Bandwidth aware on demand multipath routing in manets

International Journal of Wireless & Mobile Networks (IJWMN) Vol. 6, No. 4, August 2014
BANDWIDTH AWARE ON DEMAND
MULTIPATH ROUTING IN MANETS
Tripti Sharma1 and Dr. Vivek Kumar2
1Department of computer science & Engineering, Inderprastha Engineering College,
Ghaziabad, (U.P), India.
2Department of computer science, Gurukul Kangri Vishwavidyalaya, Haridwar, (U.K),
India.
ABSTRACT
Mobile Ad-hoc Networks (MANETs) are self configuring, decentralized and dynamic nature wireless
networks which have no infrastructure. These offer a number of advantages, however the demand of high
traffic flows in MANETs increases rapidly. For these demands, limited bandwidth of wireless network is the
important parameter that restrains the development of real time multimedia applications. In this work, we
propose a solution to utilize available bandwidth of the channel for on demand multiple disjoint paths. The
approximate bandwidth of a node is used to find the available bandwidth of the path. The source chooses
the primary route for data forwarding on the basis of path bandwidth. The simulation results show that the
proposed solution reduces the frequency of broadcast and performs well in improving the end to end
throughput, packet delivery ratio, and the end to end delay.
KEYWORDS
MANET, node-disjoint, multipath, bandwidth.
1. INTRODUCTION
Many characteristics of ad hoc networks make QoS (Quality of Service) provisioning and QoS
routing, a difficult problem. QoS routing means not only to find a route from source to
destination, but to find a quality route that satisfies the end-to-end QoS requirement, often given
in terms of bandwidth, delay or loss probability. The single path routing protocols like DSDV and
DSR, normally fail to fulfill the above requirements. The dynamic topology of MANETs provides
the existence of multiple routes between two nodes, which can be utilized to transmit the packet
for better support to real time communications. In case of route break, an alternate route can be
used to send the packets to reduce the delay and jitter. Research shows that the use of multipath
routing in ad hoc networks which are denser performs better throughput. In this paper, we are
proposing a multipath routing protocol, which is the potential improvement of the existing Ad hoc
On-demand Multipath Distance Vector (AOMDV) protocol [1], and that could be achieved when
utilizing the bandwidth of the channel and bandwidth of the respective paths.
The remaining part of the paper is organized as follows: the next part is the review of the
protocols and methodologies in the required fields of MANET. Then, we present the problems
and motivations. Since, we have modified the existing AOMDV routing protocol, it is also
discussed briefly, with its problems. And then we propose the improvements in AOMDV
followed by the simulation results of the comparison of new protocol with the existing one. In the
last section, we provide the conclusion and future scope of the proposed work.
DOI : 10.5121/ijwmn.2014.6408 101

102
2. RELATED WORK
In [3-13], various approaches to QoS provisioning and QoS routing for single path in MANETs
have been studied and derived with the aim to reduce the connection set up latency, delay and
bandwidth and to ensure guaranteed performance level to the QoS sensitive applications.
Multipath routing is more promising in ad hoc networks since it provides additional features like
load balancing, fault-tolerance, higher throughput etc., to ensure QoS assurance in ad hoc
networks.
Zhi Zhang, et al., [14] performs bandwidth estimation method with the on-demand node-disjoint
multi-path routing protocol. This approach creates the multiple node-disjoint paths during the
route discovery process and maintained those paths actively. The detector packets measure the
available bandwidth of each hop along the paths. AOMDV uses the basic idea of the popular ad
hoc on-demand distance vector (AODV) which is proposed in [2]. AOMDV extends the AODV
protocol to find the multiple paths in the route discovery process without discarding those paths.
These all multiple paths are guaranteed to be loop-free and disjoint. AOMDV has three important
issues compared to other on-demand multipath routing schemes. Firstly, like some other protocols
(e.g., TORA, ROAM [15-16]), have high coordination overhead among nodes, but its inter-nodal
coordination overheads is less. Second, the disjointness of alternate routes is guaranteed via
distributed computation without the use of source routing. Finally, this protocol computes
alternate paths over AODV with very less additional overhead; it does this by utilizing the
information which is already available with alternate paths as much as possible. There are a
number of extensions of AOMDV in various fields including dynamic route switching, stability,
load-balancing and randomization.
As the AOMDV is based on static route selection, it could not handle the change of the network
such as congestion and contention. D. Shin et al., [17], proposed A2OMDV (Adaptive AOMDV),
in this approach author resolve the problem of AOMDV, through dynamic route switching
method. A source node finds its route dynamically based on the delay of the multiple paths and
observes the quality of the alternative routes according to the change of the ad hoc network. One
idea is to accept partially disjoint paths that are more stable than other maximally disjoint paths
that could increase paths lifetime. Stability-based Partially Disjoint AOMDV (SPDA) protocol is
proposed in [18], which is a modification of the AOMDV protocol, finds partially disjoint paths
based on links stability. These Partially Disjoint paths improves MANET performance in terms of
delay, routing packets overhead, and the network throughput. M. Tekaya, et al., [19], also,
introduced a multipath routing protocol with load balancing mechanism, to develop a new
protocol called QLB-AOMDV (QoS and Load Balancing- Ad Hoc On demand Multipath
Distance Vector), with this solution we can achieve better load balancing with respect to the end-to-
end QoS requirement. A multipath routing algorithm is proposed by Pinesh A Darji, et al. in
[20], that could randomize delivery paths for data transmission also it uses secured traffic load
based on some cryptography approach, in which, randomized paths can protect data from the
intrusion of malicious nodes. In [21], author proposed an adaptive retransmission limits algorithm
for IEEE 802.11 MAC to reduce the false link failures and predict the node mobility. Since the
probability that neighbour node is still in transmission range and may be not responding due to
some problems other then mobility is maximum. In this approach the signal strength of each node
in network is monitored and, while performing transmissions to a neighbour node, if it’s received
signal strength is raised and is received recently then Adaptive MAC persists in its retransmission
attempts.

103
2.1 Overview of AOMDV
Ad-hoc On-demand Multi path Distance Vector Routing (AOMDV) protocol is an extension to
the AODV protocol for computing multiple loop-free and link disjoint paths. The routing entries
for each destination contain a list of the next-hops along with the corresponding hop counts.
Multiple paths maintained at each node for each destination have the same destination sequence
number which helps in keeping track of the route. For each destination, a node maintains the
advertised hop count, which is defined as the maximum hop count for all the paths, which is used
for sending route advertisements of the destination. The duplicate route advertisement received by
a node defines an alternate path to the same destination. Loop freedom is assured for a node by
accepting alternate paths to destination if it has a less hop count than the advertised hop count for
that destination. Because the maximum hop count is used, the advertised hop count therefore does
not change for the same sequence number. When a route advertisement is received for a
destination with a greater sequence number, the next-hop list and the advertised hop count are
reinitialized.
AOMDV can be used to find node-disjoint or link-disjoint routes. To find node-disjoint routes,
each node does not immediately reject duplicate RREQs. Each RREQs arriving via a different
neighbor of the source defines a node-disjoint path. Since the nodes cannot broadcast duplicate
RREQs, so any two RREQs arriving at an intermediate node through a different neighbor of the
source could not have traversed the same node. In an attempt to get multiple link-disjoint routes,
the destination replies to duplicate RREQs, the destination only replies to RREQs arriving via
unique neighbors. After the first hop, the RREPs follow the reverse paths, which are node disjoint
and thus link-disjoint. The trajectories of each RREP may intersect at an intermediate node, but
each takes a different reverse path to the source to ensure link disjointness. Using AOMDV
protocol is advantageous since it allows intermediate nodes to reply to RREQs, while still
selecting disjoint paths. One of the drawbacks of AOMDV is, it has more message overheads
during route discovery due to increased flooding and since it is a multipath routing protocol, the
destination replies to the multiple RREQs that results are in larger overhead.
2.2 Problems And Motivations
It is concluded from above discussion, that a reactive routing protocol generates a large number of
overhead control messages in the network during route discovery process. So, routing with QoS is
difficult in MANET due to several reasons like high overhead, dynamic nature of MANETs, and
guarantee of reserved resources.
The existing AOMDV protocol has given the improved results compared to AODV. As it does
not have large inter-nodal coordination overheads, it provides disjoint alternate routes, and these
are with minimal additional overhead over AODV. Still, there are some problems in AOMDV
which are considered in various modifications of it. Those modifications have been discussed in
section 2. The modified protocols have resolved the problems of AOMDV up to some extent, but
none of them has considered the resource utilization of the MANET. There are various problems
in AOMDV extensions, these are:
- None of the protocol has given the strategy to prioritize the alternate routes for resource
utilization.
- The selection of the alternate routes performed without comparison of performances.
- The existing protocols are not effective to compute more disjoint paths between source and
destination pairs while considering the effect of other resources on performance of the
network.

104
3. THE PROPOSED PROTOCOL
In this section, we propose an extension to AOMDV protocol, in which the channel bandwidth is
utilizes in order to improve the network performance. AOMDV allows finding many routes
between source and destination during the same route discovery procedure that guarantees loop
freedom and disjointness of alternate paths, and only one path is used to transmit data. The
routing table entry structure of AOMDV is modified for the proposed method in which only one
field is added which gives the information about the path bandwidth of the multiple paths stored
in route list entries and is given in Figure 1.
Figure 1. Routing table entry structure of proposed protocol
3.1 Route discovery
The route discovery procedure of the proposed method constructs multiple bandwidth aware paths
between a source and destination. Route request and Route reply packets now contain the existing
information and the available bandwidth of the node forwarding it. The source is able to learn the
bandwidth of the multiple paths during the route discovery by using the Maximum-Minimum
approach to measure the quality of the path. In this approach, the available bandwidth of the
entire path is just the available bandwidth of the weakest link. Once the source receives the
RREPs, it stores its next hop information and chooses the path with the greatest available
bandwidth as its primary path for data transmission. The bandwidth of the route is determined
periodically in order to find the optimal route in the change of the network topology with the help
of detector packet as explained in section 3.4. The source node will switch from its current
primary path to an alternate path if the difference in their available bandwidth is higher than the
predefined threshold in contrast to wait for its primary path to break. In Figure 2, the AOMDV
route decision procedure is summarized, which is modified for the proposed solution, and is given
in Figure 3.
If (no route to destination)
{
Initiate route discovery as in AODV;
}
If (single known route)
{
Forward data packet to specified route;
}
Else //if N routes are known from source to destination.
{
Forward data packet to best route;
// on the basis of minimum hop count.
}
Figure 2. AOMDV route decision

105
If (no route to destination)
{
Initiate route discovery as in AOMDV;
}
If (single known route)
{
Forward data packet to specified route;
}
Else
// if N routes are known from source to destination
{
Forward data packet to route with max. available bandwidth ;
}
Figure 3. Modified AOMDV route decision
We also propose to modify the route request as well as route reply packet, which are given below:
RREQ(proposed) = RREQ(AOMDV) + ABW (node)
where,
ABW (node) = MIN [ABW (RREQ recieved) , Bavail (node)]
In the proposed RREQ packet, the Bavail of node is the available bandwidth of the node sending
the packet. The estimation of Bavail is discussed in section 3.3. Upon receiving the packet, each
node will compare its own available bandwidth with the bandwidth received in the packet, and
then update the packet with the minimum bandwidth. Once the destination receives the RREQ, it
generates route reply packet ie, RREP which is also modified in the similar manner.
3.2 Hello message
The Hello packet used in AOMDV only keeps the address of the node which has generates this
packet. We modify the Hello packet for the new solution by adding the bandwidth information of
the node sending the hello packet and the neighbors of the node with their bandwidth information.
Each node broadcast this hello packet periodically, and updates all its neighbors about its
bandwidth. The format of hello packet is given in Figure 4. Where Bconsumed is the bandwidth
consumed by each node for sending packets in the network.
<SenderID, Bconsumed, timestamp> <neighborsID, Bconsumed, timestamp>
Figure 4. Format of Hello message
3.3 Bandwidth Estimation
For the forwarding of data packets where multiple routes are known from source to destination,
the maximum available bandwidth of the routes is estimated. Each node estimates its consumed
bandwidth by tracking the packets it transmits into the network. This value is recorded in the
bandwidth consumption register at the node and updated periodically. Once a node knows the
bandwidth consumption of its one-hop neighbors and its two-hop neighbors, the residual
bandwidth can be estimated as (1), the raw channel bandwidth (Braw) minus the overall consumed
bandwidth (Ball_consumed), multiplied by a weight factor. We need to multiply the residual
bandwidth by a weight factor due to overhead of IEEE 802.11 MAC, overhead of routing
protocol and overhead for the situation where a node is in sender’s interference range but it isn’t
in any of sender’s neighbors’ transmission range [14]. In this situation, the sender will never
know this node bandwidth usage. However, these instances do not happen frequently since it has

International Journal of Wireless Mobile Networks (IJWMN) Vol. 6, No. 4, August 2014
to meet strict requirements. So weight factor is used to overcome this situation. From the equation
in (1), the more interference traffic in the channel the more conservative the estimation will be.
106
Bavail = (Braw - Ball_consumed ) (1)
where, 0 0 .8865
3.4 Alternate Route Maintenance
The alternate routes constructed between a pair of source and destination is to be maintained for a
time period. The algorithm for alternate route maintenance is given below. The detector packet is
unicast from source to destination along the alternate paths. This packet contains one field apart
from source and destination addresses that is to collect the minimum bandwidth along each path.
Algorithm :
1. Source node periodically sends detector packet to the destination along each of its
alternate paths after route discovery.
2. Each node updates the bandwidth field when the detector propagates through the
alternate paths.
3. The destination records the bandwidth in the detector and sends a new detector back
to the source along the same path.
4. The bandwidth of the entire path is just the bandwidth of the weakest link.
5. The source node chooses the path with the maximum bandwidth for routing.
6. The source node will switch from its current primary path to an alternate path if the
difference in their ABW is higher than the predefined threshold in contrast to waits
for its primary path to break.
4. SIMULATION RESULTS
We study the new AOMDV performance using ns-2 [22, 23] simulations. The main objective of
our simulation is to evaluate the effectiveness of new AOMDV relative to AOMDV in the
presence of mobility-related route failures. Other objective includes evaluating the number of
alternate node disjoint paths that can be found using new AOMDV.
4.1 Simulation Environment
The simulation experiment is carried out in LINUX (ubuntu 10.4). The detailed simulation model
based on network simulator-2 (ver-2.35), is used in the evaluation. Table 1 shows the simulation
parameters. In this simulation, each packet starts its journey from a random location to a random
destination with a randomly chosen speed (uniformly distributed between 0–20 m/s). Simulations
are run for 50s, 100s, 150s, 200s, 250s and 300s simulated for 100 nodes under CBR traffic
pattern. The weight factor is defined as 0.65.
Table 1. Simulation Parameters
Parameter Values
Dimensions 1000m×1000m
Traffic type CBR
Number of nodes 100
Simulation Time 300s

107
Pause Time 50, 100, 150, 200, 250,
300s
Total Sources and
Connections
49 and 71
Maximum Speed of Nodes 20m/s
Packet rate 4pkts/s
Packet size 512 byte
Mobility model Random Waypoint Model
Channel bandwidth 2Mbps
4.2 Performance Metrics
The performance of routing protocol is evaluated using three different metrics to compare the
performance of the new protocol with the existing AOMDV routing protocol. They are:
1. Packet delivery fraction (PDF) — The packet delivery fraction is the ratio of the data
packets delivered to the destinations to those generated by the sources.
2. End-to-end delay (E2E Delay) — The end-to-end delay of data packets refers to the
time taken for a packet to be transmitted across a network from source to destination.
3. Normalized routing load (NRL) — The Normalized routing loads is computed by the
ratio of total number of routing packets sent by the number of data packets delivered
successfully.
4. Throughput (THPT) — The throughput is the amount of data packets received at the
destination per unit time.
The performance results of AOMDV and B-AOMDV for 100 nodes and the comparison of new
protocol with the existing AOMDV protocol are given below in Table 2 and 3.
Table 2. Performance Results of AOMDV and B-AOMDV for 100 nodes
Simulation
time (s)
B-AOMDV AOMDV
PDF
(%)
E2E Delay
(s)
NRL THPT PDF
(%)
E2E
Delay (s)
NRL THPT
50 90.95 0.01301 2.636 653.11 93.33 0.022 2.908 587
100 92.43 0.09289 1.502 958.33 95.206 0.0208 1.587 1027
150 79.52 1.82 1.261 1300 75.25 1.35 1.95 1295
200 64.27 3.67 1.21 1516.35 59 3.35 3.403 1490
250 56.063 5.26 1.258 1633.09 50 4.86 4.49 1629.3
300 51.101 5.99 1.402 1706.63 42 5.96 4.96 1711.6

108
Table 3. Comparison Table
Metrics B-AOMDV AOMDV
PDF High Low
E2E
Low High
Delay
NRL Low High
THPT High Low
Packet Delivery Fraction: Figure 5 compares the packet delivery fraction of AOMDV and
proposed modification in varying pause time and random node speed. The graph demonstrates
that proposed modification performs better than the AOMDV at nearly all pauses of time. The
AOMDV perform well at less pause time but degrade at high pause time, while the proposed
protocol does not degrade too much. Higher packet delivery fraction of new protocol is because
of the availability of the bandwidth utilization among alternate paths to forward the packets when
the source switched from its primary path.
Figure 5. Packet delivery fraction
End to end delay: Figure 6 compares the End to end delay of AOMDV and proposed
modification in varying pause time and random node speed. The graph demonstrates that
proposed modification results in less delay than the AOMDV at nearly all pause time. The
AOMDV perform well at less pause time but delay increase at high pause time, while the
proposed protocol does not increase the delay at almost all pause time.

109
Figure 6. End-to-End Delay
Normalized Routing Load: Figure 7 compares the Normalized Routing Load of AOMDV and
proposed modification in varying pause time and random node speed. The graph demonstrates
that proposed modification results in less normalized routing load than the AOMDV. The
AOMDV perform well at less pause time but load increase at high pause time, while the proposed
protocol results high normalized load at less pause time but normalized routing load is decreased
as the pause tme is increased.
Figure 7. Normalized Routing Load

Throughput : Figure 8 compares the throughput of AOMDV and proposed modification in
varying pause time and random node speed. The graph demonstrates that proposed modification
results in high throughput than the AOMDV. The AOMDV performs well at less pause time but
at high pause time, the new protocol results high throughput as compared to AOMDV protocol.
110
Figure 8. Throughput
5 CONCLUSION AND FUTURE SCOPE:
In this paper, we proposed an approach for multi-path routing in mobile ad hoc networks and used
bandwidth estimation by disseminating bandwidth information through detector packets. The
primary property of this approach is that it can adapt the change in network topology by
proactively estimating the available bandwidth of each path to the destination and always using
the best path. Simulation results show that the performance of the protocol is superior to the
AOMDV in all most all scenarios. Future research will focus on optimally distributing traffic over
multiple paths to upgrade the performance of the protocol.
REFERENCES
[1] Marina, M. K., and Das, S.R. (2006) “Ad hoc on-demand multipath distance vector routing”, Wireless
Communications and Mobile Computing, vol. 6, no. 7, pp 969–988,.
[2] Perkins, C.E., Royer, E. (1999) “Ad-hoc On Demand Distance Vector Routing”, IEEE Workshop on
Mobile Computing Systems and Applications (WMCSA'99), pp 90-100,.
[3] Chakrabarti, S., Mishra, A. (2001) “QoS issues in Ad Hoc Wireless Networks”, IEEE Communication
Magazine, feb, pp 142-148,
[4] Lohier, S. et.al. (2003) “A reactive QoS Routing Protocol for Ad Hoc Networks”, European
Symposium on Ambient Intelligence (EUSAI'2003), LNCS, Nov, Springer Verlag, Netherlands, 875,
pp 27-41.
[5] Cerd`a, L., et. al. (2005) “A Reservation Scheme Satisfying Bandwidth QoS Constraints for Multirate
Ad-hoc Networks”, Eurasip.org, conference proceedings, IST, jun.
[6] Chen, L. and Heinzelman, W. B. (2005) “QoS-Aware Routing Based on Bandwidth Estimation for
Mobile Ad Hoc Networks”, IEEE Journal On Selected Areas In Communications, vol. 23, no. 3, pp
561-572.
[7] Geng , R. and Li, Z. (2006) “QoS-Aware Routing Based on Local Information for Mobile Ad Hoc
Networks”, S. Madria et al. (Eds.): ICDCIT, LNCS 4317 © Springer-Verlag Berlin Heidelberg, pp
159 – 167.

[8] Chen, L. and Heinzelman, W. B. (2007) “A Survey of Routing Protocols that Support QoS in Mobile
111
Ad Hoc Networks”, IEEE Network, vol. 21, no. 6, pp 30-38.
[9] Marwaha, S., Indulska, J. and Portmann, M. (2008) “Challenges and Recent Advances in QoS
Provisioning, Signaling, Routing and MAC protocols for MANETs”, IEEE conference on
Telecommunication Networks and Applications, pp 97-108.
[10] Sharma, N. et.al, (2010) “Provisioning of Quality of Service in MANETs Performance Analysis
Comparison (AODV and DSR)”, IEEE conference on Computer Engineering and Technology, vol. 7,
pp 243-248.
[11] Shenoy, N., Pan, Y. and Reddy, V.G. (2005) “Quality of Service in Internet MANETs”, IEEE 16th
International Symposium on Personal, Indoor and Mobile Radio Communications, pp 1823-1829.
[12] Medhi, D. (2002) “Quality of Service (QoS) Routing Computation with Path Caching: A Framework
and Network Performance”, IEEE Communications Magazine, vol. 40, no. 12, pp 106–113.
[13] Medhi, D. and Sukiman, I. (2000) “Multi-Service Dynamic QoS Routing Schemes with Call
Admission Control: A Comparative Study”, IEEE Journal of Network and Systems Management,
vol. 8, no.2, pp 157–190.
[14] Zhang, Z., Dai, G. and Mu, D. (2006) “Bandwidth-Aware Multipath Routing Protocol for Mobile Ad
Hoc Networks”, J. Ma et al. (Eds.): UIC LNCS 4159 © Springer-Verlag Berlin Heidelberg, pp 322 –
330.
[15] Park, V.D. and Corson, M.S. (1997) “A highly adaptive distributed routing algorithm for mobile
wireless networks”, Annual Joint Conference of the IEEE Computer and Communications Societies.
Proceedings IEEE, vol. 3, pp 1405–1413.
[16] Raju, J. and Garcia-Luna-Aceves, J. (1999) “A new approach to on-demand loop-free multipath
routing”, Proceedings of IEEE International Conference on Computer Communications and
Networks, pp 522 – 527.
[17] Shin, D., et. al. (2009) “A2OMDV : An Adaptive Ad Hoc On-Demand Multipath Distance Vector
Routing Protocol Using Dynamic Route Switching”, Journal of Engineering Science and Technology,
vol. 4, no. 2, pp 171 – 183.
[18] AlMobaideen, W. (2009) “SPDA: Stability Based Partially Disjoint AOMDV”, European Journal of
Scientific Research, vol. 27, no. 3, pp 342-348.
[19] Tekaya, M., Tabbane, N. and Tabbane, S. (2010) “Multipath Routing with Load Balancing and QoS
in Ad hoc Network”, International Journal of Computer Science and Network Security, vol. 10, no.
8, pp 280-286.
[20] Darji, Pinesh. A. et. al., (2012) “An Improvement in AOMDV with Randomization”, International
Journal of Computer Technology Applications, vol. 3, no. 3, pp 968-972.
[21] Bhatia, G. and Kumar, V., (2010) “Adapting MAC 802.11for performance optimization of MANET
Using Cross Layer Interaction”, International Journal of Wireless Mobile Networks (IJWMN)
Vol.2, No.4, pp 31-42.
[21] McCanne, S. and Floyd, S. ns Network Simulator, http: //www. isi.edu/nsnam/ ns.
[22] Fall, K. and varadhan, k. the VINT project. The NS manual.

Bandwidth aware on demand multipath routing in manets

More Related Content

What's hot

Viewers also liked

Similar to Bandwidth aware on demand multipath routing in manets

Recently uploaded

Bandwidth aware on demand multipath routing in manets