186 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.

1, January 2009

A Review and Comparison of Reliable Unicast Routing Protocols

For Mobile Ad Hoc Networks

Hadi Sargolzaey†, Ayyoub Akbari Moghanjoughi†† and Sabira Khatun†††

Wireless Routing Research Group, Department of Computer and Communication Systems Engineering
Faculty of Engineering, University Putra Malaysia, 43400 Serdang, Selangor, Malaysia

Classification of unicast routing protocols in MANETs can

Summary be done in many ways, but most of them are depending on
An important and essential issue for mobile ad hoc networks is routing strategy and network structure. According to the
routing protocol design that is a major technical challenge due to routing strategy, the routing protocols can be categorized
the dynamism of the network. Node failures and arbitrary as proactive and reactive routing (see Fig. 1), while
movement of nodes break the routes and lead the frequent
depending on the network structure these are classified as
operation of rebuilding routes that consume lots of the network
resources and the energy of nodes. Many efforts have been made
flat, hierarchical and position based routing. Both the
to design reliable routing protocols that enhance network stability. proactive and reactive protocols come under the flat
This paper review and compare some recently published reliable routing.
unicast routing protocols, that can be used for selecting
appropriate routing protocols, which are suitable for a particular
application. Also it will maximize the reliability of the deployed
network. Furthermore the proposed methods used in the reviewed
protocols, can help designers and researchers for enhancing the
reliability of their new routing protocols.

Key words:
Reliable routing protocols, Unicast routing, Mobile Ad-hoc
Networks.

1. Introduction
Fig. 1 Classification of MANET routing protocols
The mobile ad hoc network (MANET) allows a more
flexible communication model than traditional wire line Proactive Routing Protocols: A proactive routing protocol
networks since the user is not limited to a fixed physical is also called "table-driven" routing protocol. Using a
location [1]. It is a new special network that does not have proactive routing protocol, nodes in a mobile ad hoc
any fixed wired communication infrastructure or other network continuously evaluate routes to all reachable
network equipments. With no pre-existing fixed nodes and attempt to maintain consistent, up-to-date
infrastructure, MANETs are gaining increasing popularity routing information. Therefore, a source node can get a
because of their ease of deployment and usability anytime routing path immediately if it needs one. When a network
and anywhere. So they are viewed as suitable systems topology change occurs, respective updates must be
which can support some specific applications as virtual propagated throughout the network to notify the change.
classrooms, military communications, emergency search So if we noted to network topology changes in MANETs,
and rescue operations, data acquisition in hostile the control overhead to maintain up-to-date network
environments, communications set up in Exhibitions, topology information is relatively high. Wireless Routing
conferences and meetings, in battle field among soldiers to Protocol (WRP) [3], the Destination Sequence Distance
coordinate defence or attack, at airport terminals for Vector (DSDV) [4] and the Fisheye State Routing (FSR)
workers to share files etc. [5] are all proactive routing protocols.
Host mobility can cause unpredictable network topology Reactive Routing Protocols: Reactive routing protocols for
changes in MANETs. Hence, a highly adaptive routing mobile ad hoc networks are also called "on-demand"
scheme to deal with the dynamic topology is required. routing protocols. In a reactive routing protocol, routing
Many unicast routing protocols have been proposed for paths are searched only when needed. When a source node
MANETs to achieve efficient routing [2]. wants to send packets to the destination but no route is

Manuscript received January 5, 2009

Manuscript revised January 20, 2009
IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009 187

available, it initiates a route discovery operation. In the route is generally not involved as a requirement for its
route discovery operation, the source broadcasts route selection. Consequently, route breakups will frequently
request (RREQ) packet. When the destination or a node occur, induced by nodal mobility and/ or nodal and link
that has a route to the destination receives the RREQ failures as well as by fluctuations in the communications
packet, a route reply (RREP) packet is created and transport quality experienced across the network’s
forwarded back to the source. Each node usually uses hello communications links. The latter are caused by signal
messages to notify its existence to its neighbours. interferences, fading and multi-path phenomena and other
Therefore, the link status to the next hop in an active route causes producing ambient and environmental noise and
can be monitored. When a node discovers a link signal interference processes. On the other hands, route
disconnection, it broadcasts a route error (RERR) packet to breakups lead the frequent operation of rebuilding routes
its neighbours, which in turn propagates the RERR packet that consume lots of the network resources and the energy
towards nodes whose routes may be affected by the of nodes.
disconnected link. Then, the affected source can re-initiate Many efforts have been made to design reliable routing
a route discovery operation if the route is still needed. protocols that enhance network stability. In this paper, we
Compared to the proactive routing protocols, less control review and compare some recent published reliable unicast
overhead is a distinct advantage of the reactive routing routing protocols. This paper has organized as following;
protocols. Thus, reactive routing protocols have better in section 2, some reliable routing protocols are described
scalability than proactive routing protocols. However, briefly. Section 3 presents the comparisons on the
when using reactive routing protocols, source nodes may reviewed protocols, and finally conclusion of paper comes
suffer from long delays for route searching before they can in section 4 as well.
forward data packets. Hence these protocols are not
suitable for real –time applications. The Dynamic Source
Routing (DSR) [6] and Ad hoc On-demand Distance 2. MANET Reliable unicast routing protocols
Vector routing (AODV) [7] are examples for reactive
routing protocols. 2.1 Energy Aware Reliable Routing (EARR)
Hierarchical Routing Protocols: Typically, when wireless
network size increase (beyond certain thresholds), current EARR is a cross-layer reactive routing protocol that
“flat” routing schemes become infeasible because of link reduces the route-reconstructions due to residual energy
and processing overhead. One way to solve this problem shortages [14].
and to produce scalable and efficient solutions is It is assumed that all nodes in EARR are equipped with a
hierarchical routing. Wireless hierarchical routing is based residual energy detection device and some energy
on the idea of organizing nodes in groups and then consumption model is offered for each node to estimate
assigning nodes different functionalities inside and outside whether its remaining battery capacity is sufficient to relay
of a group. The Zone Routing Protocol (ZRP) [8], Zone- the traffic. Also the traffic load from application layer is
based Hierarchical Link State routing (ZHLS) [9] and carried in the RREQ packet. In this protocol, only nodes
Hybrid Ad hoc Routing Protocol (HARP) [10] are with sufficient residual energy to complete the task will
examples for hybrid routing protocols. take part in the propagation of RREQ packet.
Position Based Routing Protocols : The advances in the Finally after all RREP packets received, source node pick
development of Global Positioning System (GPS) the optimum path with maximal bottleneck energy to route
nowadays make it possible to provide location information data packets. (Bottleneck energy refers to the minimal
with a precision in the order of a few meters. They also residual energy of all nodes along one path)
provide universal timing. While location information can In EARR, when a source node wants to send packets, if it
be used for directional routing in distributed ad hoc has already kept at least one route to the destination, then
systems, the universal clock can provide global for each path it will check whether each node on the path
synchronizing among GPS equipped nodes. In position has sufficient residual energy to complete the forwarding
based routing protocols, instead of using routing tables and task. Only when all nodes on the path have enough battery
network addresses, the routing decisions are made on the the path can be thought as valid one. Like the previous
basis of the current position of the source and the state, if several paths are valid, the one with maximal
destination nodes. Location Aided Routing (LAR) [11] bottleneck energy is chosen to transmit data packets.
and Distance Routing Effect Algorithm for Mobility Both periodical information exchange and error report are
(DREAM) [12] are typical position based routing used in EARR to maintain routes, as in many conventional
protocols proposed for mobile ad hoc networks. According reactive routing protocols. The only difference lies in that
to several experimental works, routing schemes that use when nodes exchange information with neighbours, not
positional information scale well [13]. only address information but also current residual energy
In all of the unicast routing protocols, the robustness of the info are carried by hello messages, so that each node is
188 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009

able to obtain the most up-to-date energy information of amount of time to receive other RREQ packets. Then node
neighbours. D computes the weight value using the weight function.
Afterwards, node D selects the path with the maximum
weight value as the primary routing path among all
2.2 Stable Weight-based On-demand Routing feasible paths. Finally, node D sends a RREP packet to
Protocol (SWORP) source node S along the primary routing path.
Another form of this protocol (SWORP-BR) also is
SWORP is a reactive routing protocol that uses the weight- proposed for selecting alternative paths. When a link
based route strategy to select a stable route in order to failure occurs, the alternative path selection algorithm is
improve routing quality in MANETs. The basic idea is to triggered. The method is to overhear RREP packets
select a stable routing path to reduce the routing overhead transmitted by a neighbouring node. It records this
and packet loss [15]. neighbouring node as a next hop to the destination node in
The weight of a route is decided by three factors: the its alternate route table. For example, Figure 2 shows the
Route Expiration Time (RET), the Error Count (EC), and process of constructing the primary and alternate routes.
the Hop Count (HC).
The RET is the minimum Link Expiration Time (LET) for
a feasible path. The LET can be obtained by using the
principle that two neighbours in motion will be able to
predict future disconnection time. Such a prediction can be
accomplished by the following method.
The motion parameters of two neighbouring nodes can be
obtained by using the Global Positioning System (GPS). A
free space propagation model is assumed which signal
strength solely depends on the distance to the transmitter.
Also it is considered that all nodes have their clocks
synchronized using the GPS clock. On the other hand, by
considering the motion parameters of two nodes, the
duration of time which these two nodes remain connected
can be calculated. The speed and heading of a mobile node
can be obtained from the mobile node’s own instruments
and sensors (e.g., compass, odometer, speed sensors).
When each node sends a RREQ packet, the packet
appends its location, direction, and speed. So, the next hop
of the node that receives the RREQ packet can predict the
link expiration time between itself and the previous node.
The EC is the maximum node error count for a feasible
path. The error count is used to indicate the number of
link failures caused by a mobile node.
The SWORP effectively combines all the three parameters
in a weight function with weighing factors C1, C2, and C3,
the values of which can be chosen according to the system
requirements. As an illustration, route expiration time is
very important in MANETs. Thus, the weight of that
factor can be made larger. The flexibility of changing the
factors helps in using this algorithm to select a optimum
routing path. A larger route expiration time represents
higher reliability, as do a lower error count and lower hop
count.
In the route discovery process source node S broadcasts a
RREQ packet to its neighbouring nodes. If the node is
within the transmission radius, it forwards the RREQ
packet to its own neighbouring nodes and adds its ID, the
route expiration time, the error count, and the hop count of
the RREQ packet to the packet entry. When destination
node D receives a RREQ packet, it waits for a certain
IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009 189

Fig. 2 Multiple route construction of stable weight-based on-demand service determines how routing between two nodes of the
routing protocol with backup routes (SWORP-BR): (a) destination node
D sends a route reply (RREP), (b) node L forwards the RREP, (c) the
network is performed, based on the information provided
primary route and alternate routes are established and (d) the data packet by the first service (MCS) and the software components of
is transmitted via an alternate route when the primary route is the available routing protocols. It is called the strategy
disconnected. selection service (SSS). The third service is the actual
routing service offered to the other software agents
With the proposed method, when the RREQ packet running in the mobile node, and ‘‘looks’’ like a routing
reaches the destination node D, the primary route (S, A, C, protocol. It is called the routing service (RS).
F, I, L, D) is selected. Then destination node D sends a For a graphical representation of the interconnection of the
RREP to node L. At this time, nodes Y and Z that are three services (MCS, SSS, and RS) with the lower (MAC)
within the transmission range of node D overhear the and higher (Network) layers, as well as the routing
RREP packet and insert an entry into their alternate route protocols (see Fig. 3).
table. This process is shown in Figure 2.a. Similar
processes are executed at the next hops backwards until
the RREP packet reaches the source node S. Figure 2.c
shows the state when the RREP reaches the source node S.
The protocol then builds the primary and the alternate
routes. Figure 2.d illustrates the use of an alternate path
when the primary route gets disconnected. Node i move
out of the transmission range of its neighbouring node L. It
cannot transmit data from node F to node L. Therefore,
node I must broadcast the packet to its neighbouring nodes
to salvage the data. When node W receives the packet, it
looks up its alternative route table and finds node L as its
next hop toward the destination node D. Then node W
unicast the packet to node L and builds the new connection.

2.3 Mobility Sensitive Approach (MSA)

MSA is a protocol framework that is implemented on top
of the network layer that tries to adapt to the diverse
mobility rates of the nodes as exhibited in real networks Fig. 3 Architecture of the framework – dotted lines represent inter-
[16]. Instead of providing a single routing strategy, this process communication, continuous lines represent packet flow. (AODV,
framework uses a suite of complementary protocols that RUNNERS and DSDV is routing protocol examples)
are suitable for different types of mobile networks.
In this work, a metric is proposed, named stability, which The main idea is to allow each node to execute different
is used for characterizing the relative mobility of the nodes. routing protocols concurrently and let the mobility-
The metric uses the notion of associativity, which is the sensitive framework control their operation (by
time (in beacons) that nodes are associated (i.e., they retain activating/deactivating them) and reroute the offered
a connection). According to its stability, every node is services through RS. This framework can be extended to
classified to a mobility class. Based on these mobility make this choice based on other network conditions, such
classes, a novel protocol framework is designed that as available energy, geo-location, QoS guarantees, etc.
operates on top of the network layer and, for any pair of
origin and destination nodes, determines the routing 2.4 Distributed Long Lifetime Route (DLLR)
technique (among those available to the nodes) that best
corresponds to their mobility properties. DLLR is a reactive routing protocol that discovers two of
It is assumed that every node has access to a beaconing the most desirable Long Lifetime Routes (LLRs) through
mechanism, which is either provided by the data link layer one best-effort route discovery procedure [17]. If two
or can be easily implemented using the reactive protocol. LLRs can be provided at a time, the routing protocol can
(e.g., the hello messages) save the longer LRR as a backup and use the shorter LRR,
This framework can be divided into three services (i.e., which usually has a shorter route length and is thus more
software agents). The first service is concerned with the energy-efficient, as the primary route for transmissions.
calculation of the associativity of the links and the The routing protocol can switch to the longer LRR to
classification of the nodes into mobility classes. It is called maintain the flow of traffic when the shorter LLR breaks.
the mobility classification service (MCS). The second Meanwhile, a new route discovery procedure can be
190 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009

initiated, and when the newly discovered LLRs are done through the use of a normalized cumulative path
returned, the old LLRs can be replaced. reliability factor that is calculated by the destination.
Link lifetime can be estimated through the link break In this protocol, the source node sends the RREQ packet
probability (probability that a link is broken at time t) which is similar to other reactive routing protocols but
given an estimation rule, such as from now to when the contains the following different fields: (1) RMIN: The
link break probability is higher than a certain threshold. minimum value for the reliability factor required in the
The quality of a link can be thus quantified using this path from source to destination. (2) RCUM: The
estimated link lifetime. Using this quantified link lifetime cumulative reliability factor of the path that is being
estimations, the route lifetime is simply the minimum discovered. (3) NH: It contains a list of the next hop host
lifetime of the n links. candidates that satisfy the reliability requirements for the
The main procedure of DLLR is similar to a typical path that is being discovered. (4) MAX NH: Maximum
reactive on-demand routing protocol. The difference lies in number of nodes in the NH list. This parameter is intended
the implementation details such as RREQ packet format, to control the flooding of the RREQ packet during route
routing update rules and RREQ forwarding delay rules. discovery. (5) BACKUP PATHS: This is the maximum
In DLLR, RREQ packet contains a primary route that number of backup paths that can be included in the routing
expands while the RREQ propagates throughout the table of the source node.
network. In addition, it contains an auxiliary route (backup When an intermediate node receives a RREQ packet from
route) which does not have any effect on RREQ another node, it builds the NH list with all of the
forwarding decisions. Also, included in the RREQ are the neighbours whose reliability factor is above RMIN. It then
primary and auxiliary route lifetimes. These route lifetimes adds its reliability factor to the cumulative path reliability
are calculated at each intermediate node by choosing the factor RCUM and propagates the RREQ packet to its
minimum from the composed estimated link lifetimes. The neighbours. Each of the neighbours whose ID is in the NH
source node broadcasts this RREQ packet, which initially, list will further propagate the RREQ packet according to
the two routes only contain the source node and their the same algorithm.
lifetimes are set as 0. When and if the destination receives the RREQ packets, it
When an intermediate node receives an RREQ for the first selects the path with the highest normalized path reliability
time, it appends itself into the prim/aux routes in the factor PRF=RCUM/n, where n is the number of
packet and records the request locally. Then it adjusts the intermediate nodes in the path (not including the source
lifetimes by choosing the minimum of the previous route and destination nodes), and unicast a RREP packet back to
lifetime and its link lifetime with the previous node. Next, the source. If backup paths are required then the paths with
it schedules a local delay time for forwarding this modified the highest reliability factors will have their corresponding
RREQ. When the delay time is up, it forwards this RREQ RREP packets sent to the source.
packet. If the intermediate node receives a duplicate Figure 4 shows a detailed example that illustrates the route
RREQ, it will update the prim/aux routes in its recorded discovery process using the RAS routing protocol between
RREQ based on the LLR update rule (the node picks the the source node A and the destination node G.
one with the longest lifetime from the shortest routes as
the primary route, and it picks the one with the longest
lifetime from the second shortest routes as the auxiliary
route).
When destination node receives RREQ, it waits enough
time for receiving other RREQs and then it will unicast an
RREP packet to the source node using the primary route
with the auxiliary route attached, just as in the normal
reactive procedure.

2.5 Reliable Source Routing (RSR)

RAS is a reactive routing protocol that achieves increased
reliability through the maintenance of a reliability factor
by the nodes in the network [18].
During the route discovery process, the request messages
are only propagated to nodes with a reliability factor that is
above a threshold value specified by the application.
Fig. 4 An example of the route discovery process in RAS
Selection among multiple possible discovered routes is In this case, the required path to transmit the data has a
IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009 191

minimum reliability factor RMIN = 5, and MAX NH = 2. table driven routing strategy. A special and natural
Node A propagates the RREQ packet to nodes B and P. In switching mechanism is adopted in LRHR. LRHR mostly
turn, each of the nodes further propagates the RREQ works in the table-driven state. If all the following
packet according the above algorithm and accumulates the conditions are met, LRHR will transit from table-driven
reliability factor for each node in the path cumulative path state to on-demand state. In other words, the route
reliability factor. Finally, the RREQ packet, arrives at the discovery procedure will be initiated: I) Source node S has
destination node G with the discovered path A − B − H − I packet sending to destination node D, but S has no route to
− G and a final cumulative reliability factor of 27. D. II) the interval between the new route discovery and the
Similarly, a RREQ packet is propagated through nodes latest one is bigger than the minimum route request
A−P −N −T −O −G with a cumulative reliability factor of interval. (When there is no route to the destination node,
28. When the destination node G receives both RREQ instead of simply dropping the packet, LRHR will buffer
packets for the two discovered paths, it divides each path’s the packet and initiates route discovery)
final cumulative reliability factor with the number of In LRHR, the node operates in a promiscuous receive
nodes in the path to calculate the normalized path mode. In the promiscuous mode, a node can overhear a
reliability factor PRF = RCUM/n. Here, the destination packet that is not destined for it. Therefore, it can obtain
node G determines that the discovered paths A−B−H−I−G, more routes than a node that operates in the ordinary mode.
and A − P − N − T − O − G have normalized reliability LRHR mainly consists of the following modules: route
factors of 27/3 = 9 and 32/4 = 8 respectively. Therefore it discovery, route maintenance, route advertisement and
chooses the more reliable path A − B − H − I − G with the edge weight management. Figure 5 shows the relationships
higher normalized path reliability factor of 9. So, unicast a among these modules.
RREP packet back to the source node A along the
discovered intermediate nodes in the path list.
The reliability factor of a particular node is a measure of
its past performance in being a part of successful data
transmissions. There are several different algorithms that
can be used to maintain the reliability factor:
Self maintenance of the reliability factor by the node: Each
time a node successfully transmits a number of packets to
another node along the path it increases its reliability
factor.
Maintenance of the reliability factor by the Fig. 5 Framework of LRHR
acknowledgements: Using this policy, the destination
sends positive acknowledgments to the source. The
reliability factor is incremented when an intermediate node In LRHR, link reliability between nodes is quantified as an
receives positive acknowledgements from the destination. edge weight. (The link from one node to another is defined
Maintenance of the reliability factor by 1-hop neighbours: as an edge) Contrary to the well-known Dijkstra’s
Each node maintains a reliability factor for each of its n shortest-path algorithm, LRHR selects the route that has
neighbours. This set of reliability factors are a measure of the maximum edge weight sum as the primary route. In
the reliability of each of them in its view based on their this way, the selected route is of the maximum reliability.
past performance in successful data transmissions. In addition, unlike the traditional single path routing
Maintenance of the reliability factor using link-based strategy, multiple paths are established between a pair of
acknowledgement: If the next hop node did follow up and source-destination nodes.
forward the packet to its successor node, then it is In LRHR, each edge is assigned a value between 0 and 10.
rewarded by increasing its reliability factor in previous The value represents the reliability of the edge. (The
node. Otherwise, if that node did not forward the data, then greater the value, the higher the reliability) Edge weight
it is penalized by not increasing its reliability factor. management mainly consists of increasing and decreasing
the edge weight. Each new established edge is assigned the
weight of 10. If the edge is not used, with time elapsing,
2.6 Link Reliability based Hybrid Routing (LRHR) the edge reliability will be decreased gradually. So the
more frequently the edge is used, the higher the reliability.
LRHR is a hybrid routing protocol that contrary to the In the hybrid routing protocol LRHR, the link status
traditional single path routing strategy, multiple paths are database is periodically broadcasted to the neighbour
established between a pair of source-destination nodes [19]. nodes in route advertisement by one node. To reduce route
Hybrid routing protocols have to deal with the problem of advertisement overhead, LRHR divides route
how to switch between reactive on-demand and proactive advertisement into full route advertisement and partial
192 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009

route advertisement. The full route advertisement includes stable routes and uses power factor to determine active
the entire link status database, whereas, the partial route nodes to participate in routing.
advertisement only advertises the new edges. The period Selection of backbone nodes is made upon availability of
of partial route advertisement is comparably short. It is nodes and battery status and each route table has an entry
triggered on-demand in some sense. If no route for number of backbone nodes attached to it and their
advertisement is received in the predefined time interval battery status.
from a certain node X, node X is considered to be in the The proposed scheme is explained with the help of an
inactive state. example shown in Figure 7.
When each node forwards a packet (the packet may be the
data packet, route request or route reply), it can obtain
routing information from the source route list or route
reply list carried by the packet. Then the route table can be
updated. This is an active learning procedure. Also,
because the node operates in the promiscuous receive
mode, a node can overhear a packet that is not destined for
it. In this way, the node can obtain route information from
the overheard packet. This is a passive learning procedure.
As wireless networks are inherently less reliable than
wired networks, LRHR utilizes a hop-by-hop
acknowledgment to provide detection and retransmission
of lost or corrupted packets. After sending a packet to the
next hop node, the sender may be able to hear that node
transmitting the packet again, on its way further along the
path, if it operates in promiscuous receive mode. For Fig. 7 An Example of routing with SBNRP
example, in Fig.6, node X may be able to hear Y’s
transmission of the packet to Z. This type of Assume that the node with index 1 is the source while
acknowledgment is known as implicit acknowledgment. destination is the node with index 4. If the node with index
In the route table, there exists more than one path from a 3 is having power status in critical or danger zone, then
source to a destination. In general, two packets generated though the shortest path is 1—2—3—4 but the more stable
for a common destination by a source may be routed path 1—2—5—8—9—10—4 in terms of active power
through different paths. After arranging these paths status is chosen. This may lead to slight delay but
according to the edge weight sum along the path in improves overall efficiency of the protocol by sending
descending order, the path along which the edge weight more packets without link break than the state when some
sum is the maximum, called primary path. The other paths node is unable to process route due to inadequate battery
are called backup paths. power.
The process also helps when some intermediate node
moves out of the range and link break occurs, in that case
backbone nodes take care of the process and the route is
established again without much overhead. In Figure7 if the
node with index 8 moves out, the new established route
will be 1—2—5—11—9—10— 4. Here the node with
index 11 is acting as backbone node (BN) for the node
with index 5 and the node with index 8. Similarly the node
with index 12 can be BN for the nodes with indices 7, 10
and 4. BN has been selected at one hop distance from the
said node.
Fig. 6 Implicit acknowledgment The protocol is divided into three phases. Route Request
(REQ), Route Repair (REP) and Error Phase (ERR).
2.7 Stable and Energy efficient Routing (SBNRP) REQ phase is like route discovery operation of reactive
routing protocols. (Query and Reply procedure)
SBNRP is a reactive on demand routing protocol that uses An Intermediate node with an active route (in terms of
a new concept of backbone nodes with power factor [20]. power and Backbone Nodes), upon receiving a new RREQ
In this protocol two different concepts have been joined packet, records the previous hop and the source node
together to make an efficient protocol. (Stability and information in its route table i.e. backward learning. It then
Energy Efficiency) The scheme uses backbone nodes for broadcasts the packet or sends back a RREP packet to the
IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009 193

source if it has an active route to the destination. sessions or file transfers [22].
The destination node sends a RREP packet via the selected Under the RFAR scheme, the concept is to select a route
route when it receives the first RREQ or subsequent that will stay intact (with sufficiently high probability) for
RREQ packets that traversed a more stable route. Stability a certain period of time. This time duration represents the
of route depends upon two major aspects as: Lifetime and time it takes to transport across the network the data
Power status. associated with a transaction that is executed by an
Taking advantage of the broadcast nature of wireless underlying time critical application.
communications, a node promiscuously overhears packets For the calculation of the robustness indices along a route
that are transmitted by their neighbouring nodes. When a the RREQ packet of RFAR contains in its header Route
node that is not part of the route overhears a RREP packet Vulnerability Index (RVI) and underlying session holding
not directed to itself transmit by a neighbour (on the time H. At source node s, RVI set to 0.In each node the
primary route), it records that neighbour as the next hop to value of RVI updated according to the speeds of that node
the destination in its alternate route table. From these and the previous node and the effective communications
packets, a node obtains alternate path information and range(at an acceptable bit-error –rate level).
makes entries of these backbone nodes (BN) in its route In this protocol, a node receiving such a RREQ packet will
table. proceed to forward (flood) it to its neighbours (if it has not
When a node detects a link break, it performs a one hop yet forwarded such a RREQ packet) only if:
data broadcast to its immediate neighbours. The node a) Robustness control: it is desired for the selected
specifies in the data header that the link is disconnected route to exhibit a lifetime longer than the
and thus the packet is candidate for alternate routing. expected session holding time.
Upon receiving this packet route maintenance phase starts b) Capacity control: The link capacity (or delay, or
by selecting alternate path and checking power status. queue-size) level that it can offer to packets to be
During local repair data packets will be buffered at local generated by this flow, if admitted, is acceptable.
originator. Upon receipt of RREQ packets, the intermediate or the
If, at the end of the discovery period, the repairing node destination node waits for a prescribed period of time in
has not received a reply message REP it proceeds in by collecting several, if any, RREQ packets. It then proceeds
transmitting a route error ERR to the originating node. to examine the cumulative robustness status indices carried
in the header of these packets (indicating the
2.8 Cross-Layer Reliable Routing (CLRR) corresponding expected lifetime of the route travelled by
each RREQ packet), as well as other indices, such as the
CLRR is a cross-layer reactive routing protocol that used hop-length of the travelled route.
received signal strength information to choose reliable The destination node then, for example, selects an RREQ
links to form stable routes [21]. packet that provides acceptable robustness level while
In this cross-layer design, the received signal strength is offering the shortest route among all such received RREQ
measured at physical layer and transferred to the top layers. packets.
In CLRR, each node stored/ updated the received signal In the destination node, the selected RREQ packet induces
strength (RSS) of RREQ packet in Neighbour Table (NT) the issue of a route reply (RREP) packet that is transported
of routing protocol, against the address of the towards the source node across the selected route,
neighbouring node from which the RREQ is received. configuring at this phase the forwarding entries at the
Hence, whenever a node receives RREQ packet, by using nodes located along the route.
the current received signal strength and the previous value In the RFAR, if no acceptable route is discovered or
that is available in neighbour table, the moving direction calculated, the admission of the flow is rejected or blocked
of sending and receiving nodes can be found. That is, if by the source node.
current value is greater than previous one, the nodes are So if a route is selected in this protocol, the route lifetime
approaching; else they are moving apart. is sufficiently long to carry, with high probability, the
The RREQ packet forwarding decision is based on flow’s transaction without early interruption.
whether the sending and receiving nodes are approaching In addition, the selected route should offer the admitted
or moving away, in addition to having an adaptive signal flow packets with sufficient capacity resources so that
strength threshold with reference to moving speed of the prescribed (per application type) end-to-end packet delay
nodes. (mean and jitter) levels are met.

2.9 Robust Flow Admission and Routing (RFAR)

RFAR is a reactive on-demand robust routing protocol
selectively discovers routes that are probabilistically
assured to survive for the duration of the underlying
194 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009

ratio is important as it describes the packet delivery

efficiency of the routing protocol.
3. Protocols Comparison Less time for route discovery, reduced no. of packets
dropped and lower route breakup rates are considered
Table 1 summarizes the major features of the reliability enhancements too.
aforementioned reliable unicast routing protocols. Here the On the other hands, reduced routing load, less control
protocols are discussed and compared according to their overhead, reduced path discovery overhead and Reduced
features. routing overhead are all about the reduction of the network
In the table, the first column shows the type of each overhead of the protocols.
protocol. As it can be seen, most of the protocols are Balanced energy consumption is the only different
reactive on-demand protocols. enhancement that exists in EARR. The good result of this
Simple implementation and low control overhead are enhancement is prolonging the network lifetime.
important advantage of reactive type. But as it is Finally in the last column, it is shown that which protocols
mentioned in the introduction, when wireless network size is cross-layer.
increases, flat routing schemes (like reactive protocols) Basically, the cross-layer design enables the network
become infeasible. protocols and the applications to observe and respond to
LRHR is the only hybrid method that can be used for large the changing network and channel conditions. So, in order
scale networks. Also MSA is a protocol framework that to meet the changes in mobile ad hoc networks topology,
can be used by all types of protocols, so it can be applied the various layers can be considered together, which leads
with all kind of networks. to cross-layer designs.
The second column of the table shows the most important
information about the protocols. The reliability metric
determines how the protocols select a route from the 4. Conclusion
existing routes between source and destination node. This This paper review and compare some recent published
is the main difference between protocols and each protocol reliable unicast routing protocols for mobile ad hoc
usually uses different metric. networks, that can be used for selecting appropriate
Network size, density of nodes, mobility level and the routing protocols which are suitable for a particular
type of nodes and applications must be considered when application. Furthermore it maximizes the reliability of the
the suitable metric for a network is selected. deployed network. As a conclusion, a perfect reliable
The protocols select single path or multipath? It is shown unicast routing protocol for mobile ad hoc networks is
in the third column. difficulty of design. Each protocol has its own advantages
If notice to more storage, that is used in multipath for and disadvantages. A suitable reliable unicast routing
routing table, these protocols are better choices for protocol should be chosen based on network conditions
reliability. The routing protocol can switch to the backup and application demands. The protocols comparison table
routes to maintain the flow of traffic when the main provides a guide line for such choice. Also the proposed
primary route breaks. So, the delay reduces very well. methods used in the reviewed protocols, can help
In the fourth column, assumptions or requirements of each designers and researchers for enhancing the reliability of
protocol is showed. their new routing protocols.
Before implementing the protocols this items must be
prepared. As it can be seen, some of the protocols have no
assumption or requirement that means it can be used
directly.
Next column of the table, determine the network
operation enhancements for each protocols.
The information showed in this column is obtained from
the paper related to each protocol and verified by using a
simulator and comparison with other routing protocols.
Although the main goal of all protocols is increasing the
reliability of the network, some of them have other
enhancement too.
Higher packet delivery ratio is the main reliability
enhancement of almost all of the protocols. Packet
delivery ratio is the ratio between the number of packets
received by the destination node and the number of
packets originated by the source node. The packet delivery
IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009 195

Table 1: Reliable unicast routing protocols features

Single path Cross-

Assumptions and Network Operation
Protocol Type Reliability Metric or Layer
Requirements Enhancements
Multipath Design
Residual energy
Less time for route discovery,
detection device,
EARR Reactive Node residual energy Single path Reduced routing load, Balanced Yes
Energy consumption
energy consumption
model
Link expiration time, Less control overhead, Reduced
Global Positioning
SWORP Reactive Link failures caused Single path no. of packets dropped, Higher No
System(GPS)
by a node packet delivery ratio
Can be
Beaconing
MSA used with Stability of nodes Single path Higher packet delivery ratio No
mechanism
all types
Double path
Probabilistic link
DLLR Reactive ( primary ---------------- Higher packet delivery ratio No
lifetime estimation
and backup)
Node successful data Reduced path discovery
RAS Reactive Multipath ---------------- No
transmissions overhead
LRHR Hybrid Link usages Multipath ---------------- Higher packet delivery ratio No
Probabilistic link
Node battery power
SBNRP Reactive lifetime, Node power Single path Higher packet delivery ratio Yes
detection
status
Received signal Receiver signal Higher packet delivery ratio,
CLRR Reactive Single path Yes
strength strength measurement Reduced routing overhead
Probabilistic link
RFAR Reactive lifetime, Link Single path ---------------- Lower route breakup rates No
capacity

References IEEE Workshop on, 1999, p90 - p10.

[8] Z. J. Haas. “The Zone Routing Protocol (ZRP) for ad hoc
networks”, Internet Draft, Nov. 1997.
[1] J. P. Macker and M. S. Corson, "Mobile Ad Hoc
[9] M. Joa-Ng and I-Tai Lu, “A peer-to-peer zone-based two-
Networking and the IETF", ACM SIGMOBILE Mobile
level link state routing for mobile ad hoc net-works”, IEEE
Computing and Communications Reviews, Vol. 2, No. 2,
on Selected Areas in Communications, vol. 17, no. 8, pp.
pp. 9-14, January 1998.
1415 1425, 1999.
[2] Changling Liu, Jörg Kaiser , "A Survey of Mobile Ad Hoc
[10] Navid Nikaein, Christian Bonnet and Neda Nikaein,
network Routing Protocols", the University of Magdeburg ,
“HARP - Hybrid Ad Hoc Routing Protocol”, in proceeding
October 2005 .
of IST 2001: International Symposium on
[3] Murthy, S. and J.J. Garcia-Luna-Aceves, “An Efficient
Telecommunications, Iran/Tehran 2001.
Routing Protocol for Wireless Networks” , ACM Mobile
[11] Y. B. Ko and N. H. Vaidya. “Location Aid Routing (LAR)
Networks and App. J., Special Issue on Routing in Mobile
in mobile ad hoc networks”, In Proc. ACM/IEEE
Communication Networks, Oct. 1996, pp. 183-97.
MOBICOM, Oct. 1998.
[4] C. E. Perkins and P. Bhagwat, “Highly dynamic
[12] S. Basagni, I. Chlamtac, V. Syrotiuk and B. WoodWard,
Destination-Sequenced Distance-Vector Routing (DSDV)
“A Distance Routing Effect Algorithm for Mobility
for mobile computers”, ACM Computer Communication
(DREAM)”, Proc. 4th MOBCOM, 1998.
Review, Vol. 24, No.4, (ACM SIGCOMM’94) Oct. 1994,
[13] I. Broustis, G. Jakllari, T. Repantis, and M. Molle. “A
pp.234-244.
Performance Comparison of Routing Protocols for Large-
[5] G. Pei, M. Gerla and T.-W. Chen, “Fisheye State Routing
Scale Wireless Mobile Ad Hoc Networks”, Technical
in Mobile Ad Hoc Networks”, In Proceedings of the 2000
Report, University of California, Riverside, 2006.
ICDCS Workshops, Taipei, Taiwan, Apr. 2000, pp. D71-
[14] Fang Xie, Lei Du, Yong Bai, Lan Chen, “Energy Aware
D78.
Reliable Routing Protocol for Mobile Ad-Hoc
[6] D. Johnson, D. A. Maltz, “Dynamic source routing in ad
Networks”,IEEE Communication Society,WCNC 2007
hoc wireless networks”, in Mobile Computing (T.
proceedings.
Imielinski and H. Korth, eds.), Kluwer Acad. Publ., 1996.
[15] Nen-Chung Wang, Yung-Fa Huang, Jhu-Chan Chen, “A
[7] C.E. Perkins and E.M. Royer. “Ad hoc on demand Distance
stable weight-based on-demand routing protocol for mobile
Vector routing”, mobile computing systems and
ad hoc networks”, Elsevier Information Sciences 177,
applications, 1999. Proceedings. WMCSA ’99. Second
196 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.1, January 2009

2007, p 5522–5537. research interests are mainly management aspects of high-speed

[16] Athanasios Bamis, Azzedine Boukerche, Ioannis and wireless Networks, including the area of topology
Chatzigiannakis, Sotiris Nikoletseas, “A mobility aware discovering, quality of service (QoS) routing, Wireless Ad-Hoc
protocol synthesis for efficient routing in ad hoc mobile and Wireless Mesh Networks. Also swarm Intelligence is his
networks”, Elsevier Computer Networks 52, 2008 p130– specially interest research area.
154.
[17] Zhao Cheng , Wendi B. Heinzelman, “Discovering long
lifetime routes in mobile ad hoc networks”, Elsevier Ad Associate Professor Dr. Sabira
Hoc Networks 6, 2008, p661–674. Khatun is presently working at the
[18] Imad Jawhar, Zouheir Trabelsi, and Jameela Al-Jaroodi, Department of Computer &
“Towards More Reliable Source Routing in Wireless Communication Systems Engineering,
Networks”, IEEE International Conference on Networking, Faculty of Engineering, and University
Architecture, and Storage, 2008. Putra Malaysia. She is head of
[19] Xie Xiaochuan, Wei Gang, Wu Keping, Wang Gang, Jia Electronics and Communications
Shilou, “Link reliability based hybrid routing for tactical Laboratory and Research Advisor of
mobile ad hoc network”, Elsevier Journal of Systems Networks Research Group in the
Engineering and Electronics Vol. 19, No. 2, 2008, pp.259– department. She received her PhD in
267. communications and Networking from University Putra Malaysia
[20] Vinay Rishiwal, Ashwani Kush, Shekhar Verma, “Stable (UPM) in 2003.
and Energy Efficient Routing for Mobile Adhoc Networks”, She directs research activities within the Wireless
IEEE Fifth International Conference on Information Communications and Networks (WCN) group and her work in
Technology: New Generations, 2008, p1028-1033. Mobile IPv6 (MIPv6) is gaining importance in Internet-based
[21] B. Ramachandran, S. Shanmugavel1, “Received Signal applications so as to offer secure uninterrupted on-line
Strength-based Cross-layer Designs for Mobile Ad Hoc experience in interactive network applications such as in
Networks”, IETE technical review, vol 25, issue 4, 2008, entertainment, games, video conferencing or video streaming
p192-200. while on the move and has had worldwide publicity. She
[22] Izhak Rubin, Runhe Zhang, “Robust throughput and currently leads a few research projects on WCN for mobile
routing for mobile ad hoc wireless networks”, Elsevier Ad internet, medical and 3G/4G applications including UWB,
Hoc Networks 7, 2009, p 265–280. Combined C-SDMA and SDR. She was the proud recipient of
Best Inventor of the Asia Pacific Rim 2006, Korea Invention
Promotion Association (KIPA) Special Award 2006 for
commending excellent efforts to create micro-mobility related
Hadi Sargolzaey received the B.S. MIPv6 invention and Excellent Research Award 2006 from
in Electronic Engineering and M.S. Ministry of Higher Research, Malaysia.
degrees in Digital Electronic She is an active researcher of Teman project and MyREN
Engineering from Ferdowsi University Research Community. She is a member of IEEE and her research
of Mashhad and Sharif University interest spans Broadband and Wireless Communications
respectively. Currently he is PhD including Cognitive Radio, Software Defined Radio (SDR),
student in Communication Network MIMO, UWB and IPv6.
Engineering, University Putra
Malaysia. He has teaching experience
in data communication network for
more than 13 years in Islamic Azad University of Qazvin. His
research interests are mainly on wireless data communication
networks and microprocessor systems.

Ayyoub Akbari Moghanjoughi

received his B. Eng from IAU
University (2006) in Iran and currently
he doing his M.Sc. of communication
network engineering in Faculty of
Engineering at University Putra
Malaysia. Also he is postgraduate
research assistant there. He was
professional inspector of ICT training
Centers for state of Yazd in Iran
(2004-2006). Currently he is a member of International
Association of Engineers (IAEng), International Engineering
Consortium (IEC) and IEEE member of Computer,
Communication and Computational Intelligence societies. His