A Survey of Vehicular Ad hoc Networks Routing Techniques - Sudan ...

rabidwestvirginiaNetworking and Communications

Oct 26, 2013 (3 years and 11 months ago)

268 views

Internati onal Journal of Innovati on and Appl i ed Studi es

ISSN 2028
-
9324 Vol. X No. X Month 20XX, pp.

XXX
-
XXX

© 20XX Innovati ve Space of Sci enti fi c Research Journal s

http://www.i ssr
-
journal s.org/i jias/


Corresponding Author:
Fi rst Author

1



A S
urvey

of Vehicular Ad
h
oc Networks

Routing
Techniques

Marwa Altayeb

1

and

Imad Mahgoub

2

1

Department of
Computer

Informati on & Networks
,

Col l ege of
Computer

Sci ence &
Informati on
Technol ogy

Sudan Uni versi ty of Sci ence & Technol ogy
,

Khartoum
,
Sudan


2
Department of Computer & El ectri cal Engi neeri ng and Computer Sci ence
,

Col l ege of Engi neeri ng &
Computer

Sci ence,

Fl ori da Atl anti c Uni versi ty,

Fl ori da
,
USA







A
BSTRACT
:

In recent
years
,
the aspect of vehi cul ar ad hoc network (VANET) i s becomi ng an i nteresti ng research area; VANET
i s a mobi l e ad

hoc network consi dered as a speci al case of mobi l e ad hoc network (MANET). Si mi l ar to MANET, VANET i s
characteri zed as autonomous and sel f
-
confi gured wi rel ess network. However, VANET has very dynami c topol ogy, l arge and
vari abl e network si ze, and constrai ned mobi l i ty; these characteri sti cs l ed to the need for effi ci ent routi ng and resource sav
i ng
VANET protocol s, to fi t wi th di fferent VANET e
nvi ronments. These di fferences render tradi ti onal MANET’s protocol s
unsui tabl e for VANET
.
The ai m of thi s work i s to gi ve a survey of the VANETs routi ng mechani sms, thi s paper gi ves an
overvi ew of Vehi cul ar ad hoc networks (VANETs) and the exi sti ng VANET r
outi ng protocol s; mai nl y i t focused on vehi cl e to
vehi cl e (V2V) communi cati on and protocol s. The paper al so represents the general outl i nes and goal s of VANETs, i nvesti gates
di fferent routi ng schemes that have been devel oped for VANETs, as wel l as provi di n
g cl assi fi cati ons of VANET routi ng
protocol s (focusi ng on two cl assificati on forms), and gi ves summari zed compari sons between di fferent cl asses i n the context
of thei r methodol ogi es used, strengths, and l i mi tati ons of each cl ass scheme compared to other cl
asses. Fi nal l y, i t extracts
the current trends and the chal l enges for effi ci ent routi ng mechani sms i n VANETs
.


K
EY
WORDS
:

VANET, Route, Routi ng Prot
ocol s, Topol ogy
-
based, Posi ti on
-
based, V2V
.

1

I
NTRODUCTION

Vehi cul ar Ad hoc networks (VANETs) are a speci al type of mobi l e ad hoc networks; where vehi cl es are si mul ated as
mobi l e nodes.
VANET contai ns two enti ti es: access poi nts and vehi cl es, the access poi nts are fi xed and usual l y connected to
the i nternet, and t
hey coul d parti ci pate as a di stri buti on poi nt for vehi cl es [1]. VANET addresses the wi rel ess communi cati on
between vehi cl es (V2V), and between vehi cl es and i nfrastructure access poi nt (V2I). Vehi cl e to vehi cl e communi cati on (V2V)
has two types of communi ca
ti on: one hop communi cati on (di rect vehi cl e to vehi cl e communi cati on), and mul ti hop
communi cati on (vehi cl e rel
i e
s on other vehi cl es to retransmi t).
VANET al so has speci al characteri stics that di sti ngui sh i t from
other mobi l e ad hoc networks; the most i mpo
rtant character
i sti cs are: hi gh mobi l i ty, sel f
-
organi zati on, di stri buted
communi cati on, road pattern restri cti ons, and no restri cti ons of network si ze [
2
]
-
[4], al l these characteri sti cs made VANETs
envi ronment a chal l engi ng for devel opi ng effi ci ent routi ng

protocol s
.

VANETs appl i cati on
s

types are cl assi fi ed i nto
safety
and

effi ci ency
appl i cati on

[
1
]
,

[
5
]
,

[
6
].
T
here

are many di ffi cul ti es
faci ng VANETs systems desi gn and i mpl ementati on, i ncl udi ng: securi ty, pri vacy, routi ng, connecti vi ty, and qual i ty of serv
i ces.
Thi s paper wi l l focus on routi ng probl em i n vehi cl e to vehi cl e communi cati on (V2V); di scusses some proposed routi ng
sol uti ons, routi ng protocol s cl assi fi cati ons, and i l l ustrates some chal l enges and open i ssues i n VANET routi ng.

Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

2



The mai n goal for rout
i ng protocol i s to provi de opti mal paths between network nodes vi a mi ni mum overhead.
Many
routi ng protocol s have been devel oped for VANETs envi ronment, whi ch can be cl assified i n many ways, accordi ng to di fferent
aspects; such as: protocol s characteri sti cs
, techni ques used, routi ng i nformati on, qual i ty of servi ces, network structures,
routi ng al gori thms, and so on.
Some research papers cl assi fi ed VANETs routi ng protocol s i nto fi ve cl
asses: topol ogy
-
based,
posi ti on
-
based, geocast
-
based, broadcast
, and
cl uster
-
based routi ng protocol s, thi s cl assi fi cati on i s based on the routi ng
protocol s characteri stics and techni ques used [
2
]
,
[
5
]
,
[
7
]. As wel l, other papers cl assified VANETs routi ng protocol s according
to the network structures, i nto three cl asses: hi e
rarchical rout
i ng, fl at routi ng, and posi ti on
-
base routi ng. Moreover, they can
be categori zed i nto two cl asses accordi ng to routi ng strategi es: proacti ve and reacti ve [
8
]. On the other hand other papers
cl assi fi ed them i nto two categori es: geographi c
-
based

and topol ogy
-
based, accordi ng to the routi ng i nformati on used i n
packet forwardi ng [4]. Al so based on qual i ty of servi ces cl assi fi cati on, there are three types of protocol s that deal i ng wi th

network topol ogy (hi erarchi cal, fl at, and posi ti on aware), conce
rni ng wi th route di scovery (reacti ve, proacti ve, hybri d and
predi cti ve), or based on the MAC l ayer i nteracti on [9]. However al l previ ous cl assi fi cati ons di d not concern by transmi ssi on

strategi es cl assi fi cati on (such as uni cast
,
broadcast,
and
mul ti cast)
.

Thi s paper wi l l address two types of cl assi fi cati ons as shown i n
Fi g.

1
; the fi rst one i s the routi ng i nformati on whi ch used
i n packet forwardi ng, i t mai nl y focuses on topol ogy
-
based and
graphi c
-
based routi ng.
And the other cl ass i s the transmi ssi on
strat
egi es, whi ch i s we thought i t has a si gni fi cant i mpact i n protocol desi gn and network performance (i n case of network
overhead, del ay, and packet l oss).



























Fig. 1.

Classification of VANET routing protocols


Topol ogy
-
base
d

Posi ti on
-
base
d


Transmissi on Strategi es

Routi ng Informati on

VANET
Routing Protocols
Classes

Proactive

Reactive

Hybrid

OLSR

Link State

Distance

Vector

DSDV


FSR


AODV

DSR

TORA

DYMO

AODV_DB

AODV+PGB

BAODV

AOMDV

R
-
AOMDV


SD
-
AOMDV

ZRP

ZHLS


DTN

Non DTN


Hybrid

MOV
E

VADD


GeOpps

SPSR

GPCR



RIRP

HLAR


Uni cast


Broadcast



Mul ti
cast

DECA

DV
-
CAST

POCA


DADCQ


Geocast


Cluster


ROVER


Mobicast


COIN

CBDRP

First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

3



The rest of t
he paper i s organi zed as fol l ow
s
: secti on 2 presents cl ass one: the routi ng i nformati on used i n packet
forwardi ng, di scus
s
es topol ogy
-
based and posi ti on
-
based routi ng protocol, and i l l ustrate some rel ated protocol s wi th a bri ef
showi ng to thei r strengths and l i mi tati ons, and compari sons between the di fferent cl ass types. Secti on 3 just l i ke secti on 2,

represents the second cl ass types; transmi ssion strateg
i es, shows vari ous categori es of routi ng protocol s, and di scusses some
rel ated protocol s al ong wi th thei r strengths and l i mi tati ons, also gi ves bri ef compari sons between the di fferent categori es.

Secti on 3 di scus
s
es some research area and open i ssues i n
VANETs. And fi nal secti on summari zes and concl udes thi s paper.

2

R
OUTING INFORMATION U
SED IN PACKET FORWAR
DING

Thi s cl ass i s di vided i nto two subcl ass
es: topol ogy
-
based and posi ti on
-
based routi ng protocol s
.
In topol ogy
-
based routi ng,
each node shoul d
be aware of the network l ayout, al so shoul d abl e to
forward packets usi ng i nformati on about avai l abl e
nodes and l i nks i n the network.
In contrast, posi ti on
-
based routi ng shoul d be aware of the nodes l ocati ons i n the packet
forwardi ng.

2.1

T
OPOLOGY
-
B
ASED
R
OUTIN
G
P
ROTOCOL
:

Topol ogy
-
based routi ng protocol usual l y a tradi ti onal MANET routi ng protocol, i t uses l i nk's i nformati on whi ch stored i n
the routi ng tabl e as a basi s to forward packets from source node to desti nati on; i t commonl y categori zed i nto three
categor
i es (base on underl yi ng archi tecture) [
3
]
,
[
10
]: Proacti ve (peri odi c)
,
Reacti ve (on
-
demand)
and
Hybri d

2.1.1

P
ROACTIVE
R
OUTING
P
ROTOCOLS

Proacti ve protocol s al l ow a network node to use the routi ng tabl e to store routes i nformati on for al l other nodes, each
entry
i n the tabl e contai ns the next hop node used i n the path to the desti nati on, regardl ess of whether the route i s actual l y
needed or not.
The tabl e must be updated frequentl y to refl ect the network topol ogy changes,
and

shoul d be
broadcast
peri odi cal l y to
the

nei ghbors. Thi s scheme may cause more overhead especi al l y i n the hi gh mobi l i ty network. However,
routes to desti nati ons wi l l al ways be avai l abl e when needed [4]. Proacti ve protocol s usual l y depend on shortest path
al gori thms to determi ne whi ch route wi
l l be chosen; they general l y use two routi ng strategi es: Li nk state strateg
y and
di stance vector strategy.

2.1.1.1

D
ESTINATION
S
EQUENCE
D
ISTANCE
V
ECTOR
R
OUTING
(DSDV)

DSDV protocol i t i s an earl i est ad hoc routi ng protocol, i t i mpl ements the di stance vector strate
gy and uses a shortest
path al gori thm to i mpl ement onl y one route to desti nati on whi ch stored i n the routi ng tabl e, each routi ng tabl e contai ns
i nformati on about al l accessi bl e network nodes, as wel l as the total number of hops needed to reach these nodes,

and each
entry i n the routi ng tabl e i s l abel ed wi th a sequence number i ni ti ated by the desti nati on node. To mai ntai n routes rel i abi l i t
y,
each node must peri odi cal l y broadcast i ts routi ng tabl e to i ts nei ghbors. DSDV protocol guarantees the l oop free routs
,
excl udes extra traffi c caused by frequent updates, as wel l as reduces control message overhead, i t al so keeps onl y the
opti mal path to every node, rather than keepi ng mul ti paths whi ch wi l l hel p to reduce the total si ze of routi ng tabl e [
8
].

However, DSD
V i ncreases the overhead i n the l arge network; because of unnecessary updati ng broadcast even i f there i s no
change i n the network topol ogy.
Besi des
that, DSDV don't provi de mul ti routes to desti nati on node [
8
] and has no control
over the network congesti o
n whi ch decreases the routi ng effi ci ency [
11
].
As the resul t of these l i mi tati ons, Randomi zed
DSDV protocol (R
-
DSDV) i s proposed to support congesti on control over DSDV; by mai ntai ni ng nodes randomi zed deci si on
whi ch al l ows each node to make a deci si on whe
ther to forward or di scard a packet. However the R
-
DSDV produces more
overhead compared to the DSDV protocol.

2.1.1.2

O
PTIMIZED
L
INK
S
TATE
R
OUTING
P
ROTOCOL
(OLSR)


OLSR protocol i mpl ement

the l i nk state strategy; i t keeps a routi ng tabl e contai ns i nformati on about

al l possi bl e routes to
network nodes. Once the network topol ogy i s changed each node must send i ts updated

i nformati on to some

sel ecti ve
nodes, whi ch retransmi t thi s i nformati on to i ts other sel ecti ve nodes. The nodes whi ch are not i n the sel ected l i st ca
n just
read and process the packet [10].

Some researchers thought that OLSR has easy procedure whi ch al l ows i t to bui l t
-
i n di fferent operati ng systems, besi des i t
works wel l i n the dynami c topol ogy, al so i t i s general l y sui tabl e for appl i cati ons that requ
i red l ow l atency i n the data
transmi ssi on (l i ke warni ng appl i cati ons) [11]. However, OLSR may cause network congesti on; because of frequent control
packets whi ch sent to handl e topol ogy changes, moreover OLSR i gnore the hi gh resources capabi l i ti es of nodes

(l i ke
Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

4



transmi ssi on range, bandwi dth, di recti onal antenna and so on) [12].
Therefore
, some researchers propose Hi erarchi cal
Opti mi zed Li nk State Routi ng (HOLSR) protocol as enhancement of the OLSR protocol, whi ch decreases routi ng control
overhead i n the l arge si ze networks, al so maxi mi zes the routi ng performance; by the defi ni ng network
hi erarchy archi tecture
wi th mul ti pl e networks [13]. Al so some researchers propose QOLSR as a sol uti on of provi di ng a path such that the avai l abl e
bandwi dth at each node on the path i s not l ess than the requi red bandwi dth. QOLSR consi ders del ay as a second
for path
sel ecti on [12]. These protocol s usual l y provi de average enhancement for the QoS of packets. However, they cause more
compl exi ty, i ncreasi ng packet overhead, and onl y sui tabl e for some l i mi ted appl i cati ons [9].

2.1.1.3

F
ISHEYE STATE ROUTING

(FSR)

In FSR, t
he node peri odi cal l y updates i ts tabl e based on the l atest i nformati on recei ved from nei ghbori ng nodes. The
updati ng of the routi ng tabl e entri es that concern a certai n desti nati on must be broadcast by di fferent frequenci es for
nei ghbors. Tabl e entri es tha
t are further i n the di stance are broadcast wi th l ower frequency than entri es that are nearer, thi s

scheme doesn't guarantee decreasi ng broadcast overhead i n l arge di stances routi ng process.
However, i t coul d be accurate,
i f the packets come cl oser to the
desti nati on [4]
,
[
14
].
The probl em wi th the FSR i s that, the growi ng network si zes wi l l al so
i ncrease the routi ng tabl es, al so i f the topol ogy changes i ncreased, the route to a remote desti nati on becomes i naccurate.
Moreover i f the

desti nati on moves out o
f scope of source node then i t can

no
t di scover the route [
4
]
,
[
15
].

The advantage of proacti ve routi ng protocol s can be abbrevi ated to there i s no need to route di scovery process; because
the route to the desti nati on i s kept i n the background, moreover pr
oacti ve protocol s peri odi cal l y update the routi ng
i nformati on whi ch l ets these protocol s to perform wel l i n l ow mobi l i ty networks. However, they have degraded performance
i n hi ghl y mobi l i ty and densi ty network that when compare them wi th the reacti ve routi
ng protocol s, moreover unused
routes consume the avai l abl e bandwi dth and i ncrease the network overhead [
2
].

Recent studi es show that proacti ve routi ng protocol s (such as OLSR) general l y outperform the reacti ve protocol s i n terms
of network throughput and
end to end del ay [1
6
].
However; there i s no much research i n the proacti ve routi ng protocol s for
VANET compared wi th exi sti ng VANET reacti ve protocol s researches.

2.1.2

R
EACTIVE
R
OUTING
P
ROTOCOLS

Reacti ve routi ng protocol s (al so cal l ed on
-
demand) reduce the network overhead; by mai ntai ni ng routes onl y when
needed, that the source node starts a route di scovery process, i f i t needs a non exi sti ng route to a desti nati on, i t does thi
s
process by fl oodi
ng the network by a route request message.
After the message reaches the desti nati on node (or to the node
whi ch has a route to the desti nati on), thi s node wi l l send a route repl y message back to the source node usi ng uni cast
communi cati on [1
7
].

Reacti ve r
outi ng protocol s are appl i cabl e to the l arge si ze of the mobi l e ad hoc networks whi ch are
hi ghl y mobi l i ty and frequent topol ogy changes [1
8
]. Many reacti ve routi ng protocol s have been devel oped
,
the

fol l owi ng
secti ons wi l l i l l ustrate characteri sti c of
some

reacti ve
protocol s, as wel l as i l l ustrates the exi sti ng enhancement protocol s.

2.1.2.1

A
D HOC ON
-
DEMAND DISTANCE VECT
OR
(AODV)

AODV routi ng protocol i s proposed for mobi l e ad hoc network, i t has been eval uated i n several researches and shows
good resul ts compared

to rel ated routi ng protocol s; so i t has a good documentati on [1
9
].
AODV offers l ow network overhead
by reduci ng messages fl oodi ng i n the network; that when compared to proacti ve routi ng protocol s,
besi des

reduci ng the
requi rement of memory si ze; by mi ni mi
zi ng the routi ng tabl es whi ch keep onl y entri es for recent acti ve routes, al so keeps
next hop for a route rather than the whol e route.
It al so provi des dynami call y updates for adapti ng the route condi ti ons and
el i mi nates l oopi ng i n routes; by usi ng desti na
ti on sequence numbers. So AODV i s fl exi bl e to hi ghl y dynami c network top
ol ogy
and l arge
-
scal e network [20
].
However, i t causes l arge del ays i n a route di scovery, al so route fai l ure may requi re a new route
di scovery whi ch produces addi ti onal del ays that dec
rease the data transmi ssi on rate and i ncrease the network overhead
[1
7
].
Moreover, the redundant broadcasts wi thout control wi l l consume extra bandwi dth (broadcast storm probl em), thi s
probl em grows as the number of network nodes i ncreases, that besi des co
l l i si ons whi ch l ead to packet l ost probl em [1
9
].
There are several protocol s have been proposed to enhance AODV protoc
ol; by decreasi ng i ts probl ems.

2.1.2.1.1

A
D
-
HOC
O
N
-
DEMAND
M
ULTIPATH
D
ISTANCE
V
ECTOR
R
OUTING
(AOMDV)


AOMDV protocol i s a mul ti path on
-
demand protocol s comes as an extensi on of the AODV protocol, i t di scovers many
paths from source to desti nati on i n a si ngl e route di scovery process. Mul ti path On
-
demand protocol s perform better than
si ngl e path protocol s
especi al ly i n reduci ng route di scovery retransmi ssion. It stores mul ti paths to desti nati on usi ng a si ngl e
route di scovery process; therefore no need to di scover a new route i f onl y a si ngl e path i s fai l i ng; i t's easy use any one of

the
First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

5



exi sti ng redundancy

paths.

In mul ti path protocol s a new route di scovery i s requi red i f and onl y i f al l repl i cated paths to the
desti nati on are fai l i ng.
In contrast, di fferent thi ng happens i n the si ngl e path protocol s whi ch establ i sh a new route di scovery
every ti me a si ngl
e path from the source to the desti nati on i s fai l i ng.
Thi s made mul ti path protocol s have a better
performance i n term of uni nterrupted communi cati ons for the packet transmi ssi on, and provi de l ower overhead; due to
decreasi ng frequent route di scovery trans
mi ssi on.

AOMDV protocol i s an enhancement of AODV protocol, i t uses same
control messages used i n AODV, i t just adds extra fi el ds for AODV routi ng control messages; that to reduce the overhead
occurri ng by di scoveri ng mul ti pl e paths.
Moreover di scoveri ng m
ul ti pl e paths doesn't i ncrease the del ay of the route
di scovery process; because the l atency of a route di scovery i s defi ned by the total source wai ti ng ti me, before the source
recei ved the fi rst path. AOMDV keeps al l avai l abl e paths i n the routi ng tabl e,
then the source chose one of the stored paths;
the preferred path wi l l be the fi rst establ i shed one

[
21
]
.


2.1.2.1.1.1

E
NHANCING
AOMDV

R
OUTING
P
ROTOCOL FOR
V2V

C
OMMUNICATION
(SD
-
AOMDV
)


SD
-
AOMDV i s proposed as an enhancement of the AOMDV protocol; to deal wi th VANET c
haracteri sti cs. SD
-
AOMDV
appends new factors (speed and di recti on) to the hop count fi el d to determi ne the next hop duri ng the rout di scovery
process.
The next hop node i s an i ntermedi ate node, sel ected based on two factors: i ntermedi ate nodes that can mo
ve i n
the same di recti on of the source and the desti nati on, or the same di recti on of the source, or the same di recti on of the
desti nati on, and i ts speed i s equal or near to the average speed of the source and desti nati on. The protocol merges these
two fact
ors wi th a hop count fi el d to choose a route [22].

2.1.2.1.1.2

A

C
ROSS
-
L
AYER
AOMDV

R
OUTING
P
ROTOCOL FOR
V2V

C
OMMUNICATION IN
U
RBAN
VANET

(R
-
AOMDV)

R
-
AOMDV i s a protocol bui l t on AOMDV; i t uses a method whi ch merges transmi ssi on count and hop counts at the MAC
l ayer, t
aki ng i nto account mi ni mi zi ng del ay and performance of i ntermedi ate l i nks. In the route di scovery process, R
-
AOMDV
i s si mi l ar to AOMDV; i t depends on route request and route repl y control packets. Thi s protocol adds two fi el ds to route
repl ay message fi el d
s; to compute qual i ty of the whol e path, one of these fi el ds i s the maxi mum retransmi ssi on count (MRC)
whi ch i s computed by the MAC l ayer, and the other i s the total hop count whi ch i s computed by the network l ayer. In R
-
AOMDV, a source node sends a route
di scovery packet when i t hasn't a path to the requi red desti nati on, stored i n i ts route
tabl e. When a route repl ay packet recei ved to the source, an i ntermedi ate node updates i ts retransmi ssi on count val ue; i n
case i f i t was greater than the current MRC. S
o, when the route repl ay packet arri ves to the source, the source can i denti fy
whi ch path has maxi mum MRC. R
-
AOMDV protocol i nheri ts al l good properti es of mul ti path routi ng protocol s, such as
reduci ng rebroadcast route di scovery. Thi s protocol shows bett
er performance than AOMDV i n both rural and ci ty vehi cul ar
networks [22]
-
[23], as wel l as enhances the routi ng operati ons by getti ng i nformati on about route's qual i ty based on the
nei ghbors IP addresses.
However, the techni que based on IP addresses i s not
conveni ent for VANET; because i t sends the
packets to nodes IPs, even though they change thei r l ocati ons; i n thi s case the source node must search for a new
i ntermedi ate forwardi ng node; and as a resul t thi s may l ead to i ncrease the packet del ays and packe
t l oss.
Thi s probl em has
l argel y appeared i n ci ty vehi cul ar networks whi ch have mul ti pl e paths, several i ntersecti ons, di fferent node densi ty, and hi g
h
congesti on [23].

2.1.2.1.2

T
HE
DYMO

R
OUTING
P
ROTOCOL IN
VANET

S
CENARIOS

Dynami c MANET On
-
demand (DYMO) protocol
i s a reacti ve mul ti hop routi ng protocol. Li ke AODV protocol, i t uses the
sequence numbers to provi de l oop free routes; i t al so has two essenti al processes: Route Di scovery and Route Mai ntenance.

DYMO i s di fferent from AODV protocol i n other characteri sti c
; that i n DYMO a new route request process has to mai ntai n
i nformati on about al l i ntermedi ate nodes, however i n AODV, i t just col l ects i nformati on about the desti nati on node and the
next hop, moreover i n DYMO, every node whi ch parti ci pates i n a recent rout
e di scovery process shoul d col l ect i nformati on
about a requested node and al l other i ntermedi ate nodes i n the new path. Parti cul arl y at hi gher node densi ti es, whi ch
commonl y occurred i n VANETs, routi ng protocol s and transport protocol s may i ncrease the net
work overhead. When
establ i shi ng a new route i s requi red i n the congested network, the use of the si mpl e retry mechani sm wi l l onl y cause
furthered
congesti on [24].

2.1.2.1.3

AODV_DB

It wi l l be a chal l enge i n VANET i f a route was fai l i ng; accordi ng to thi s, AODV prot
ocol spent a l ong ti me i n di scoveri ng a
new route; by sendi ng a route request and route repl ay messages. However, VANET topol ogy i s hi ghl y dynami c change,
whi ch requi res an establ i shment of a new route redi scovery. The route cachi ng scheme that used by the

on
-
demand protocol
has a bad performance i n the hi gh mobi l i ty envi ronment. In thi s case the usi ng of fl oodi ng mechani sm i s the more sui tabl e
Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

6



for routes mai ntenance; however i t produces many redundant messages and network overhead.

AODV_BD di scovers and
ma
i ntai ns a route by broadcasti ng the data packet i nstead of the route request packet. The data packet header contai ns the
reverse path, so i ntermedi ate nodes wi l l store the reverse path and rebroadcast the data packet, fi nal l y the desti nati on wi l l

recei ve t
he data packet and at the same ti me send the route repl ay to the source.
Thi s approach wi l l reduce the ti me of route
setup packet transmi ssi on [2
5
]. However i t l eads to heavy network overhead by fl oodi ng data packet broadcast. The probl em
ari ses i f there i
s no route to desti nati on.

2.1.2.1.4

A
D HOC ON
-
DEMAND DISTANCE VECT
OR
P
REFERRED
G
ROUP
B
ROADCASTING
(AODV+PGB)


Thi s protocol enhances the AODV protocol by Preferred Group Broadcasti ng (PGB) al gori thm, thi s al gori thm ai ms to
reduce control message overhead i n addi ti on to offer routes avai l abi l i ty whi ch i s an i mportant feature i n VANET
envi ronment, as the reduci ng ro
uti ng overhead i s a si gni fi cant i ssue i n ad hoc networks, al so the routes consi stency i s a
desi rabl e i ssue i n fast movi ng envi ronment. There are many i ssues that cri ti cally decreasing ad hoc network
performance

can
be abbrevi ated i n

[19]
:



The probl em of hi
dden termi nal whi ch ari ses i f the si gnal from the source to the desti nati on i s weaker; thi s makes
easy to i nterrupt the communi cati on between two nodes by a hi dden termi nal.



No parti cul ar scheme i s used to sel ect i ntermedi ate hops. A l arge number of hops i
nvol ve the short di stance
sel ecti on; however the l i nk can si mpl y be fai l i f one of the i ntermedi ate nodes goes out of the range, otherwi se the
weak si gnal may be changed.




The l arger numbers of errors may reduce the qual i ty of l i nks; whi ch l ead to decreasi
ng network throughput.



Al so i f the data transmi ssi on rate i s adapted accordi ng to the network congesti on, i t coul d be affected by the l arge
data error rate decreases the data transmi ssi on rate and may cause a bottl eneck i n the current node.

PGB tri ed to d
eal wi th al l these i ssues vi a permi ts some parti cul ar nodes to re
-
broadcast a route request packet.
However, i f the node that al l owed rebroadcasti ng the route request i s not the nearest node to the desti nati on, then the
route di scovery coul d be l onger than

i t shoul d. Al so broadcast can be hal ted i f there i s no speci fi c node whi ch had a
rebroadcast permi t (case i n l i ght networks). Moreover packet dupl i cati on may occur i f any two nodes rebroadcast the same
packet at the same ti me.

2.1.2.1.5

T
HE
B
US
A
D
H
OC
O
N
-
DEMAND
D
ISTANCE
V
ECTOR
(BAODV)

BAODV focused on extendi ng the AODV protocol to make i t more sui tabl e for VANETs, i t has been desi gned to sel ect a
route wi th the mi ni mum number of vehi cl es (buses), and i t mi ni mi zed end
-
to
-
end packet del ay i n the network. The proto
col
al so al l ows dri vers to avoi d congested roads whi ch contai n a l arge number of buses.
The AODV protocol i s not an effi ci ent
VANET protocol; i t has a l i mi tati on on when i t works for VANET; whi ch mai nl y shows i n the frequent routi ng tabl e updates.
The mai n

i dea i n the desi gni ng BAODV i s to overcome AODV protocol l i mi tati ons that made BAODV works on sensi ng of the
vehi cl e characteri stics and behavi ors. In the AODV protocol, the update occurs accordi ng to two i mportant factors: the newer
sequence number and t
he l ower hop count. But BAODV consi ders the l ower hop count factor al one (mean l ower number of
vehi cl es) i s i nsuffi ci ent for many VANET envi ronments, so i t uses addi ti onal factors l i ke vehi cl e type and behavi or to achi ev
e
better routi ng performance. The BA
ODV protocol modi fi ed ADOV messages: RREP and RREQ; by addi ng new fi el ds defi nes
the number of vehi cl es (bus, car, trucks, Trai l ers and so on) i n the route by addi ng new fi el ds i n order to cal cul ate the num
ber
of buses on the route. Thi s i nformati on shoul d

be stored i n the routi ng tabl e; thus i t requi red modi fyi ng the routi ng tabl e
al so. The si mul ati on resul ts show the BAODV protocol performed better than the tradi ti onal AODV i n terms of end
-
to
-
end
packet del ay i n ci ty areas whi ch i s a si gni ficant factor i n

VANETs, parti cul arly for warni ng messages. The protocol al so enabl es
dri vers to choose preferabl e routes that hel p them i n avoi di ng congesti on

[26]
. However, there i s a shortcomi ng of BAODV
document
ati on
, i t di dn't i l l iterate careful l y the modi fi cati on
method used i n the routi ng tabl e and what i s the al gori thm used
to sel ect a sui tabl e route. Al so the protocol focused on sol vi ng end
-
to
-
end del ay, but negl ected the frequent changeabl e
topol ogy, so the sel ecti on of routes based on the l ess hop count may no
t work properl y i n l ong di stance between nodes,
because i f any i ntermedi ate node moved out of the range, the route wi l l break, al so i t may cause the probl em of weak si gnal
.

2.1.2.2

D
YNAMIC
S
OURCE
R
OUTING PROTOCOL
(DSR)

DSR protocol ai ms to provi de a hi ghl y reacti v
e routi ng process; by i mpl ementi ng a routi ng mechani sm wi th an extremel y
l ow overhead and fast reacti on to the frequent network changes, to guarantee successful data packet del i very regardl ess of
network changes. DSR i s a mul ti hop protocol; i t decreases t
he network overhead by reduci ng peri odi c messages. Thi s
protocol has two mai n processes: route di scovery and route Mai ntenance. In the route di scovery, when a source node needs
an unavai l able route, i t i ni ti ally broadcasts a route request message. Al l i nte
rmedi ate nodes whi ch recei ved thi s message wi l l
First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

7



rebroadcast i t, except i f i t was the desti nati on node or i t has a route to the desti nati on; i n thi s case the node wi l l send a

route
repl ay message back to the source, l ater the recei ved route i s cashed i n the

source routi ng tabl e for future use. If a route i s
fai l i ng, the source node wi l l be i nformed by a route error message.

In DSR protocol, every data packet contai ns a compl ete
l i st of the i ntermedi ate nodes; so the source node shoul d del ete the fai l ed rout
e from i ts cache, and i f i t stores other
successful route to that desti nati on i n i ts cache, i t wi l l exchange the fai l ed one by the other successful route. But i f ther
e i s no
al ternati ve route, i t wi l l i niti ate a new route di scovery process [2
7
]. The benefi
t of DSR protocol i s cl earl y shown i n a network
wi th l ow mobi l i ty; because i t can use the al ternati ve route before starts a new process for route di scovery. However,
the
mul ti routes may l ead to addi ti onal routi ng overheads by addi ng al l route i nformati on
to every data packet, besi des, as the
network span l arger di stance and i ncl udi ng more nodes, the overhead wi l l frequentl y i ncrease and as resul t network
performance wi l l be degraded [28].

2.1.2.3

T
EMPORALLY
O
RDERED
R
OUTING
A
LGORITHM
(TORA)

TORA i s a di stri buted ro
uti ng protocol usi ng mul ti hop routes; i t i s desi gned to reduce the communi cati on overhead
rel ated to adapti ng frequent network changes. Thi s protocol does not i mpl ement a shortest path al gori thm; thus the routi ng
structure does not represent a di stance.
T
ORA constructs a di rected graph whi ch contai ns the source node as the tree root.
Packets shoul d be runni ng from hi gher nodes to l ower nodes i n the tree. Once a node broadcasts a packet to a parti cul ar
desti nati on, i ts nei ghbor wi l l broadcast a route repl ay

i f i t has a downward l i nk to the desti nati on, i f not, i t just drops the
packet. TORA ensures mul ti path l oop free routi ng; si nce the packet al ways fl ows downward to the desti nati on and don't
fl ow upward back to the sendi ng node

[2
9
]. The advantages of TOR
A are that i t offers a route to every node i n the network,
and reduces the control messages broadcast. However, i t causes routi ng overhead i n mai ntai ni ng routes to al l network
nodes, especi al l y i n hi ghl y dynami c VANETs [4]
,

[1
5
].

2.1.3

H
YBRID
R
OUTING
P
ROTOCOLS

Hybri d protocol i s a mi xture of both proacti ve and reacti ve protocol s; i t ai ms to mi ni mi ze the proacti ve routi ng protocol
control overhead and reduce the del ay of the route di scovery process wi thi n on
-
demand routi ng protocol s. Usual l y the
hybri d protocol d
i vi des the network to many zones to provi de more rel i abi l i ty for route di scovery and mai ntenance
processes. Each node di vi des the network i nto two regi ons: i nsi de and outsi de regi ons; i t uses a proacti ve routi ng mechani sm
to mai ntai n routes to i nsi de regi o
n nodes and usi ng a route di scovery mechani sm to reach the outsi de regi on nodes [
3
].

2.1.3.1

Z
ONE
R
OUTING
P
ROTOCOL
(ZRP)

ZRP i s the fi rst protocol devel oped as a hybri d routi ng protocol, i t al l ows a network node to di vi de the network i nto zones
accordi ng to many factors; l i ke: power of transmi ssi on, si gnal strength, speed and many other factors. The area i nsi de the
zone i s t
he routi ng range area for the node and vi ce versa for outsi de zone. ZRP uses the reacti ve routi ng schemes for outsi de
the zone and the proacti ve routi ng schemes for i nsi de the zone; wi th a vi ew to keep the l atest route i nformati on wi thi n the
i nsi de zone. I
n the l ocal i nsi de the zone, the source node uses a proacti ve cached routi ng tabl e to i ni ti ate a route to a
desti nati on, whi ch can be hel ped i n transmi tti ng packets di rectl y wi thout del ay. ZRP uses i ndependent protocol s i nsi de and
outsi de the zone; i t may
use any exi sti ng proacti ve and reacti ve routi ng protocol s. For outsi de zone, the ZRP reacti vel y
di scover a route; that the source node transmi ts a route request
packet to the border nodes
of i ts routi ng
zone;

the packet
i ncl udes a uni que sequence number, the source address and the desti nati on address. When the border node recei ves a route
request packet, i t l ooks for the desti nati on wi thi n i ts i nsi de zone. If the desti nati on i s found, i t sends a route repl y on

reverse
path to the source node; el se i f i t doesn't fi nd the desti nati on i n i ts l ocal zone, the border node adds i ts address to the r
oute
request packet and forwards i t to i ts own border nodes. After the source recei ved a repl y, i t stores the path i ncl ude
d i n the
route repl y packet to use i t for data transmi ssion to the desti nati on [
30
]. The weakness of ZRP protocol i s that i t performs l i ke
a pure proacti ve protocol parti cul arly for l arge si ze zones; however for smal l zones i t performs si mi l ar to a reacti v
e protocol
[1
7
]. Thus ZRP protocol i s not appl i cabl e for l arge si ze VANET wi th hi ghl y dynami c topol ogy and

frequentl y change
envi ronment.

2.1.3.2

Z
ONE
-
BASED HIERARCHICAL L
INK STATE
(ZHLS)

ZHLS protocol di vi des the network i nto non overl appi ng zones; every network
node has i ts own ID and a zone ID, whi ch i s
measured by a GPS. There are two l evel s for structural topol ogy: zone l evel topol ogy and node l evel topol ogy. In ZHLS there
i s no posi ti on admi ni strator or cl uster head are used to manage the data communi cati on;
that means there i s no traffi c
bottl eneck. Besi des that the ZHLS reduces the transmi ssi on overheads when compared i t wi th the reacti ve protocol s. ZHLS
broadcast scheme showed l ower overhead compared to the fl oodi ng scheme i n pure reacti ve protocol s. Al so
i n ZHLS, the
Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

8



routes i s fl exi bl e to the dynami c topol ogy because i t requi red onl y the zone ID and the node ID of the desti nati on node for
routi ng; that means there i s no need to search for the l ocati on, i f the desti nati on node does not move to another zone.

The
shortcomi ng of ZHLS, i t needs a stati c zone map i nto each node, and thi s may not be suffi ci ent for a network wi th dynami c
zone edges. Moreover, i t i s not appropri ate for hi ghl y dynami c topol ogi es [1
7
].

General l y, the hybri d routi ng protocol s have a hi
gher scal ability than pure proacti ve and pure reacti ve protocol s; because
of reduci ng the number of rebroadcast messages whi ch achi eved by al l owi ng network nodes to work together and the most
appropri ate nodes are used to setup a route [1
7
]. However, pure
proacti ve and pure reacti ve routi ng protocol s coul d be
more sui tabl e to some hi ghl y dynami c l evel i n a network envi ronment.

O
THER TOPOLOGY
-
BASED PROTOCOLS

Some researchers [
31
] assumed i t’s more effi ci ent to devel op a routi ng protocol based on the topol ogy

of roads; that
mean road to road transmi ssi on rather than the conventi onal node to node routi ng scheme; they justi fy that because of two
reasons: the vehi cl es hi ghl y mobi l i ty and the data del i very i s constrai ned by road pattern. However, the chal l engi ng w
i l l be
the transmi ssi on of packets at i ntersecti ons. So they proposed Buffer and Swi tch (BAS) protocol whi ch al l ow each road to
store packets al ong wi th many transmi tti ng copi es to offer addi ti onal chances for a packet to swi tchi ng at i ntersecti ons.
Di ffer
ent from conventi onal protocol s i n VANETs, BAS i s a bi di recti onal dupl icate transmi ssi on. Al so, BAS control l ed dupl i cate
transmi ssi on by spati otemporal l y, whi ch l eads to si gni fi cantl y mi ni mum cost compared by other fl oodi ng protocol s. BAS
performance shows

better than the tradi ti onal protocol s, mai nl y for network wi th l i mi ted resources. However, i t may cause
transmi ssi on del ay, and packet l oss due to packet expi red.

2.2

P
OSITION
-
B
ASED
R
OUTING
P
ROTOCOL

Posi ti on or geographi c routi ng protocol i s based on the pos
i ti onal i nformati on i n routi ng process; where
the source
sends a packet to the desti nati on usi ng i ts geographi c posi ti on rather than usi ng the network address
. Thi s protocol requi red
each node i s abl e to de
ci de
i ts l ocati on and the l ocati on of i ts nei ghbor
s through the Geographi c Posi ti on System (GPS)
assi stance. The node i denti fi es i ts nei ghbor as a node that l ocated i nsi de the node’s radi o range. When the source need to
send a packet, i t usual l y stores the posi ti on of the desti nati on i n the packet header

whi ch

wi l l hel p i n forwardi ng the packet
to the desti nati on wi thout needs to route di scovery, route mai ntenance, or even awareness of the network topol ogy [
3
]
,
[4].

Thus the posi ti on routi ng protocol s are consi dered to be more stabl e and sui tabl e for VANE
T wi th a hi gh mobi l i ty
en
vi ronment, compared to topol ogy
-
base
d

routi ng protocol s.
Geographi c routi ng protocol s commonl y cl assi fi ed i nto three
cl asses: Del ay Tol erant Network (DTN) Protocol s, Non Del ay Tol erant Network (Non DTN) Protocol s and hybri d [4
]
.

2.2.1

D
ELAY
T
OLERANT
N
ETWORK
(DTN)

P
ROTOCOLS

DTN i s a wi rel ess network desi gned to perform effi ci entl y i n networks wi th some characteri sti cs; l i ke frequent
di sconnecti on communi cati on, l arge scale, l ong unavoi dabl e del ays, l i mi ted bandwi dth, power

constrai nts a
nd hi gh bi t faul t
rates [1
5
]. In thi s network, al l nodes hel p each other to forward packets (store and forward scheme). These nodes may have
a l i mi ted transmi ssion range; so packets transmi ssi on wi l l take l arge del ays. Commonl y, the DTN node i s a mobi l e no
de, so i t
establ i shes routes to other nodes when they reach i ts transmi ssi on range. In DTN protocol, there i s no guarantee of
unbroken end to end connecti vi ty, so the packets may be cached for a ti me at i ntermedi ate nodes
[4]
,
[1
4
]
,
[
3
]
.

To desi gn of a
rou
ti ng protocol for DTN network wi th these characteri sti cs i s a si gni fi cant probl em. Thi s secti on, revi ew many DTN routi ng
protocol s that fal l under thi s category.

2.2.1.1

M
OTION
V
ECTOR
R
OUTING
A
LGORITHM
(MOVE)


MOVE al gori thm i s desi gned for l i ght networks, especi a
l l y for road si de vehi cl e communi cati on. Thi s protocol assumes
that each node has gl obal l ocati ons i nformati on, that's besi de the knowl edge of a mobi l e router speed and i ts nei ghbori ng
nodes vel oci ty. From thi s i nformati on the node can esti mate the nodes w
hi ch are the cl osest di stance to the desti nati on [1
4
].
In thi s protocol each node

regul arl y

broadcasts a

HELLO

message; and i ts nei ghbor

repl ays by a RESPONSE

message
; by thi s
repl ayed message the node wi l l know i ts nei ghbors and thei r l ocati ons. Gi ven

thi
s i nformati on, the node

can esti mate the
shortest di stance to desti nati on, i n that case the

node

deci des how to

forward

the

message

accordi ng to the i nformati on
about nodes whi ch are currentl y

l ocated nearby the desti nati on.

MOVE protocol uses

l ess

memory
si ze compared wi th

Non
DTN
posi ti on
-
based routi ng; i t al so has a hi gher data transmi ssi on rate i n l i ght envi ronments [
30
]. However, Non DTN
posi ti on
-
based routi ng coul d have better performance

onl y i f the routes are stabl e and consi stent [
3
].

2.2.1.2

VADD:

V
EHICLE
-
A
SSISTED
D
ATA
D
ELIVERY IN
V
EHICULAR
A
D
H
OC
N
ETWORKS

First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

9



VADD protocol desi gned to handl e frequentl y di sconnected vehi cul ar networks and hi ghl y mobi l i ty probl ems. It
i mpl ements the store and forward scheme; whi l e a node i s movi ng i t stores the packet, unti l a
new node arri ves to i ts zone
range, and then i t forwards the stored packet to thi s new node. Thi s protocol predi cts node mobi l i ty based on two factors:
network traffi c and route type; that hel p a node to di scover the next forwardi ng node. VADD protocol s us
ual l y del i ver the
packet to the path wi th the l east transmi ssi on del ay; fol l owi ng three mai n pri nci pl es

[4]
,
[14]
:



Conti nue use the avai l abl e wi rel ess channel



Del i ver the packet to the hi gher speed node i n the route to carry i t



VANET i s a hi gh mobi l i ty
envi ronment, so i t's di ffi cul t to esti mate packet del i very by a predefi ned opti mal path,
whi ch may l ead to frequent di scover a new opti mal path to transmi t a packet.

To break the routi ng l oop, each node adds i nformati on about i ts former hop/hops before for
wardi ng the packet,
contai ni ng i ts own i nformati on as a former hop. Once the packet recei ved to a node, i t l ooks at the previ ous hops
i nformati on to avoi d forwards the packet to the previ ous hops and try to fi nd other avai l abl e hop; so that may avoi d the
r
outi ng l oop probl em. To forward a packet, VADD i mpl ements four di fferent schemes

[4]
,
[14]
:



Locati on Fi rst Probe (L
-
VADD): i t used to del i ver the packet to the cl osest node to the desti nati on wi thout
consi derati on of the movement di recti on. The drawback i
n thi s scheme the occurri ng of the routi ng l oop.



Di recti on Fi rst Probe (D
-
VADD): the sel ecti on of the next hop

i s

based on the node has the same movement
di recti on as the
desti nati on, whi ch hel ps

i n avoi di ng the route l oop.



Mul ti
-
Path Di recti on Fi rst i s the Probe VADD (MD
-
VADD): i t provi des a mul ti path rather than one path; however, i t
consumes the bandwi dth by redundancy packets.



Hybri d Probe VADD (H
-
VADD): i t i s a hybri d scheme that takes the advantages of L
-
VADD and D
-
VA
DD, to del i ver a
packet, i t i ni ti al l y uses the L
-
VADD; but i f a route l oop i s i denti fi ed, i t changes to D
-
VADD. As a resul t thi s scheme
performs better than pure L
-
VADD and D
-
VADD.

2.2.1.3

G
EOGRAPHICAL

O
PPORTUNISTIC
R
OUTING
(G
E
O
PPS
)


GeOpps

i s a forwardi ng protoco
l uses the avai l abl e navi gati on system i n col l ecti ng i nformati on about geographi cal
posi ti on; thi s i nformati on i s used to sel ect vehi cl es that are cl osest to a certai n desti nati on.
The protocol uses store and
forward techni que, i t works just l i ke the Move
and Non DTN protocol s but i t uses navi gati on system to provi de effi ci ent
packet del i very.

In the GeoOpps, to send a packet from the source to the desti nati on, there are three mai n steps used to
sel ect the next hop of the i ntermedi ate nodes

[3]
,
[5]
:



Each n
ei ghbor
i ng

node at the esti mated routes cal cul ates the future cl osest poi nt to the desti nati on whi ch i t wi l l
reach soon.



Each nei ghbor node then cal cul ates esti mated shortest del ay ti me to reach the speci fi ed packet's desti nati on.



Use the esti mated shortes
t ti me cal cul ated by each nei ghbor node; that any node esti mated to be cl oser to the
desti nati on i n l owest del ay ti me, shoul d be sel ected to become the next hop carri er to transmi t the packet faster to
the speci fi ed desti nati on.

The protocol concerned some

cases whi ch affect i ts effi ci ency

[3]
,
[5]
:



The node i gnores the esti mated cal cul ated route and fol l ows other di fferent path; i n thi s case the system wi l l
forward the hol di ng packet to any nei ghbor node.



The node stops i ts movement (swi tch off the engi ne

or l ong pause ti me); i n thi s case i ts packets shoul d be forward
to another nei ghbori ng node.


The benefi t of GeOpps does not requi re al l nodes to cal cul ate the routes; and GeOpps

transmi ssi on rate depends onl y on
the route topol ogy and the mobi l i ty of the
nodes.
However, i t has some compl exi ti es i n cal cul ati ng del ay ti me dependi ng on
a navi gati on system measurement.

2.2.2

NON
D
ELAY
T
OLERANT
N
ETWORK
(N
ON
DTN)

P
ROTOCOLS

The non
-
DTN proto
col s are geographi c routi ng protocol s, but i t does not consi der a di s
-
connecti vi ty i ssue; i t assumes
there are al ways a number of nodes to achi eve the successful communi cati on; so, thi s protocol i s onl y sui tabl e for hi gh
densi ty network. In these protocol s
, the node forwards i ts packet to the cl osest nei ghbor to the desti nati on, but thi s
approach may be unsuccessful i f there i s no cl osest nei ghbor to the desti nati on rather than the current node i tsel f. Many
non
-
DTN routi ng protocol s handl e thi s fai l ure; by
di fferent strategi es wi l l be shown

i n the fol l owi ng secti ons [1].

Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

10



2.2.2.1

G
REEDY
P
ERIMETER
S
TATELESS
R
OUTING
(GPSR)


GPSR i s a famous greedy routi ng protocol i n VANETs. In thi s protocol, each node forwards packets to other i ntermedi ate
nodes that are constantl y ne
arer to the packet's desti nati on (greedy forward), unti l the packet reaches i ts fi nal desti nati on. If
there i s no nei ghbori ng node cl ose to the desti nati on, i t uses peri meter forwardi ng to deci de to whi ch node i t wi l l del i ver t
he
packet. GPSR i s a statel es
s protocol that keeps i nformati on of i ts fi rst hop nei ghbor's posi ti ons, whi ch coul d i ncrease protocol
scal abi l i ty more than shortest path ad hoc routi ng protocol s. Another advantage i s the dynami c forwardi ng packet deci si on
[
3
]. However, GPSR coul d face a

l i nk fai l ure due to the hi gh mobi l i ty network and frequent topol ogy changes (i t hol ds ol d
posi ti on i nformati on). Thi s probl em can be handl ed by peri meter forwardi ng, but i t may cause hi gh packet l oss and more
l atency ti me due to the l arge number of hops i
n peri meter forwardi ng mode. Moreover, i f the desti nati on node moves to a
new l ocati on, i ts i nformati on whi ch embedded i n the packet header wi l l never be updated [
2
].

2.2.2.2

G
REEDY
P
ERIMETER
C
OORDINATOR
R
OUTING
(GPCR)

GPCR protocol i s desi gned to be sui tabl e for
the
hi gh mobi l i ty envi ronments (as i n ci ty) based on the greedy forwardi ng
techni que; thi s techni que ai ms to forward the packet to a nei ghbor node whi ch i s cl osest to the l ocati on of the desti nati on.
Each node has to be aware of i ts l ocati on gotten by a na
vi gati on system, i t knows i ts nei ghbor by peri odi c beaconi ng, and the
posi ti on of the desti nati on i s obtai ned from the l ocati on servi ce. When a node forwards a packet, the packet wi l l be spread
over the road unti l i t reaches the next i ntersecti on. The mai n
tenance process covers two components: deci si on maki ng, to
deci de whi ch i ntermedi ate node the packet wi l l be passed on the i ntersecti on (a coordi nator node sel ecti on), and forwardi ng
the packet to the next i ntersecti on. The coordi nator node deci des to whi c
h route the packet wi l l be forwarded. But i f no
coordi nator node found i n the route, the packet wi l l be forwarded to furthest node [
10
]. GPCR does not need any gl obal
i nformati on; however i t i s based on the connecti vi ty of the desti nati on node and the dens
i ty of the next roads, i t cou
l d not
connect the desti nati on
i f

the node densi ty i s l ow, whi ch wi l l i ncrease the transmi ssi on del ay [32].

2.2.2.3

R
ELIABILITY
-
I
MPROVING
P
OSITION
-
BASED
R
OUTING
(RIRP)


RIPR i s a posi ti on
-
base
d

routi ng al gori thm desi gned for VANETs, i t

ai ms to sol ve the probl ems of
l i nks
fai l ures that found
i n a posi ti on
-
based routi ng; whi ch appear

due to stori ng
ol d i nformati on about a stal e i ntermedi ate node
. RIPR predi cts the
vehi cl e speeds and thei r movi ng di recti ons, as wel l as esti mates the charac
teri sti cs of the ci ty road. In thi s protocol, the
sender sel ects an i ntermedi ate node to forward i ts packet, based on the mobi l i ty esti mati on for nei ghbori ng nodes that done
by i ni ti al l y deci di ng whether a nei ghbor node exi sts or not. The sender creates a
posi ti on record for each nei ghbori ng node,
thi s record contai ns the recent posi ti on of the node and i ts mobi l i ty speed; that hel ps i n the sel ecti on of the forwarder nod
e
whi ch i s done based on the route characteri sti cs and the node posi ti on record whi ch ar
ranged after the exchange of beacon
messages. Thi s record avoi ds the l ocal probl em whi ch prevents a node to sel ect a nei ghbor node as a forwarder node; that
happens because there i s no node that i s cl osest to the desti nati on [3
3
].
RIPR protocol i s si mi l ar
to GPSR protocol uses two
modes: a greedy mode and peri meter mode, as wel l as the route characteri stics considerati on, and the posi ti on of the nodes.
Therefore, RIPR can sol ve the l i nk fai l ure probl em caused by stori ng i nformati on about a stal e i ntermedi at
e node; so i t can
reduce the possi bi l i ty of l i nk fai l ure [32].

2.2.3

H
YB
RID
P
OSITION
-
B
ASED
R
OUTING

Posi ti on routi ng protocol reduces control routi ng overhead, i t doesn't need to construct or mai ntai n a routi ng tabl e;
because i t onl y uses the l ocati on i nformati o
n about the nei ghbors and desti nati on nodes, these i ssues made pos
i ti on
-
based
routi ng protocol s scal abl e. However, posi ti on routi ng protocol s have many l i mi tati ons that restri ct thei r usage; these
l i mi tati ons can be summari zed i n the fol l owi ng poi nts [
6
]:



The performance of posi ti on routi ng can be si gni fi cantly decreased accordi ng to the l ocati on accuracy; because the
accurate l ocati ons i nformati on i s an essenti al factor to get a good performance i n posi ti on routi ng.



Posi ti on routi ng coul d be fai l i ng, i f th
ere i s no any nei ghbor node whi ch i s cl oser to the desti nati on (nul l area).



Posi ti on routi ng sol ves the absence of cl osest nei ghbor toward the desti nati on, by the backup process. However, i t
requi red packets to travel l arger di stances to reach desti nati on
s, al so packets coul d be travel i n a cl ose ci rcl e, or
coul d be dropped.

So no exi sti ng routi ng protocol performs effi ci entl y i n al l ci rcumstances.
Therefore, many researchers devel oped hybri d
schemes, they merge character
i sti cs of two or more posi ti on
-
bas
ed routi ng protocol s (non
-
DTN and DTN schemes),
someti mes they merge one or more topol ogy routi ng protocol s (reacti ve, proacti ve an
d hybri d schemes) wi th posi ti on
-
based
First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

11



routi ng. The hybri d posi ti on routi ng protocol i s a mi xture protocol that takes advantag
e of more than one protocol schemes.
The next secti on wi l l i l l ustrate HLAR protocol whi ch bei ng an
exampl e of i s a hybri d posi ti on
-
base
d

routi ng protocol
.

2.2.3.1

H
YBRID
L
OCATION
-
B
ASED
A
D HOC
R
OUTING
P
ROTOCOL
(HLAR):

HLAR i s a hybri d posi ti on routi ng protocol
desi gned to effi ci entl y use al l the avai l abl e l ocati on i nformati on and to
mi ni mi ze the routi ng control overhead. Thi s protocol i s pl anned to swi tch to the on
-
demand routi ng when suffi ci ent l ocati on
i nformati on i s unavai l abl e or l i mi ted, i t al so deal s wi th
the probl em of no cl osest nei ghbor to the desti nati on (voi d regi ons),
and so i t i s al most a scal abl e protocol.

HLAR works as a reacti ve protocol i n the route di scovery process, however i f there i s
no route to the desti nati on node, the source node adds i nformati on about i ts l ocati on and the l ocati on of the desti nati on i n
the route request packet then i t searches for

a cl oser node near the desti nati on. If the node fi nds a nei ghbor whi ch i s cl ose to
the desti nati on then i t forwards the request packet to i t. But i f no cl oser nei ghbor node i s found, i t fl oods the route reque
st
packet to al l i ts nei ghbors. The source node

repeats these steps unti l i t reaches the desi red desti nati on.

The si mul ati on
resul ts showed that the HLAR protocol mi ni mi zes the routi ng control overhead compared wi th the on
-
demand routi ng
protocol s, furthermore i t general l y provi des a fresh l arge si ze
l ocati on i nformati on [
6
]. However, HLAR doesn't guarantee the
best rel i abl e route; because the i ntermedi ate node doesn't have a reverse l i nk to the source, and coul d not i nform other
nei ghbori ng nodes i f i t fi nds a better route to source [3
4
].

Actual l y, al
l categori es of routi ng protocol s have the same objecti ves; that they ai m to decrease the network overhead,
mi ni mi zi ng the transmi ssi on del ay and i ncreasi ng the network throughput. However, i n VANETs i t i s more di ffi cul t to fi nd a
speci fi c routi ng protocol

that works effi ci entl y i n al l network envi ronment si tuati ons; that some protocol may be sui tabl e for
the hi gh mobi l i ty envi ronment but suffer from end to end del ay, i n contrast other protocol s coul d provi de fast packet
del i very, but unsui tabl e for the hi g
h mobi l i ty envi ronment, and so on. So i t coul d be not easy work to preci sel y compare the
exi sti ng VANETs routi ng protocol s, or even cl ai ms whi ch one i s the best i n al l envi ronment si tuati ons; however some research
papers anal yzed the two cl asses and compar
e them usi ng some rel ated protocol s; and thei r resul ts

i s concl uded that the
posi ti on
-
based routi n
g performs better than topol ogy
-
based routi ng for both urban and rural scenari os [
35
]. Tabl e 1
i l l ustrates a compari son betw
een topol ogy
-
based and posi ti on
-
ba
se
d

routi ng, focusi ng on strengths, l i mi tati ons, and
methods used i n each cl ass.

3

T
RANSMISSION
S
TRATEGIES USED IN PA
CKET FORWARDING

Del i very of i nformati on from a source to a desti nati on can be cl assified i nto four types: uni cast, broadcast, mul ti cast, and
geocast, however the mul ti cast and geocast can be merged i n one cl ass because geocast usual ly i s a speci al type of mul ti cast
transmi ssi on.

3.1

U
NICAST
R
OUTING
P
ROTOCOLS

Uni cast routi ng refers to i nformati on del i very from a si ngl e source to a si ngl e desti nati on

usi ng the wi rel ess mul ti hop
scheme; where the i ntermedi ate nodes are used to forward data from the source to the desti nati on, or by usi ng the store
and forward scheme. It i s the most cl ass that wi del y used i n the general ad hoc networks.
Thi s scheme requ
i red the source
vehi cl e to hol d i ts data for a ti me and then forward i t [
36
]. There are many uni cast routi ng protocol s proposed f
or VANETs;
most of the topol ogy
-
based routi ng protocol s bel ong to a uni cast cl ass; such as VADD, AODV, DSR and many other, whi c
h
presented i n the previ ous secti ons.

3.2

B
ROADCAST
R
OUTING
P
ROTOCOLS

Broadcasti ng routi ng enabl es packets to fl ood i nto the network to al l avai l abl e nodes i nsi de the broadcast domai n.
Broadcasti ng routi ng i s wi del y i n VANETs, i t mai nl y used i n the route di sco
very process, some protocol s (l i ke AODV) al l ow
nodes to rebroadcast the recei ved packets. Thi s routi ng scheme al l ows packets to del i ver vi a many nodes whi ch may achi eve
a rel i abl e packet transmi ssi on, however i t coul d consume the network bandwi dth by sendi
ng repl i cated packets, so each
node need to i denti fy whi ch packet i s repl i ca (i t has recei ved i t before)

to di scard.


Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

12



Table 1.

Comparison of Topo
logy
-
based Routing and Position
-
based Routing

3.2.1

D
ENSITY
-
AWARE RELIABL
E BROADCASTING PROTO
COL
(DECA)

DECA i s a densi ty aware protocol; i t uses beacon messages to get knowl edge about i ts nei ghbori ng nodes and to share
i nformati on between nodes. It i s a rel i able broadcast protocol uti l i zes store and forward transmi ssi on scheme.
When a node
broadcasts a pack
et, i t i ni ti al l y chooses a next hop to rebroadcast the packet; the next hop sel ecti on i s based on the amount
of node i nformati on; that means the next hop node wi l l be the node that has the l argest densi ty i nformati on, after the next
hop sel ecti on, the node

adds the next hop ID, to the packet then broadcast the packet. Other nodes whi ch aren't next hops,
shoul d store the packet and startup a wai ti ng ti mer; i f the ti me i s over and no rebroadcast packet recei ved then they
rebroadcast the packet by themsel ves.
Any nei ghbori ng node whi ch recei ved the broadcasted packet, wi l l add i ts ID to the
regul ar beacon, to enabl e other nodes to determi ne whi ch one of i ts nei ghbors that haven't recei ved the broadcasted packet,
i n order to rebroadcast the packet to i t

[10]
. Ma
i nl y, DECA protocol doesn't use any gl obal posi ti on i nformati on (l i ke GPS) i n
i ts processes; that hel p i t to be more fl exi bl e and do wel l i n many network envi ronments [
10
]. However, transmi ssi on of
peri odi c beaconi ng coul d cause a broadcast storm probl em w
hi ch i ncreases the network overhead and decreases the
performance. Al so i f the wai ti ng ti me i s ended wi thout recei vi ng any broadcasted packet, the network wi l l fl ood by
rebroadcast from nei ghbori ng nodes.

3.2.2

P
OSITION
-
AWARE RELIABLE BROAD
CASTING PROTOCOL
(POCA
)

POCA si mi l ar to DECA protocol, i t sel ect certai n nei ghbor nodes to rebroadcast a packet, however i n thi s protocol the
sel ecti on of rebroadcast nodes i s based on thei r posi ti on; other unsel ected nodes stores the packet and startup a wai ti ng
ti mer, i f the
ti me i s over and no rebroadcast recei ved, they rebroadcast the packet by themsel ves. POCA al so depends on
adapti ve beaconi ng whi ch mi ni mi zes beacon overhead, to obtai n i nformati on about nei ghbors' l ocati on
s
, thei r speed, and
thei r connecti vi ty status. Thus

nodes recogni ze i f thei r nei ghbors di d no recei ved some packets and rebroadcast to them.
POCA provi ded a good rel i abi l i ty i n hi gher densi ty [
10
]. However, i t wi l l be a probl em i f the wai ti ng ti me i s over; the network
wi l l fl ood by rebroadcast from nei ghbo
ri ng nodes.

VANET Routing
Protocols

Methods Used

Strengths

Limitations

Comments

Topology
-
based
Routing

Li nk's i nformati on
stored i n the routi ng
tabl e as a basi s on
forwardi ng a packet


The shortest route
from source to
desti nati on


Support of messages
uni cast, mul ti cast and
broadcast


Less resource
consumpti on


Beaconl ess


Save bandwi dth

More overhead


Routes di scover and
mai ntai ni ng del ays


Fa
i l to di scover a
compl ete path
(frequent network
changes)


Unnecessary fl oodi ng

These protocol s
general l y are
proposed for MANETs


Can hel pful for smal l
networks (l ess
overhead)

Position
-
based
Routing

Beaconi ng


Vehi cl es posi ti on
i nformati on


Gl obal
posi ti oning
servi ce

No need to create and
mai ntai n gl obal routes

More stabl e i n hi gh
mobi l i ty envi ronment

More fi tti ng for
network di stri buted
nodes

Lowest overhead

More scal abl e

Obstacl es i n hi ghway
scenari o

Deadl ock probl em i n
l ocati on server


Posi ti on servi ces may
fai l i n tunnel or
obstacl es (mi ssing
satel l i te si gnal)

More sui tabl e for
VANETs; but need
more researches for
smal l networks and
control congesti on

First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

13



3.2.3

D
ISTRIBUTED VEHICULAR

BROADCAST PROTOCOL
(DV
-
CAST)

DV
-
CAST i s a broadcast routi ng protocol uses mul ti hop scheme. In thi s protocol, each node moni tors the stat
u
s of i ts
nei ghbori ng connecti vi ty al l the ti me, i n order to broadcasts to them. DV
-
C
AST deal s wi th di fferent cl asses accordi ng to many
aspects; such as: traffi c state, connected state of the nei ghbori ng nodes, l i ght traffi c, and normal traffi c. It uses the per
i odi c
beacon messages to get i nformati on about the network topol ogy. In a smal l e
r amount of the connected nodes, the node can
be rebroadcast al ong wi th nodes movi ng i n the same way. In di sconnected case of nei ghbori ng nodes, the source node
shoul d use the store and forward scheme; that i t store the broadcasti ng packet unti l i t fi nd an
other node move i nto i ts
broadcast domai n, but i f there i s no node, i t wi l l di scard the packet after the packet l i ve ti me i s ended. The protocol al so
enabl es network nodes to deci de i f the packet i s recei ved before or not; that by usi ng fl ag parameter [
33
]
. Certai nl y, DV
-
CAST
protocol mi ni mi zes the broadcasti ng overhead, whi ch made the protocol appropri ate for both of l i ght and crowded traffi c
si tuati ons. However, i t coul d cause a hi ghl y control overhead and i ncrease the end to end del ay i n the data transmi
ssion [
14
].

3.2.4

D
ISTRIBUTION
-
A
DAPTIVE
D
ISTANCE WITH
C
HANNEL
Q
UALITY
(DADCQ)

DADCQ ai ms to provi de a wel l performed adapti ve mul ti hop broadcast protocol for l arge networks wi th hi gh node
di stri buti on. It sel ects forwardi ng nodes to rebroadcast packets
accordi ng to posi ti onal i nformati on. In rebroadcast deci si on,
when a node recei ved a packet, i t fi rst checks i ts di stance from the desti nati on; i f i t was very cl ose there then no need to
rebroadcast; because i ts rebroadcast wi l l not cover a further area. H
owever, i f thi s di stance i s l arge, then the node has to
rebroadcast the packet.

DADCQ has a mi ni mum transmi ssi on overhead, because i t depends on the node a di stri buted
measurement whi ch i s the l ower changes than network topol ogy changes [
37
].

However i t ma
y cause a
l arge message
overhead.

3.3

M
ULTICAST
-
B
ASED
R
OUTING
P
ROTOCOL

Mul ti cast i s defi ned by sendi ng packets from a si ngl e source to speci fi c group members by mul ti hop communi cati on
[
36
]. Mul ti cast routi ng i n VANETs can be cl assi fi ed i nto two
categori es: ge
ocast and cl uster
-
based routi ng
.

The fol l owi ng
secti on i l l ustrate
s

each
cl ass

i n more detai l.

3.3.1

G
EOC
AST
-
B
ASED
R
OUTING
P
ROTOCOL

Geoc
ast

routi ng protocol s bel ongs to a mul ti cast routi ng protocol whi ch based on sendi ng packets from a source to a
parti cul ar group of desti nati on
s. Some publ i cati ons remark
g
eoc
ast routi ng i
s actual ly a mul ti cast posi ti on
-
based routi ng [
2
]
,
[
7
]. In VANETs, the
geocast routi ng protocol i s a mul ti cast servi ce whi ch enabl es a si ngl e vehi cl e to transmi t a packet to al l
other vehi cl es l ocated i n the speci fi c geographi cal area whi ch l abel ed zone of rel evance (ZOR) [
7
]. Nodes are el ements i n a
one ZOR group, i f they l o
cated i n the same and a speci fi c geographi cal area. The node membershi p i s changed when the
node moves out of the defi ned geographi cal area scope, and i n thi s case i t drops the packet. A zone of forwardi ng (ZOF) i s
defi ned as the geographi c area whi ch vehi
cl es i n thi s area must del i ver the packets to other ZOR vehi cl es. ZOF ai ms to
achi eve a rel i abl e packet's del i very i n hi ghl y dynami c topol ogy. It provi des a peri odi c retransmi ssion, to deal wi th the netw
ork
changes.

The o
ne drawback of g
eoc
ast i s packet tr
ansmi ssi on del ay that caused by network di sconnecti on. There are a
vari ety of proposed Geo cast routi ng protocol s avai l abl e
.

3.3.1.1

R
OBUST
V
EHICULAR
R
OUTING
(ROVER):

ROVER i s a geographi cal mul ti cast protocol; whi ch permi ts each vehi cl e to del i ver a packet to al l

vehi cl es i nsi de a speci fi c
ZOR; usi ng on
-
demand routi ng to di scover packets i nsi de a ZOR. ROVER si mi l ar to AODV protocol; i t onl y fl oods control
pac
kets i n the network, and uni
casts the data packets, thi s scheme ri si ng the consi stency and effi ci ency. Thi s

protocol
assumes vehi cl es have i denti fi cati on numbers, di gi tal map, and gl obal l ocati ons i nformati on.

The source node starts di scover
a route by fl oodi ng i ts ZOR by route request packet, thi s packet i ncl uded source ID, i ts l ocati on, i ts recent ZOR, and a
sequence number of the route. When a vehi cl e recei ved the route request packet, i t accepts the packet; i f i t was nearl y cl ose

to the source and l ocated i nsi de the ZOF and ZOR. If the vehi cl e was outsi de the ZOR, i t doesn't send a repl y. After a vehi cl
e
acc
epts the route request packet, i t sends back a repl y packet contai ns i ts ID to one
-
hop nei ghbors, besi des recorded the
route request packet i nformati on i n i ts routi ng tabl e. And then retransmi t the route request packet

[1]
.

ROVER i s di fferent from AODV i n

that i t sends the repl y back to the node whi ch transmi tted the route request packet not
to the source. So each node can construct a tree of mul ti cast routi ng whi ch has the source node as the tree root. After the
tree i s constructed, the data packet
i s the
n broadcasted i n the tree.

ROVER has achi eved a rel i abl e geographi cal mul ti cast
Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

14



routi ng scheme for VANETs. However, the control packet overhead i s i ncreased as wel l as data del i very del ay; due to the
i ncreased number of retransmi ssi on packets [
2
].

3.3.1.2

M
OBILE
J
UST IN TIME
M
ULTICASTING PROTOCOL

(M
OBICAST
)


Mobi cast

i s a mul ti cast geographi cal protocol, di fferent to conventi onal geocast routi ng protocol,
M
obi cast

routi ng
protocol takes i nto account the ti me aspect. It's desi gned to provi de a management for spati otemporal needs i n VANETs;
that by transmi tti ng a
M
obi cast packet to Vehi cl es i nsi de a ZOR at ti me t (ZORt). Al l vehi cl es bel ong to the ZOR at a ti me t
sho
ul d stay connected to preserve the communi cati on of the real
-
ti me data among the enti re ZOR vehi cl es. The
communi cati on of ZOR i s fai l i ng i f any ZOR vehi cl e unexpectedl y speeds up or sl ows down i ts speed. A l ocati on provi der
(GPS) i s used to know the l oca
ti on of each vehi cl e.

When the network i s temporal fragmented, vehi cl e wi thi n a ZOR may not
effi ci entl y recei ve
M
obi cast packets. Mobi cast protocol desi gned to sol ve thi s probl em usi ng ZOF, and effi ci entl y di ssemi nate
M
obi cast packets. ZOFt i s a ZOF that e
nabl es to di ssemi nate a
M
obi cast packet to every vehi cl e found i n the ZORt. The si ze of
the ZOFt may not usual l y be the opti mal one.
That i f the ZOFt si ze i s bi gger than the ZOFt opti mal si ze, i n thi s case many
unrel ated vehi cl es can be requested to del i ve
r
M
obi cast packets. As wel l i f the ZOFt si ze i s smal l er than the opti mal ZOFt si ze,
i t wi l l cause a temporal network fragmentati on probl em; actual l y the ZOFt i s not easy to accuratel y determi ne i t, especi al ly
in
the hi gh mobi l i ty envi ronment, whi ch may l ea
d to mi suse of the network resources. Mobi cast

esti mates the ZOFt si ze to a
val ue i s cl ose to the opti mal, as wel l as determi nes the structure, and l ocati on of the ZOFt, to dynami cal l y form fl exi bl e ZO
F

[32].

However, Mobi cast

has a l i mi tati on because of i
t rel i es on the l ocati on provi der (GPS) to know a gl obal knowl edge about
the network densi ty; because i t may not perform wel l i n a l arge network wi th a dynami c and densi ty envi ronment. Al so the
ZOF

i s may be l arge wi thout the need (due to i ncorrect confi gu
rati on esti mati on), thi s may i ncrease the network overhead.

3.4

C
LUSTER
-
B
ASED
R
OUTING
P
ROTOCOL

Thi s protocol di vi des the network to cl usters, where nodes have the same characteri sti cs, l i ke same di recti on or same
vel oci ty, or so on. Each cl uster has a cl uster

head, i ts task i s to manage communi cati on processes i nsi de, and to outsi de i ts
cl uster. Nodes i nsi de the cl uster communi cate by di rect paths, but thei r communi cati on wi th other nodes outsi de the cl uster
i s achi eved by thei r cl uster header, and thi s create
s a vi rtual i nfrastructure for networks. Thi s scheme can provi de a good
scal abi l i ty for l arge networks; however i t may i ncrease network overhead and del ays i n hi ghl y dynami c network [
2
]
,
[1
4
].

3.4.1

COIN:

C
LUSTERING FOR
O
PEN
IVC

N
ETWORK

COIN i s a cl usteri ng
mechani sm desi gned to i mprove network scal abi l i ty, i t di vi des the network to cl usters; but not l i ke
conventi onal other cl usteri ng protocol s, COIN sel ects cl usters accordi ng to three parameters: mobi l i ty of nodes, nodes
posi ti ons and behavi or of nodes. The
protocol provi des each cl uster speci fi c ti me whi ch i s a ti me to l i ve; i n order to decrease
control overhead.

Inter vehi cl es communi cati on system (IVC) deal s wi th the unstabl e di stances of i nter vehi cl es. To enabl e a
head of cl uster node and the cl uster mem
ber node stay conti nue communi cate, thei r mobi l i ty shoul d be l ow and rel ated to
the mobi l i ty of each, i n thi s case they can resi de i n radi o contact for a l onger ti me [
6
]
,
[1
4
].

3.4.2

C
LUSTER
-
B
ASED
D
IRECTIONAL
R
OUTING
P
ROTOCOL
(CBDRP)

In CBRP the network i s
di vi ded i nto several cl usters. Each cl uster has a cl uster head whi ch i s responsi bl e for the routi ng
procedure. These cl uster heads communi cate wi th each other vi a gateway nodes whi ch are nodes that have more than one
cl uster head. When a source node reques
ts a route, i t fl oods the network by request packet. The CBRP cl ustered structure
reduces tra
ffi
c overhead, because request packet onl y passes among cl uster heads. However, i n densi ty networks, the packet
overhead wi l l i ncrease and the transmi ssi on del ay w
i l l i ncrease; because i nformati on of each node i n the route shoul d be
added to the routed packet whi ch i ncreases the si ze of the packet. One more l i mi tati on of the CBRP, i t mai ntai ns
uni di recti onal l i nks, whi l e most of l i nk l ayer protocol s support onl y bi d
i recti onal l i nks [3
8
].

General l y, cl uster
-
based routi ng protocol s may perform wel l i n network scal abi l i ty for l arge
networks;

however
, they may
cause more
networks

overhead due to structure of cl usters and cl uster heads

[3
4
]

4

VANET
S
R
OUTING
O
PEN
I
SSUES

Thr
ough our l i terature revi ew i n VANETs routi ng protocol s, we found there are sti l l some open i ssues and chal l enges i n
VANETs routi ng, whi ch i t i s one of the most acti ve topi cs i n VANETs research area, i t has several recent publ i cati ons. Thi s
First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

15



secti on represen
ts some open research i ssues i n VANETs routi ng probl em (
for exampl e, but not l i mi ted to). It requi red
desi gni ng a si ngl e routi ng protocol that can:



Works effi ci entl y for both urban and rural, has the abi l i ty to i mproved networks throughput and packet del i v
ery rati o.
Al so reduces resource consumpti on, and guarantee opti mal paths.



Scal abl e; has the abi l ity to handl e dynami c connecti vi ty for broken l i nks, as wel l as deal s wi th the condi ti ons of a si ngl e
network, l i ke crowded, congesti on, avai l abl e bandwi dth, t
ransmi ssi on i nterference, al l owed speed, and so on.



Adapti ve rel i abl e broadcasti ng/ mul ti cast transmi ssi ons, that works effi ci entl y i n sendi ng packets for al l nodes wi th
mi ni mum overhead, dupl i cati ons, col l i si ons and congesti ons.



Can sol ve hi dden termi nal
probl ems; to avoi d out range col l i si on.

Intel l i gent to adapt and deal wi th unexpected condi ti ons; l i ke dri ver behavi or, si gnal l oss, i nterference by tunnel s, hi gh
bui l di ng, and i ntersecti ons con
di ti on.

Bri efl y, an expected VAENETs routi ng protocol shoul d b
e abl e to provi di ng
communi cati on wi th mi ni mal overhead and del ay, hi ghest scal abi l i ty and adaptabl e for VANETs envi ronments; by usi ng
opti mal route sel ecti on and powerful reconfi gurati on al gori thm.


Table 2.

Comparison of
VANET Transmission Strategies


Transmission

strategies

Methods Used

Strengths

Limitations

Comments

Unicast


Information delivery
from a single source to
a single destination

Less network
overhead


More privacy


Minimum packet delay

Links should be
frequently configured
and maintain


Less
reliability


Packet loss

Need more researches
to enhance reliability,
packet retransmission,

scalability and avoid
collision

Broadcast


Packets flooding to all
network nodes inside
the broadcast domain

More reliable data
transmission


Less packet loss


Consumes bandwidth


Routes loop


Network congestion


Less network
throughput


More packet delay


Packet collisions

Required reducing
bandwidth
consumption


Could be useful for
alert messages


Need some packets
flooding constrains

Multicast


Geocast

sending
packets from a source
to a group of
destinations using
geographic addresses


Cluster divides the
network to clusters,
each cluster has a
cluster head to
manage
communication inside
and outside the
cluster

Efficient routing by
sending one copy to
m
ultiple nodes


Minimum network
consumption



Minimum packet
delivery delay


Easy to implement


Transparent to
changeable addresses
(no requirement to
recei ver’s address)

ConVumeV banTwi TWh


More overheaT i n
Ti vi Ti ng neWwork
noTeV i nWo groupV


RouWeV l oop

卣al abi 汩ly conWrol 景r
Tynami c groupV


The cl uVWer may noW
very e晦i ci enW becauVe
晲equenW changi ng
heaTV

(l i ke Mobi l e rouWerV i n
neWwork mobi l i Wy buW
wi WhouW a guaranWee
Whe neWwork noTeV wi 汬
Wravel aV one uni W)

Sample paper of International journal of Innovation and Applied Studies



ISSN: 2028
-
9324

Vol. X No. X, Month 20XX

16



5

C
ONCLUSION

Thi s paper has presented

an overvi ew of Vehi cul ar s ad hoc networks (VANETs), i l l ustrates thei r moti vati on and
characteri sti cs, i t studi ed i n detai l VANETs routi ng probl em, mai nl y vehi cl e to vehi cl e (V2V) communi cati on, provi di ng two
cl assi fi cati ons of VANETs routi ng protocol s th
at exi st i n the l ast few years, i nvesti gated them and showi ng how do they work
and thei r mai n advantages and l i mi tati ons

The paper al so summari zed compari sons between the mai n cl asses.

Thorough thi s study of di fferent VANETs routi ng protocol s, many rel ate
d open i ssues and research chal l enges are found
and represented, these i ssues sti l l requi red more effort and research to address them. We hope that the i nstrument
presented i n thi s paper to be useful and hel pful to students and researchers i n the fi el d.

R
E
FERENCES

[1]

Rami n Kari mi, Norafi da Ithni n ,Shukor Abd Razak

and

Sara Najafzadeh,

Non DTN Geographi c Routi ng Protocol s for
Vehi cul ar

Ad Hoc Networks
,”

International Journal of Computer Science Issues

(
IJCSI
)
,
v
ol. 8, Issue 5,
n
o
.

3, September
2011.

[2]

Bi jan

Paul, Mohammed J. Isl am,

Survey over VANET Routi ng Protocol s for V
ehi cl e to Vehi cl e Communi cati on
,


IOSR
Journal of Computer Engineering (IOSRJCE)
,
ISSN: 2278
-
0661, ISBN: 2278
-
8727
v
ol
.

7, Issue 5 (Nov
-
Dec.
2012),
pp.

01
-
09
.

[3]

Moath Muayad Al
-
Doori, "Di rec
ti onal Routi ng Techni ques i n VANET,"
PhD Thesis, Software Technology Research
Laboratory, De Montfort University, Leicester United Kingdom
, November 2011.

[4]

Lee, Kevi n C., Ui chi n Lee and Mari o Gerl a., "Survey of Routi ng Protocol s i n Vehi cul ar Ad Hoc Networks
,"
Advances in
Vehicular Ad
-
Hoc Networks: Developments and Challenges reference
.
IGI Global
, 2010,
pp.
149
-
170, Web. 25 Mar. 2013.

[5]

Sal i m Al l al and Saadi Boudji t
, “
Geocast Routi ng Protocol s for VANETs: Survey and Geometry
-
Dri ven Scheme Proposal
,


Journal
of Internet Services and Information Security (JISIS)
, vol
.
3, n
o
.

1/2, pp. 20
-
36
,
February 2013
.

[6]

Mohammad Al
-
Rabayah and Robert Mal aney, "A New H
ybri
d Locati on
-
based Ad Hoc Routi ng Protocol
,
"

Global
Telecommunications Conference (GLOBECOM 2010)

IEEE
, vol.

1,

no.

6,
pp.
6
-
10
,

Dec. 2010
.

[7]

Rakesh Kumar

and

Mayank Dave
, "
A Comparati ve Study of Vari ous Routi ng Protocol s i n VANET
,"
International Journal of
Computer Science Issues

(
IJCSI
)
,

v
ol. 8, Issue 4,
n
o
.

1, Jul y 2011
.

[8]

Vi jayal askhmi M., Avi nash Patel,
Li nganagouda Kul karni, “QoS Parameter Anal ysi s on AODV and DSDV Protocol s i n a
Wi rel ess Network
,


International Journal of Communication Network & Security
,
v
ol.

1,
n
o.

1, 2011.

[9]

l ajos hanzo and rahi m tafazol l i, "Asurvey of QoS Routi ng Sol uti ons for Mobi l e

Ad hoc Networks
,
"
IEEE C
ommunications
S
urveys

&
T
utorials • 2nd quarte.

2007
.

[10]

James Bernsen and D. Mani vannan,"Greedy Routi ng Protocol s for Vehi cul ar Ad Hoc Networks
,
"

Wireless
Communications and Mobile Computing Conference
. IWCMC 08. Internati onal
,

vol.
632,

no.

637,
pp.

6
-
8 Aug. 2008
.

[11]

Jamal Toutouh, Enri que Al ba, "Opti mi zi ng OLSR i n VANETS wi th Di fferenti al Evol uti on: A Comprehensi ve Study
,
"

ACM

DIVANet’11,

Miami, Florida, USA.
,
November 4, 2011
.

[12]

Sonam Jai n and Sandeep Sahu, "Topol ogy
vs.

Posi ti on based Routi ng Protocol s i n Mobi l e Ad hoc Networks: A Survey
,
"
International Journal of Engineering Research & Technology (IJERT)

v
ol. 1
, Issue 3, May 2012.

[13]

Y. Lachari te and M. Wang, "Hi erarchi cal OLSR draft
-
l achari te
-
manet
-
hol sr
-
02
,
"

Mobile Ad ho
c Networking (MANET)
, IETF

Internet
-
Draft,

Jul y 13, 2009
,

url

http://tool s.i etf.org/html/draft
-
l achari te
-
manet
-
hol sr
-
02
.

[14]

Yatendra Mohan Sharma

and

Saurabh Mukherjee
, "A Contemporary
Proporti onal Expl orati on of Numerous Routi ng
Protocol i n VANET",
International Journal of Computer Applications
,

v
ol
.
50
, n
o.

21,
pp. 14
-
21,
Jul y 2012.

[15]

Sandhaya Kohl i, Bandanjot Kaur and
Sabi na Bi ndra,

"
A

comparati ve study of Routi ng Protocol s i n VANET",
International
Journal of Computer Science, Issues
-
IJCSI
,
Proceedings of ISCET
,
2010
.

[16]

Haerri, J., Fi l al i, F., and Bonnet, C., "Performance compari son of AODV and OLSR i n VANETs urban envi ronments under
real i sti c mobi l i ty patterns",
In Proceedings 5th IFIP M
editerranean Ad
-
Hoc Networking Workshop
, June 14
-
17, 2006,
Li pari, Ital y.

[17]

M. Abol hasan, T. Wysocki and E. Dutki ewi cz, “A revi ew of routi ng protocol s for mobi l e ad hoc networks”,
Ad Hoc
Networks 2, Elsevier
,
pp. 1

22,
2004
.

[18]

Raphael Frank, Eugeni o Gi ordano,
Thomas Engel and Mari o Gerl a "Performance Bound for Routi ng i n Urban Scenari os",
ACM AINTEC’11
, November 9

11, 2011, Bangkok, Thai l and.

[19]

Naumov, V., Baumann, R., Gross, T. (2006), “An eval uati on of Inter
-
Vehi cl e Ad Hoc Networks Based on Real i sti c
Vehi cul ar
Traces,”
Proc. ACM MobiHoc’06 Conf.
, May, 2006.

[20]

S. Das, C. Perki ns, E. Royer, "Ad hoc on demand di stance vector (AODV) routi ng",
Internet Draft, draft
-
ietf
-
manet
-
aodv
-
13.txt
,
Mobile Ad Hoc Networking Working Group
,

2003
.

[21]

M. K. Mari na, and S. R. Das, “On
-
demand mul ti path di stance vector routi ng i n Ad Hoc networks,”
Proc. 9th International
Conference on Network Protocols, IEEE Press
,
pp. 14
-
23,
Nov
ember. 2001.

First Author, Second Author, Third Author, Fourth Author and
Fifth Author



ISSN: XXXX
-
XXXX

Vol. X No. X, Month 20XX

17



[22]

Hafez Maowad and Eman Shaaban
, "Enhanci ng AOMDV Routi ng Protocol for V2V Communi cati on",
In Proceedings of
the 6th international conference on Communications and Information Technology, World Scientific and Engineering
Academy and Society (WSEAS)
, url
http://dl.acm.org/ci tati on.cfm?i d=2209535.2209538
, 2012
.

[23]

Boangoat Jarupan and Eyl em Eki ci, "A survey of cross
-
l ayer desi gn for VANETs
,
"
Ad Hoc Networks (Elsevier) Journal
Special Issue on Vehicular Networks
,
doi:10.1016/j.a
dhoc.2010.11.007, 2010
.

[24]

Sommer, C., Dressl er, F. "The DYMO Routi ng Protocol i n VANET Scenari os
,
"
Vehicular Technology Conference
,
VTC
-
2007
fall
. IEEE 66th
,
Pp:

16

20, 2007.

[25]

Annu Mor, "
A

STUDY OF IMPROVED AODV ROUTING PROTOCOL IN VANET
,
"
International
Journal of Computer
Applications & Information Technology
,
v
ol. II, Issue I, January 2013 (ISSN: 2278
-
7720).

[26]

Sufyan T. Faraj Al
-
Janabi, Yaseen Sal eem Yaseen, and Bob Askwi th, "The Bus Ad Hoc On
-
demand Di stance Vector
(BAODV) Routi ng Protocol",
PGNet
ISBN:
978
-
1
-
902560
-
26
-
7,

2012
.

[27]

Johnson, Davi d B., Davi d A. Mal tz, and Josh Broch. "DSR: The dynami c source routi ng protocol for mul ti
-
hop wi rel
ess ad
hoc networks,
"
Ad hoc networking

5
, pp. 139
-
172, 2001.

[28]

Gurl een Kaur Wal i a, "A Survey on Reacti ve Routi ng Protoco
l s of the Mobi l e Ad hoc Networks
,
"
International Journal of
Computer Applications
,

v
ol
.

64
, n
o.22,
pp. 45
-
51,
February 2013.

[29]

V. Park, S. Corson, "Temporal l y
-
Ordered Routi ng Al gori thm (TORA) Versi on 1 Functi onal Speci fi cati on
,
"
IETF Internet
draft, 2001
. UR
L:
http://tool s.i etf.org/pdf/draft
-
i etf
-
manet
-
tora
-
spec
-
04.pdf

[30]

Jan Schaumann, "Anal ysi s of the Zone Rout
i ng Protocol," avai l abl e at:

http://www.netmei ster.org/mi sc/zrp/zrp.pdf
,
December 8, 2002.

[31]

C
hao Song, Mi ng Li u, Yonggang Wen, Ji annong Cao and Gui hai Chen "Buffer and Swi tch: An Effici ent Road
-
to
-
road
Routi ng Scheme for VANETs",
Seventh International Conference on Mobile Ad
-
hoc and Sensor Networks
,

2011
.

[32]

Si
Ho Cha1, Mi n
-
Woo Ryu and Kuk
-
Hyun Cho, "A Survey of Greedy Routi ng Protocol s for Vehi cul ar Ad Hoc Networks
,
"
Smart Computing Review
, vol. 2, no. 2, Apri l 2012.

[33]

M.
W. Ryu,

S. H.

Cha, J.
G. Koh, S. Kang and
K.
H.
Cho, “Posi ti on
-
based routi ng al gorithm for i mprovi ng rel i abi lity of i nter
-
vehi cl e communi cati on
,


Transactions on Internet and Information Systems (TIIS)
, vol. 5, no. 8, Aug. 2011.
url
http://freepatentsonl i ne.com/arti cl e/KSII
-
Transacti ons
-
Internet
-
Informati on
-
Systems/272246267.html
.

[34]

James. Jong Hyuk Park, Vi ctor C.M. Leung,

Cho
-
Li. Wang and Taeshi k Shon,
"
Future Informati on technol ogy,
Appl i cati ons, and Servi ces"
,
FutureTech
Springer
,

v
ol
.
1, pp. 82
-
98,

2012
,
url
http://books.googl e.com/books?i d=pggzUBA
SE4IC&pri ntsec=copyri ght#v=onepage&q&f=fal se
.

[35]

Akhtar Husai n, Ram Shri ngar Raw, Brajesh Kumar and Ami t Doegar, "Performance Compari son of Topol ogy and Posi ti on
Based Routi ng Protocol s i n Vehi cul ar Network Envi ronments",
International Journal of Wireless & M
obile Networks
(IJWMN)
,

v
ol. 3,
n
o. 4, August 2011.

[36]

Y
un

L
i n
,
Y
uh
C
hen

and

S
i ng
L
ee
,"

Routi ng Protocol s i n Vehi cul ar Ad Hoc Networks: A Survey and Future Perspecti ves",
J
ournal of

I
nformation

S
cience and

Engineering
, vol.
26,
pp. 913
-
932, 2010.

[37]

Mi chael
Sl avi k

and

Imad Mahgoub, "Spati al Di stri buti on and Channel Qual i ty Adapti ve Protocol for Mul ti hop Wi rel ess
Broadcast Routi ng i n VANET
," IEEE Transactions on Mobile Computing
, vol. 12, no. 4, pp. 722
-
734, Apri l, 2013
.


[38]

Ti m Dani el Hol l erung, "The Cl uster
-
Bas
ed Routi ng Protocol",
project group Mobile Ad
-
Hoc Networks Based on Wireless
LAN
,
University of Paderborn
,
2003/2004.