en -child ela ionship wi h i s neighbo s, look-
ing o he bes pa h o he oo . The numbe abo e e e y
node ep esen s he MAC add ess. Node 0 -shown in yel-
low colo in igu e 4- is assumed o be he oo and i s adio
ange is depic ed wi h a ci cle. I ’s placed in he cen e o
he ne wo k opology.
The simula ion was gene a ed wi h a unslo ed
CSMA-CA channel access con igu a ion, and packe s wi h
and wi hou acknowledgemen . Ou objec i es we e o
measu e he pe o mance issues ela ed in sec ion 2 and
o ob ain e icien ly SCADA applica ion da a om all he
nodes.
We ha e c ea ed se e al expe imen s, inc emen -
ing he numbe o ne wo ked nodes un il 127, al hough
OLIMPO hasn’ es ic ions on he maximal numbe o ne -
wo ked nodes.
A e he simula ion we ha e ob ain some ele an e-
sul s. The ime o spanning ee c ea ion is ela ed in able
1. These imes depend on he con igu a ion imes (e.g.:
Table 1. Time o Spanning T ee C ea ion
Numbe o nodes Time o C ea ion (sec)
7 7
14 11
21 23
28 25
37 31
64 68
127 130
ime o beacon). Anyway, we conside i o be easonable
good esul because his o ma ion ime is a e y sho ime
compa ed wi h he ypical SCADA applica ion ime. A e
he spanning ee was c ea ed, we execu ed he SCADA ap-
plica ion p oposed. Da a applica ion we e collec ing by he
UC success ully, ob aining he ollowing ne wo k pe o -
mance esul s:
0 5 10 15 20 25 30 35 40
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Numbe o nodes
A e age collision ime (%)
Figu e 6. A e age Collision Time
0 5 10 15 20 25 30 35 40
0
0.5
1
1.5
2
2.5
3
3.5
4
Numbe o nodes
A e age du y ime (%)
Figu e 7. A e age Du y Time

The a e age da a packe deli e y end- o-end delay
pe o mance o 20 hou s o simula ion un ime is
shown in igu e 5. We can see ha his ime depend
di ec ly on he numbe o hops. Once we inc ease he
numbe o nodes, he numbe o hops also inc eases,
and, consequen ly, he a e age da a packe deli e y
end- o-end delay inc eases oo.

Figu e 6 shows he a e age collision ime o he all
nodes. Collision ime ne e is abo e 0.6 % o all
nodes and 1.5 % o base s a ion.

Da a h oughpu is ob ained applying he equa ion 1

 

"!!#%$'&(
)*!#+"!!#%$'&,!-(/.10
2345!
0
!#
(1)
Wi h a aw bi a e o 4800 bps, we ha e ob ain a
h oughpu o app oxima ely 4655 bps.

The simula ion shows an a e age du y ime o 2.4-
3.6% as depic ed in igu e 7.
6 Conclusion
We ha e p esen ed OLIMPO, a ne wo k simula o , as an
use ul ool o de elop SCADA applica ions o e wi eless
senso ne wo ks. We ha e de eloped a ne wo k p o ocol
ha cons uc s and main ains he spanning ee.
To illus a e he capabili ies o OLIMPO, we ha e
gene a ed se e al expe imen s, p esen ing simula ion e-
sul s showing ne wo k o ma ion ime, da a packe deli -
e y esponse ime, collision ime, h oughpu and a e age
du y ime. We ha e ob ained ele an esul s o SCADA
applica ions.
7 Fu u e wo ks
Fo he p oposed simula o , we a e de eloping new p o-
ocols, such as Low Du y cycle MAC (LD-MAC), slo ed
CSMA-CA channel access, Media ion De ice (MD) p o o-
col and Dis ibu ed MD p o ocol.
Due o he good esul s, ob ained wi h OLIMPO, we
ha e begun o wo k on he implemen a ion o some SCADA
applica ions o e eal senso nodes.
Re e ences
[1] Ins i u e o Elec ical and Elec onics Enginee s.
IEEE 802.11-1999(ISO/IEC 8802-11:1999) IEEE
S anda d o In o ma ion Technology - Telecomuni-
ca ions and In o ma ion Exchange be ween Sys ems -
Local and Me opoli an A ea Ne wo ks - Speci ic Re-
qui emen s - Pa 11: Wi eless LAN Medium Access
Con ol (MAC) and Physical Laye (PHY) Speci ica-
ions. IEEE P ess, New Yo k, USA, 1999.
[2] Jos´e A. Gu i´e ez. IEEE 802.15.4: A de eloping s an-
da d o low-powe , low-cos wi eless pe sonal a ea
ne wo ks. In IEEE Ne wo k, numbe 5 in . 15, pages
12–19, New Yo k, USA, Oc obe 2001. IEEE, IEEE
P ess.
[3] D. Es in e al. Ins umen ing he wo ld wi h wi eless
senso ne wo ks. In Acous ics, Speech, and Signal
P occesing, . 4, pages 2033–2036, New Yo k, USA,
2001. IEEE, IEEE P ess.
[4] Ins i u e o Elec ical and Elec onics Enginee s.
IEEE 802.15.4-2003 IEEE S anda d o In o ma ion
Technology - Telecomunica ions and In o ma ion Ex-
change be ween Sys ems - Local and Me opoli an
A ea Ne wo ks - Speci ic Requi emen s - Pa 15.4:
Wi eless Medium Access Con ol (MAC) and Physi-
cal Laye (PHY) Speci ica ions o Low-Ra e Wi eless
Pe sonal A ea Ne wo ks (LR-WPANs). IEEE P ess,
New Yo k, USA, 2003.
[5] A. Bond. The wa e indus y (au oma ic me e ead-
ing). IEE Colloquium on ’Low Powe Radio and

Me e ing’. IEE, London, UK, (Diges No.1994/060),
1994.
[6] B. Adams and M. Philips. T ends owa ds s anda d
communica ions o me e ing. Nin h In e na ional
Con e ence on Me e ing and Ta i s o Ene gy Sup-
ply. IEE, London, UK, (Con . Publ No.462), 1999.
[7] A.M. Fox. The business case o adio based AMR.
IEE Colloquium on ’Low Powe Radio and Me e -
ing’. IEE, London, UK, (Diges No.1994/060), 1999.
[8] S. Mak and D. Rad o d. Design conside a ions o
implemen a ion o la ge scale au oma ic me e ead-
ing sys ems. IEEE T ansac ions on Powe Deli e y,
10(Diges No.1994/060), Jan 1995.
[9] F. J. Molina, M. G. Go dillo, J. Luque, and J. Ba os.
Radio ne wo k a chi ec u e o au oma ic me e ead-
ing. Con ´
e ence In e na ionale des G andes R´
eseaux
´
Elec iques, K akow POLAND, 1999.
[10] R. Es in, J. Heideman, and S. Kuma . Nex cen-
u y challenges: Scalable coodina ion in senso ne -
wo ks. In ACM/IEEE In e na ional Con e ence on
Mobile Compu ing and Ne wo king, . 5, pages 263–
270, 1999.
[11] S. Tilak, N. Abu-Ghazaleh, and W. Heinzelman. A
axonomy o wi eless mic osenso ne wo k models.
In ACM Mobile Compu ing and Communica ions Re-
iew, MC2R, 2002.
[12] Pa ick Kinney. Zigbee echnology: Wi eless con ol
ha supply wo ks. In Communica ion Design Con-
e ence. IEEE, 2003.
[13] L. Hes e , Y. Hung, O. And ic, A. Allen, and P. Chen.
neuRFon

Ne Fo m. A sel -o ganizing wi eless
senso ne wo k. 2003.
[14] Jian Huang, Yan Huang, Edga Callaway, Qicai Shi,
and Bob 0’Dea. Simula ion o a low du y cycle p o o-
col. In OPNETWORK 2001, Washing on, D.C., Au-
gus 2001.
[15] Yan Huang, Jian Huang, Lance Hes e , An hony
Allen, Oleg And ic, P iscilla Chen, and Bob 0’Dea.
Opne simula ion o a mul i-hop sel -o ganizing wi e-
less senso ne wo k. In OPNETWORK 2002, Wash-
ing on, D.C., Augus 2002.
[16] J. B och, D.A. Mal z, D. B.Johnson, Y.C. Hu, and
J. Je che a. A pe o mance compa ison o mul i-hop
wi eless ad hoc ne wo k ou ing p o ocols. MOBI-
COM, Dallas, TX, Aug 1998.
[17] Edga d H. Callaway J . Wi eless Senso Ne wo k.
AUERBACH PUBLICATIONS, Boca Ra on, USA,
2003.
[18] J. J. Ga cia-Luna-Ace es and M. Spohn. Sou ce- ee
ou ing in wi eless ne wo ks. P oc. IEEE ICNP 99,
7 h n l. Con e ence on Ne wo k P o ocols, To on o,
Canada, Oc 1999.
[19] Cou o and B Aguayo. Pe o mance o mul ihop wi e-
less ne wo ks: Sho es pa h is no enough. P oceed-
ings o he Ho Ne s, P ince on, New Je sey, Oc 2000.
[20] R.W. Heinzelman, A. Chand akasan, and H. Balak -
ishnan. Ene gy-e icien ou ingp o ocols o wi eless
mic osenso ne wo ks. In Hawaii In e na ional Con-
e ence on Sys em Sciences (HICSS ’00), Jan 2000.
[21] F ancisco J. Molina, Julio Ba bancho, and Joaqu´ın
Luque. Applica ion p o ocol o SCADA on a
wi eless me opoli an a ea ne wo k. In C.E. Palau
Sal ado , edi o , Communica ion Sys ems and Ne -
wo ks, pages 227–232, Benalm´adena, Spain, Sep em-
be 2003. IASTED, ACTA P ess.
[22] F ancisco J. Molina, Julio Ba bancho, and Joaqu´ın
Luque. Au oma ed me e eading and SCADA ap-
plica ion o wi eless senso ne wo ks. In Samuel
Pie e, Michel Ba beau, and E angelos K anakis, edi-
o s, Ad-Hoc, Mobile and Wi eless Ne wo ks, numbe
2865 in Lec u es No es in Compu e Science, pages
223–234, Mon ´eal, Canada, Oc obe 2003. ADHOC-
NOW 2003, Sp inge -Ve lag.
[23] H. Zimme mann. OSI e e ence model - he ISO
model o a chi ec u e o open sys ems in e connec-
ion. In IEEE T ans.Commun., numbe 4 in . COM-
28. IEEE, 1980.
[24] h p://www.opne .com/nsnam/ns/.
[25] h p://www.isi.edu.
[26] C. Ulme . h p://use s.ece.ga ech.edu/ g i-
mace/ esea ch/senso simii/. 2000.