**5. Simulation results and discussion**

To evaluate our protocol, we carried out two simulations with different scenarios, as follows.

#### **5.1. Simulation 1:Compare our protocol with multi-radio routing protocol**

In this simulation, the efficiency of AODV-MRCR compared to AODV-MR [24] was evaluated using ns-2 [22]. Each mesh router is equipped with four wireless interfaces that statically turn to non-overlapping channels using the same channel allocation scheme. Concurrent UDP flows are established between the randomly selected source and the gateway while keeping the other parameters, as in Table 1. We run AODV-MRCR with twelve non-overlapping channels, and AODV-MR with four channels statically dedicated to the node interfaces. We carry out different scenarios to show the various factors affecting the performance of our protocol. The scenarios are: varying the number of flows, effect of number of radio per node, and evaluating the performance of the protocol using TCP traffic.

#### *5.1.1. Scenario 1: Varying number of the flow*

12 Will-be-set-by-IN-TECH

Traffic type Constant Bit Rate (CBR)

1. Packet Delivery Ratio(PDR): The ratio between the number of data packets successfully received by the destination nodes and the total number of data packets sent by the source

2. Aggregate goodput: The total number of application layer data bits successfully

3. End-to-end delay of data packets: The delay between the time at which the data packet originated at the source and the time when it reaches the destination. It includes all of the possible delays caused by queuing for transmission at the node, buffering the packet and

4. Routing overhead: The ratio of the total number of packets generated to the total number

5. Packet loss: The number of packets that were lost due to unavailable or incorrect routes,

To evaluate our protocol, we carried out two simulations with different scenarios, as follows.

In this simulation, the efficiency of AODV-MRCR compared to AODV-MR [24] was evaluated using ns-2 [22]. Each mesh router is equipped with four wireless interfaces that statically turn to non-overlapping channels using the same channel allocation scheme. Concurrent UDP flows are established between the randomly selected source and the gateway while keeping the other parameters, as in Table 1. We run AODV-MRCR with twelve non-overlapping channels, and AODV-MR with four channels statically dedicated to the node interfaces. We carry out different scenarios to show the various factors affecting the performance of our protocol. The scenarios are: varying the number of flows, effect of number of radio per node,

**5.1. Simulation 1:Compare our protocol with multi-radio routing protocol**

retransmission delays. This metric represents the quality of the routing protocol.

MAC layer collisions or through the saturation of interface queues.

and evaluating the performance of the protocol using TCP traffic.

Packet rates 20 packet per second

Simulation area 1000 <sup>×</sup> 1000 meter<sup>2</sup>

Simulation time 250 second

Transmission range 250 meter

Packet size 512 bytes

Number of nodes 100 Number of connection 50

The simulation provides the following five performance metrics:

transmitted in the network per second.

of data packets that are successfully received.

**5. Simulation results and discussion**

**Table 1.** Simulation parameters **4.1. Performance metrics**

nodes.

In this scenario, we investigate the impact of traffic load in the network. The number of generated flows varied from 10 to 50 flows with increments of 10. The packet size for each connection was fixed to 512 bytes with a 20 packet per second data rate. Fig. 4 (a-d) shows the packet loss, aggregated goodput, PDR and end-to-end delay performance. As observed from Fig 4, performance of the AODV-MRCR is comparable to that of the AODV-MR at a lower traffic load, such as 10 or 20. This is because AODV-MR can utilize the four channels to minimize the number of links on the same channel within the path as well as create channel diversity along the path.

**Figure 4.** Simulation 1: results for scenario 1 (Varying the number of flows).

However, when the flow increases, the network becomes saturated. Thus, the AODV-MR end-to-end delay and packet loss increase as shown in Fig. 4(a, d) while the PDR and aggregated goodput decrease (Fig. 4(b, c). This is because AODV-MR is unable to avoid the congestion area as it does not have an intelligent way to route the packet through a less congested area. Moreover, minimum channel diversity cannot achieve due to using the hop count as a metric. In contrast, 4 shows an improvement in the performance of the AODV-MRCR under increasing traffic load relative to the AODV-MR. In addition, it shows that the lower packet losses incurred by our protocol enables it to achieve a significantly higher packet delivery ratio.

This is because the protocol assigns a unique list of channels to every RREQ\_I received at the gateway during the route-establishing stage. Hence, the single collision domain divided to many collision domains as the number of flows increase.

Fig. 4(d) As the number of the flows increase, our protocol achieves better average end-to-end delay than the AODV-MR. This is because our reservation scheme allows the AODV-MRCR protocol to assign different non-overlapping channels for each node for the gateway traffic. Resulting in, reduces the contention time at the MAC layer as well as allowing the node to receive and send simultaneously. Moreover, It is interesting to note that at high traffic load such as 50 flows, the AODV-MR's end-to-end-delay (Fig. 4(d)) reduces compare to other flows such as 30 and 40 flows. The reason for that is at high traffic load the network is saturated and the aggregated throughput exceeds the actual bandwidth, hence the collision probability of the multi-hop packets becomes high as the number of flows increases. Accordingly, a few multi-hop packets will be received at the destination while most of the received packets are single hop packets.

inter-flow interference leads to increase in the packet end-to-end delay due to an increase in the MAC contention time and packet retransmission at the MAC layer, see Fig. 5(d). In all of the cases considered, AODV-MRCR performs significantly better than the AODV-MR routing

High Throughput Path Establishment for Common Traffi c in Wireless Mesh Networks 241

We study the performance of the AODV-MRCR and AODV-MR when there are 100 nodes distribute on area of 100 1000 <sup>×</sup> 1000 meter<sup>2</sup> and twelve of non-overlapping channels. Table 1

We analyze the performance of the proposed protocol in two main scenarios. In the first scenario, we study the effective of TCP traffic load by varied the number of the flow from 10

Fig. 6 shows the performance of the proposed protocol and the AODV-MR. The results of Fig. 6 show that the AODV-MR achieves poor performance as the number of flows or the packet size increases. This result occurs because AODV-MR suffers from many problems. First, it used hop count metric as path selection metric, hence, AODV-MR not being able to avoid the hot spot area, it may also select paths with small channel diversity and route the packet through high-congestion areas. The result is that the links frequently get saturated and suffer

Second, the AODV-MR using the same route and same channel to forward the TCP data and the acknowledgement, which leads to an increased Round Trip Time (RTT). In contrast, The AODV-MRCR minimizes the packet round trip time by assignees different channels for reverse and forward routes per node, which allows the node to become a full-duplex node. This procedure reduces the contention and transmission time at the MAC layer. Moreover, the AODV-MRCR assigns a unique list of channels for each flow received at the destination, which leads to minimizing the interference and reducing the packet drops due to packet collision,

106 ×

106 ×

0 0.5 1 1.5 2 2.5 3

Aggregate Goodput (bps)

AODVͲMR AODVͲMRCR

AODVͲMR AODVͲMRCR

Aggregate Goodput (bps**)**

**Figure 6.** Simulation 1: results for scenario 3 test TCP traffic (Varying the number of flows and Varying

128 256 512 1024 1440

20 30 40 50

Number of Flows (b)

Packet Size (Byte) (d)

to 50, and in the second scenario, we varied the packet size from 128 to 1440 bytes.

protocol.

*5.1.3. Scenario 3: Varying TCP traffic*

from multi-flow interference.

0 0.5 1 1.5 2 2.5 3

0 0.5 1 1.5 2 2.5 3

Routing Packet Overhead

packet size).

Routing Packet Overhead

20 30 40 50

AODVͲMR AODVͲMRCR

Number of Flows (a)

128 256 512 1024 1440

AODVͲMR AODVͲMRCR

Packet Size (Byte) (c)

shows the simulation parameters for this scenario.

#### *5.1.2. Scenario 2: Effect of number of radio per node*

5 depicts the performance of AODV-MR and proposed algorithm versus the number of radio. The number of radios in each mesh router is varied from three to eight in increments of one along with all other simulation parameters, as per Table 1. This scenario carried out to determine the optimal number of radios to be placed in each mesh router. The results shown in Fig. 5 show that AODV-MR has an improvement in performance up to eight radios in each mesh router. The reason for this is that AODV-MR cannot utilize all the non-overlapping channels, unless they are assigned to the interface. Hence, adding a new interface to the mesh router means adding a new channel to the AODV-MR spectrum utilization.

**Figure 5.** Simulation 1: results for scenario 2 (Varying the number of radios in each mesh router).

Fig. 5 also shows that a limited number of radios per node, about 7, are sufficient for AODV-MRCR to achieve its maximum performance improvements. However, increasing the number of radios beyond seven seems to only achieve a marginal improvement. This result obviously depends on the number of available channels, which are used to choose the reserved channel list for each RREQ\_I received at the destination. The AODV-MRCR statically assigns a unique channel for each mesh router interface and half of these channels consider as used channel. Consequently, increasing the number of interfaces per mesh router, minimizes the non-distributed channels in the unused channel list, which increase the packet loss ratio, see Fig. 5(a). However, increase the number of interfaces per node lead to minimize the number of channel in the unused channel list which leads to increase the inter-flow and intra-flow interferences for the gateway traffic. Moreover, increasing the intra-flow and inter-flow interference leads to increase in the packet end-to-end delay due to an increase in the MAC contention time and packet retransmission at the MAC layer, see Fig. 5(d). In all of the cases considered, AODV-MRCR performs significantly better than the AODV-MR routing protocol.

#### *5.1.3. Scenario 3: Varying TCP traffic*

14 Will-be-set-by-IN-TECH

such as 50 flows, the AODV-MR's end-to-end-delay (Fig. 4(d)) reduces compare to other flows such as 30 and 40 flows. The reason for that is at high traffic load the network is saturated and the aggregated throughput exceeds the actual bandwidth, hence the collision probability of the multi-hop packets becomes high as the number of flows increases. Accordingly, a few multi-hop packets will be received at the destination while most of the received packets are

5 depicts the performance of AODV-MR and proposed algorithm versus the number of radio. The number of radios in each mesh router is varied from three to eight in increments of one along with all other simulation parameters, as per Table 1. This scenario carried out to determine the optimal number of radios to be placed in each mesh router. The results shown in Fig. 5 show that AODV-MR has an improvement in performance up to eight radios in each mesh router. The reason for this is that AODV-MR cannot utilize all the non-overlapping channels, unless they are assigned to the interface. Hence, adding a new interface to the mesh

> 0 0,5 1 1,5 2 2,5 3 3,5

0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6

EndͲtoͲEnd Delay (seconds) **106 ×**

Aggregate Goodput (bps)

345678

AODV-MR-10 AODV-MRCR-10 AODV-MR-30 AODV-MRCR-30 AODV-MR-50 AODV-MRCR-50

> Number of Radios (b)

AODV-MR-10 AODV-MRCR-10 AODV-MR-30 AODV-MRCR-30 AODV-MR-50 AODV-MRCR-50

345678

Number of Radios (d)

router means adding a new channel to the AODV-MR spectrum utilization.

345678

345678

AODV-MR-10 AODV-MRCR-10 AODV-MR-30 AODV-MRCR-30 AODV-MR-50 AODV-MRCR-50

**Figure 5.** Simulation 1: results for scenario 2 (Varying the number of radios in each mesh router).

Fig. 5 also shows that a limited number of radios per node, about 7, are sufficient for AODV-MRCR to achieve its maximum performance improvements. However, increasing the number of radios beyond seven seems to only achieve a marginal improvement. This result obviously depends on the number of available channels, which are used to choose the reserved channel list for each RREQ\_I received at the destination. The AODV-MRCR statically assigns a unique channel for each mesh router interface and half of these channels consider as used channel. Consequently, increasing the number of interfaces per mesh router, minimizes the non-distributed channels in the unused channel list, which increase the packet loss ratio, see Fig. 5(a). However, increase the number of interfaces per node lead to minimize the number of channel in the unused channel list which leads to increase the inter-flow and intra-flow interferences for the gateway traffic. Moreover, increasing the intra-flow and

Number of Radios (c)

Number of Radios (a)

AODV-MR-10 AODV-MRCR-10 AODV-MR-30 AODV-MRCR-30 AODV-MR-50 AODV-MRCR-50

single hop packets.

*5.1.2. Scenario 2: Effect of number of radio per node*

0 0,5 1 1,5 2 2,5

Packet Delivery Ratio (%) **105 ×**

Packet Loss (packet) We study the performance of the AODV-MRCR and AODV-MR when there are 100 nodes distribute on area of 100 1000 <sup>×</sup> 1000 meter<sup>2</sup> and twelve of non-overlapping channels. Table 1 shows the simulation parameters for this scenario.

We analyze the performance of the proposed protocol in two main scenarios. In the first scenario, we study the effective of TCP traffic load by varied the number of the flow from 10 to 50, and in the second scenario, we varied the packet size from 128 to 1440 bytes.

Fig. 6 shows the performance of the proposed protocol and the AODV-MR. The results of Fig. 6 show that the AODV-MR achieves poor performance as the number of flows or the packet size increases. This result occurs because AODV-MR suffers from many problems. First, it used hop count metric as path selection metric, hence, AODV-MR not being able to avoid the hot spot area, it may also select paths with small channel diversity and route the packet through high-congestion areas. The result is that the links frequently get saturated and suffer from multi-flow interference.

Second, the AODV-MR using the same route and same channel to forward the TCP data and the acknowledgement, which leads to an increased Round Trip Time (RTT). In contrast, The AODV-MRCR minimizes the packet round trip time by assignees different channels for reverse and forward routes per node, which allows the node to become a full-duplex node. This procedure reduces the contention and transmission time at the MAC layer. Moreover, the AODV-MRCR assigns a unique list of channels for each flow received at the destination, which leads to minimizing the interference and reducing the packet drops due to packet collision,

**Figure 6.** Simulation 1: results for scenario 3 test TCP traffic (Varying the number of flows and Varying packet size).

#### 16 Will-be-set-by-IN-TECH 242 Wireless Mesh Networks – Effi cient Link Scheduling, Channel Assignment and Network Planning Strategies High Throughput Path Establishment for Common Traffic in Wireless Mesh Networks <sup>17</sup>

enabling the route to be effective for long durations while minimizing the number of route discovery messages.

Fig. 6(c, d) shows the performance of both protocols when varying the packet sizes. Our protocol outperforms AODV-MR in terms of aggregate throughput and routing overhead as the packet size increases. This is because our protocol reduces the round trip time at the intermediate node and reduces the transmission time. Reducing the RTT time leads to an increase in the network throughput due to the TCP packet generation depends on successfully receiving ACKs. Furthermore, similar to [14], Fig. 6 shows that a larger packet size can lead to an increase in link failure at the MAC layer. Similarly, a small packet size can reduce the duration of capture, resulting in frequent opportunities for channel access. However, a small packet size also increases the control overhead and can increase the number of collisions at the link layer.

0 0.5 1 1.5 2 2.5

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

EndͲtoͲEnd Delay (seconds)

packets.

traffic.

105 ×

Packet Loss (packet)

10 20 30 40 50

10 20 30 40 50

**Figure 7.** Simulation 2: results for scenario 1 (Varying the number of flows).

Number of Flows (c)

involved in the path during the routing discovery process.

*5.2.2. Scenario 2: Varying the number of available channels*

Routing Packet Overhead

However, our protocol sends a copy of broadcasting message on all node's interfaces; in a multi-channel network the number of channels are more than the number of interfaces. Moreover, our protocol does not use the hello message to inform the neighbors about the node channel, as in the MCR. This is because our scheme distributes the channel to the nodes

To investigate the impact of the number of channel in AODV-MRCR, we varied the number of available channels from 5 to 12. The other simulation parameters are fixed as in Table 1. Fig. 8 shows the results of the MCR with respect to the AODV-MRCR for a different number of available channels. As observed from Fig. 8(b), AODV-MRCR shows higher PDR than MCR regardless of the number of channels. The performance difference becomes large with the increase in the number of channel. The reason of this performance improvement of AODV-MRCR is can be explained by the fact that adding more channel will increase the number of concurrent transmission. Moreover, using more channels allows our protocol to maximize the channel diversity along the path as well as minimize the channel use between multiple concurrent transmission flows. However, as the number of channels increases, the node interface becomes congested which can limit the MCR performance. This is because that the MCR routing protocol spent more time by the switchable interface in sending broadcast

It is interesting to note that at 5 available channels, our protocol has a higher routing overhead than the MCR routing protocol. This is because our protocol selects the reserved channel list from the unused channel list. However, small available channels while keeping the number of interfaces per mesh router at three will minimize the number of channels in the unused channel list. For example, at 5 available channels the unused channel list will be four channels, as at least one channel will be used to keep the network connectivity and to support the local

Packet Delivery Ratio (%)

10 20 30 40 50

MCR AODVͲMRCR

Number of Flows (b)

10 20 30 40 50

Number of Flows (d)

MCR AODVͲMRCR

High Throughput Path Establishment for Common Traffi c in Wireless Mesh Networks 243

Number of Flows (a)

TXͲRate MCR AODVͲMRCR

MCR AODVͲMRCR

#### **5.2. Simulation 2: Compare our protocol with multi-radio multi-channel routing protocol**

In this simulation, we compare the performance of our protocol with the performance of the MCR routing protocol [17]. Many centralized approaches have been proposed to assign a channel to node interfaces in multi-radio multi-channel networks. Most of these approaches need a global view of the network such as traffic profile or node position. Even though, our approach is a centralized approach that establishes high throughput paths for the gateway traffic, it does not require prior knowledge about the network. The MCR routing protocol is multi-radio multi-channel routing protocol that is similar to our protocol in some aspects, such as, it randomly selects a channel with no prior knowledge about the network and uses the routing management messages to inform the neighbors about the selected channel. For all scenarios, varying number of flows, varying packet size, studying the impact of node density, studying the impact of local traffic, and varying the number of non-overlapping channels available in this simulation, we keep the number of interfaces per mesh router to three interfaces and each interface is statically dedicated to a channel using the common channel assignment approach. We carried out different scenarios as described below:

#### *5.2.1. Scenario 1: Varying number of the flows*

In this scenario, we evaluate the impact of varying the number of the flows in the network. The number of CBR flows is varied from 10 to 50 with an increment of 10 flows. We keep the other parameters as in Table 1. Under heavy traffic load beyond twenty flows, the proposed approach performs better than MCR as shown in Fig. 7. Moreover, with high traffic load, the MCR end-to-end delay increases as the number of concurrent flows increase see Fig. 7(c). This is because the MCR routing protocol adds extra delay overheads for every received packet. In contrast, our protocol reduces the packet end-to-end delay. This is due to the fact that in proposed scheme, a node becomes full duplex transmission. Furthermore, the scheme only assigns channel to the active nodes, which considerably reduces the interference and contention time at the MAC layer and hence the delay reduces. The delay increment has direct impact in the aggregated goodput and number of packet loss see figure. Hence, increase the number of packet loss lead to increase in the of control message. Since the MCR routing protocol sends a copy of the broadcasting message on every channel, the MCR overhead (RREQ, RREP, HELLO) increases as the number of the flows increases as show in Fig. 7(d).

**Figure 7.** Simulation 2: results for scenario 1 (Varying the number of flows).

However, our protocol sends a copy of broadcasting message on all node's interfaces; in a multi-channel network the number of channels are more than the number of interfaces. Moreover, our protocol does not use the hello message to inform the neighbors about the node channel, as in the MCR. This is because our scheme distributes the channel to the nodes involved in the path during the routing discovery process.

#### *5.2.2. Scenario 2: Varying the number of available channels*

16 Will-be-set-by-IN-TECH

enabling the route to be effective for long durations while minimizing the number of route

Fig. 6(c, d) shows the performance of both protocols when varying the packet sizes. Our protocol outperforms AODV-MR in terms of aggregate throughput and routing overhead as the packet size increases. This is because our protocol reduces the round trip time at the intermediate node and reduces the transmission time. Reducing the RTT time leads to an increase in the network throughput due to the TCP packet generation depends on successfully receiving ACKs. Furthermore, similar to [14], Fig. 6 shows that a larger packet size can lead to an increase in link failure at the MAC layer. Similarly, a small packet size can reduce the duration of capture, resulting in frequent opportunities for channel access. However, a small packet size also increases the control overhead and can increase the number of collisions at

**5.2. Simulation 2: Compare our protocol with multi-radio multi-channel routing**

In this simulation, we compare the performance of our protocol with the performance of the MCR routing protocol [17]. Many centralized approaches have been proposed to assign a channel to node interfaces in multi-radio multi-channel networks. Most of these approaches need a global view of the network such as traffic profile or node position. Even though, our approach is a centralized approach that establishes high throughput paths for the gateway traffic, it does not require prior knowledge about the network. The MCR routing protocol is multi-radio multi-channel routing protocol that is similar to our protocol in some aspects, such as, it randomly selects a channel with no prior knowledge about the network and uses the routing management messages to inform the neighbors about the selected channel. For all scenarios, varying number of flows, varying packet size, studying the impact of node density, studying the impact of local traffic, and varying the number of non-overlapping channels available in this simulation, we keep the number of interfaces per mesh router to three interfaces and each interface is statically dedicated to a channel using the common

channel assignment approach. We carried out different scenarios as described below:

In this scenario, we evaluate the impact of varying the number of the flows in the network. The number of CBR flows is varied from 10 to 50 with an increment of 10 flows. We keep the other parameters as in Table 1. Under heavy traffic load beyond twenty flows, the proposed approach performs better than MCR as shown in Fig. 7. Moreover, with high traffic load, the MCR end-to-end delay increases as the number of concurrent flows increase see Fig. 7(c). This is because the MCR routing protocol adds extra delay overheads for every received packet. In contrast, our protocol reduces the packet end-to-end delay. This is due to the fact that in proposed scheme, a node becomes full duplex transmission. Furthermore, the scheme only assigns channel to the active nodes, which considerably reduces the interference and contention time at the MAC layer and hence the delay reduces. The delay increment has direct impact in the aggregated goodput and number of packet loss see figure. Hence, increase the number of packet loss lead to increase in the of control message. Since the MCR routing protocol sends a copy of the broadcasting message on every channel, the MCR overhead (RREQ, RREP, HELLO) increases as the number of the flows increases as show in Fig. 7(d).

*5.2.1. Scenario 1: Varying number of the flows*

discovery messages.

the link layer.

**protocol**

To investigate the impact of the number of channel in AODV-MRCR, we varied the number of available channels from 5 to 12. The other simulation parameters are fixed as in Table 1.

Fig. 8 shows the results of the MCR with respect to the AODV-MRCR for a different number of available channels. As observed from Fig. 8(b), AODV-MRCR shows higher PDR than MCR regardless of the number of channels. The performance difference becomes large with the increase in the number of channel. The reason of this performance improvement of AODV-MRCR is can be explained by the fact that adding more channel will increase the number of concurrent transmission. Moreover, using more channels allows our protocol to maximize the channel diversity along the path as well as minimize the channel use between multiple concurrent transmission flows. However, as the number of channels increases, the node interface becomes congested which can limit the MCR performance. This is because that the MCR routing protocol spent more time by the switchable interface in sending broadcast packets.

It is interesting to note that at 5 available channels, our protocol has a higher routing overhead than the MCR routing protocol. This is because our protocol selects the reserved channel list from the unused channel list. However, small available channels while keeping the number of interfaces per mesh router at three will minimize the number of channels in the unused channel list. For example, at 5 available channels the unused channel list will be four channels, as at least one channel will be used to keep the network connectivity and to support the local traffic.

that as the local traffic increase, the AODV-MRCR improves both type of the traffic and still can get higher results than MCR. The figure also shows that the AODV-MRCR's performance metric for gateway traffic does not decrease when the number of the local traffic increase. The improved performance of AODV-MRCR can be explained as following. Our protocol differentiates between the local traffic and internet traffic by reserve a list of channels for the gateway traffic and these channels cannot be used to transmit the peer-to-peer traffic. Moreover, the AODV-MRCR protocol assigns different channel for the reverse and forward path which lead to better channel diversity than MCR routing protocol. Fig. 9 also shows that for the local traffic( peer-to-peer) of ten flows, the performance of both protocols in term of PDR and end-to-end delay are decreased. The reason for that is the randomly distributed traffic between mesh router nodes as well as the node random distribution may cause the

High Throughput Path Establishment for Common Traffi c in Wireless Mesh Networks 245

In this chapter, we proposed a channel reservation scheme, which establishes a high throughput path for the gateway traffic by utilizing the WMN characteristics, such as multi-radio mesh router and most of the traffic toward the gateway. The channel reservation and assignment are integrated with the gateway routing discovery process. This scheme reduced the influence of local traffic on the performance of the gateway traffic. Moreover, the scheme minimized the number of nodes using the channel by only assigning channels to the node that is involved in the gateway path route discovery process. The performance of the proposed scheme is evaluated with respect to the metrics, such as packet delivery ratio, end-to-end delay, aggregate throughput, packet loss and routing overhead. The results obtained show that the proposed scheme is better than the existing schemes with respect to these metrics. Currently, the protocol designed in this chapter is mainly for infrastructure wireless mesh networks, which assumes that most of the traffic is towards the gateway. This proposed protocol could be further enhanced to support a more general wireless network, such as hybrid wireless mesh network. Moreover, it can be enhanced by assigning the channel to traffic based on the traffic load. Another possible extension is to consider multiple transmission rates. Different transmission rates can be achieved by using different modulation schemes, for example, IEEE 802.11b transmissions support four different data rates 1Mbps, 2Mbps, 5.5Mbps, and 11 Mbps. Finally, the proposed protocol can be further investigated

The research was partially supported by the Research University, FRGS

Hassen A. Mogaibel, Mohamed Othman, Shamala Subramaniam and Nor Asilah Wati Abdul

*Department of Communication Technology and Network, Universiti Putra Malaysia, 43400 UPM,*

traffic to be located in the same area.

with different MAC protocols for different radio interfaces.

**6. Conclusion**

**Acknowledgement**

/1/11/SG/UPM/01/1

*Serdang, Selangor D.E., Malaysia*

**Author details**

Hamid

**Figure 8.** Simulation 2: results for scenario 2 (Varying the number of channels).

#### *5.2.3. Scenario 3: Impact of local traffic*

To investigate the impact of the local traffic on the gateway traffic, we varied the number of the peer-to-peer traffic from 5 to 20 flows. The number of the gateway traffic is the subtraction of the total number flows (50 flows) in the network and number of the local traffic per each scenario. The simulation parameters for this scenario are as shown in Table 1. Fig. 9 shows the end-to-end delay and PDR for both peer-to-peer and gateway traffic. This figure Shows

**Figure 9.** Simulation 2: results for scenario 3 (Impact of local traffic).

that as the local traffic increase, the AODV-MRCR improves both type of the traffic and still can get higher results than MCR. The figure also shows that the AODV-MRCR's performance metric for gateway traffic does not decrease when the number of the local traffic increase. The improved performance of AODV-MRCR can be explained as following. Our protocol differentiates between the local traffic and internet traffic by reserve a list of channels for the gateway traffic and these channels cannot be used to transmit the peer-to-peer traffic. Moreover, the AODV-MRCR protocol assigns different channel for the reverse and forward path which lead to better channel diversity than MCR routing protocol. Fig. 9 also shows that for the local traffic( peer-to-peer) of ten flows, the performance of both protocols in term of PDR and end-to-end delay are decreased. The reason for that is the randomly distributed traffic between mesh router nodes as well as the node random distribution may cause the traffic to be located in the same area.

#### **6. Conclusion**

18 Will-be-set-by-IN-TECH

> 0 0.5 1 1.5 2 2.5 3 3.5

Routing Packet Overhead

To investigate the impact of the local traffic on the gateway traffic, we varied the number of the peer-to-peer traffic from 5 to 20 flows. The number of the gateway traffic is the subtraction of the total number flows (50 flows) in the network and number of the local traffic per each scenario. The simulation parameters for this scenario are as shown in Table 1. Fig. 9 shows the end-to-end delay and PDR for both peer-to-peer and gateway traffic. This figure Shows

**(b) Peer-to-Peertraffic**

Packet Delivery Ratio (%)

**(a) Gatewaytraffic**

Packet Delivery Ratio (%)

Packet Delivery Ratio (%)

5 7 912

5 7 9 12

MCR AODVͲMRCR

Number ofChannels (d)

5 10 15 20

5 10 15 20

MCR AODVͲMRCR

Number of PeerͲtoͲPeer Flows

Number of PeerͲtoͲPeer Flows

MCR AODVͲMRCR

Number of Channels (b)

MCR AODVͲMRCR

0 0.5 1 1.5 2 2.5

0 0.1 0.2 0.3 0.4

EndͲtoͲEnd Delay (seconds)

*5.2.3. Scenario 3: Impact of local traffic*

0 0.05 0.1 0.15 0.2

0 0.05 0.1 0.15 0.2

EndͲtoͲEnd Delay (seconds)

EndͲtoͲEnd Delay (seconds)

5 10 15 20

**Figure 9.** Simulation 2: results for scenario 3 (Impact of local traffic).

5 10 15 20

Number of PeerͲtoͲPeer Flows

MCR AODVͲMRCR

MCR AODVͲMRCR

Number of PeerͲtoͲPeer Flows

106 ×

Packet Loss (packet)

5 7 912

5 7 912

**Figure 8.** Simulation 2: results for scenario 2 (Varying the number of channels).

Number of Channels (c)

Number of Channels (a)

MCR AODVͲMRCR

TXͲRate MCR AODVͲMRCR

In this chapter, we proposed a channel reservation scheme, which establishes a high throughput path for the gateway traffic by utilizing the WMN characteristics, such as multi-radio mesh router and most of the traffic toward the gateway. The channel reservation and assignment are integrated with the gateway routing discovery process. This scheme reduced the influence of local traffic on the performance of the gateway traffic. Moreover, the scheme minimized the number of nodes using the channel by only assigning channels to the node that is involved in the gateway path route discovery process. The performance of the proposed scheme is evaluated with respect to the metrics, such as packet delivery ratio, end-to-end delay, aggregate throughput, packet loss and routing overhead. The results obtained show that the proposed scheme is better than the existing schemes with respect to these metrics. Currently, the protocol designed in this chapter is mainly for infrastructure wireless mesh networks, which assumes that most of the traffic is towards the gateway. This proposed protocol could be further enhanced to support a more general wireless network, such as hybrid wireless mesh network. Moreover, it can be enhanced by assigning the channel to traffic based on the traffic load. Another possible extension is to consider multiple transmission rates. Different transmission rates can be achieved by using different modulation schemes, for example, IEEE 802.11b transmissions support four different data rates 1Mbps, 2Mbps, 5.5Mbps, and 11 Mbps. Finally, the proposed protocol can be further investigated with different MAC protocols for different radio interfaces.

#### **Acknowledgement**

The research was partially supported by the Research University, FRGS /1/11/SG/UPM/01/1

#### **Author details**

Hassen A. Mogaibel, Mohamed Othman, Shamala Subramaniam and Nor Asilah Wati Abdul Hamid

*Department of Communication Technology and Network, Universiti Putra Malaysia, 43400 UPM, Serdang, Selangor D.E., Malaysia*

20 Will-be-set-by-IN-TECH 246 Wireless Mesh Networks – Effi cient Link Scheduling, Channel Assignment and Network Planning Strategies High Throughput Path Establishment for Common Traffic in Wireless Mesh Networks <sup>21</sup>

#### **7. References**

[1] Adya, A., Bahl, P., Padhye, J., Wolman, A. & Zhou, L. [2004]. A multi-radio unification protocol for ieee 802.11 wireless networks, *Broadband Networks, 2004. BroadNets 2004. Proceedings. First International Conference on*, pp. 344–354.

[16] Kim, S.-H. & Suh, Y. J. [2008]. Local channel information assisted channel assignment for multi-channel wireless mesh networks, *Vehicular Technology Conference (VTC) Spring*

High Throughput Path Establishment for Common Traffi c in Wireless Mesh Networks 247

[17] Kyasanur, P. & Vaidya, N. H. [2006]. Routing and link-layer protocols for multi-channel multi-interface ad hoc wireless networks, *SIGMOBILE Mob. Comput. Commun. Rev.*

[18] Lee, J., Kim, D., Garcia-Luna-Aceves, J., Choi, Y., Choi, J. & Nam, S. [2003]. Hybrid gateway advertisement scheme for connecting mobile ad hoc networks to the internet, *Vehicular Technology Conference, 2003. VTC 2003-Spring. The 57th IEEE Semiannual*, Vol. 1,

[19] Marina, M. K., Das, S. R. & Subramanian, A. P. [2010]. A topology control approach for utilizing multiple channels in multi-radio wireless mesh networks, *Comput. Netw.*

[20] Mogaibel, H. A., Othman, M., Subramaniam, S. & Hamid, N. A. W. A. [2012]. On-demand channel reservation scheme for common traffic in wireless mesh networks,

[21] Nandiraju, D. S., Nandiraju, N. S. & Agrawal, D. P. [2009]. Adaptive state-based multi-radio multi-channel multi-path routing in wireless mesh networks, *Pervasive Mob.*

[23] Pediaditaki, S., Arrieta, P. & Marina, M. [2009]. A learning-based approach for distributed multi-radio channel allocation in wireless mesh networks, *Network Protocols,*

[24] Pirzada, A. A., Portmann, M. & Indulska, J. [2006]. Evaluation of multi-radio extensions to aodv for wireless mesh networks, *MobiWac '06: Proceedings of the 4th ACM international workshop on Mobility management and wireless access*, ACM, New York, NY, USA, pp. 45–51. [25] Raniwala, A. & Chiueh, T. [2005]. Architecture and algorithms for an ieee 802.11-based multi-channel wireless mesh network, *INFOCOM 2005. 24th Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings IEEE*, Vol. 3, pp. 2223–2234. [26] Raniwala, A., Gopalan, K. & Chiueh, T. [2004]. Centralized channel assignment and routing algorithms for multi-channel wireless mesh networks, *ACM Mobile Computing*

[27] Ratanchandani, P. & Kravets, R. [2003]. A hybrid approach to internet connectivity for mobile ad hoc networks, *Wireless Communications and Networking, 2003. WCNC 2003. 2003*

[28] Sarkar, T., Ji, Z., Kim, K., Medouri, A. & Salazar-Palma, M. [2003]. A survey of various propagation models for mobile communication, *Antennas and Propagation Magazine, IEEE*

[29] Shui, G. & Shen, S. [2008]. A new multi-channel mac protocol combined with on-demand routing for wireless mesh networks, *Computer Science and Software Engineering, 2008*

[30] So, J. & Vaidya, N. [2004a]. Multi-channel mac for ad hoc networks: Handling multi-channel hidden terminals using a single transceiver, *In ACM MobiHoc*, pp. 222–233. [31] So, J. & Vaidya, N. [2004b]. A routing protocol for utilizing multiple channels in

[32] Subramanian, A., Buddhikot, M. & Miller, S. [2006]. Interference aware routing in multi-radio wireless mesh networks, *Wireless Mesh Networks, 2006. WiMesh 2006. 2nd*

multi-hop wireless networks with a single transceiver, *Technical report*.

URL: *http://www.sciencedirect.com/science/article/pii/S1084804512000318*

*2008, IEEE, 2008*, pp. 2611–2615.

*Journal of Network and Computer Applications* (0): –.

[22] *NS. The Network Simulator, http://www.isi.edu/nsnam/ns/* [1989].

*2009. ICNP 2009. 17th IEEE International Conference on*, pp. 31–41.

10: 31–43.

pp. 191–195.

54: 241–256.

*Comput.* 5: 93–109.

*and Communications Review* 8: 50–65.

*International Conference, vol. 4, 2008*, pp. 1036–1039.

*IEEE*, Vol. 3, pp. 1522–1527.

*IEEE Workshop on*, pp. 55–63.

45(3): 51–82.


20 Will-be-set-by-IN-TECH

[1] Adya, A., Bahl, P., Padhye, J., Wolman, A. & Zhou, L. [2004]. A multi-radio unification protocol for ieee 802.11 wireless networks, *Broadband Networks, 2004. BroadNets 2004.*

[2] Akyildiz, I. F., Wang, X. & Wang, W. [2005]. Wireless mesh networks: a survey, *Comput.*

[3] Avallone, S., Akyildiz, I. & Ventre, G. [2009]. A channel and rate assignment algorithm and a layer-2.5 forwarding paradigm for multi-radio wireless mesh networks,

[4] Bahl, P., Chandra, R. & Dunagan, J. [2004]. Ssch: slotted seeded channel hopping for capacity improvement in ieee 802.11 ad-hoc wireless networks, *Proceedings of the 10th annual international conference on Mobile computing and networking*, MobiCom '04, ACM,

[5] Broch, J., Maltz, D. A. & Johnson, D. B. [1999]. Supporting hierarchy and heterogeneous interfaces in multi-hop wireless ad hoc networks, *Proceedings of the 1999 International Symposium on Parallel Architectures, Algorithms and Networks*, ISPAN '99, IEEE Computer

[6] Chaudhry, A., Hafez, R., Aboul-Magd, O. & Mahmoud, S. [2010]. Fault-tolerant and scalable channel assignment for multi-radio multi-channel ieee 802.11a-based wireless mesh networks, *Technical report*, GLOBECOM Workshops (GC Wkshps), 2010 IEEE. [7] Chen, J., Jia, J., Wen, Y., zhao, D. & Liu, J. [2009]. A genetic approach to channel assignment for multi-radio multi-channel wireless mesh networks, *GEC '09: Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation*, ACM, New

[8] Chiu, H. S., Yeung, K. & Lui, K.-S. [2009]. J-car: An efficient joint channel assignment and routing protocol for ieee 802.11-based multi-channel multi-interface mobile ad hoc

[9] Ding, Y., Pongaliur, K. & Xiao, L. [2009]. Hybrid multi-channel multi-radio wireless mesh networks, *Quality of Service, 2009. IWQoS. 17th International Workshop on*, pp. 1–5. [10] Draves, R., Padhye, J. & Zill, B. [2004]. Routing in multi-radio, multi-hop wireless mesh networks, *MobiCom '04: Proceedings of the 10th annual international conference on Mobile*

[11] Elizabeth, Y. S., Sun, Y., Belding-Royer, E. M. & Perkins, C. E. [2002]. Internet Connectivity for Ad hoc Mobile Networks, *International journal of wireless information*

[12] Gong, M. X., Midkiff, S. F. & Mao, S. [2009]. On-demand routing and channel assignment

[13] Hamidian, A. A. [2003]. *A study of internet connectivity for mobile ad hoc networks in ns2*, Master's thesis, Department of Communication Systems, Lund Institute of Technology,

[14] Jiang, R., Gupta, V. & Ravishankar, C. V. [2003]. Interactions between tcp and the ieee 802.11 mac protocol, *DARPA Information Survivability Conference and Exposition,*

[15] Jönsson, U., Alriksson, F., Larsson, T., Johansson, P. & Maguire, Jr., G. Q. [2000]. Mipmanet: mobile ip for mobile ad hoc networks, *Proceedings of the 1st ACM international symposium on Mobile ad hoc networking & computing*, MobiHoc '00, IEEE Press, Piscataway,

networks, *Wireless Communications, IEEE Transactions on* 8(4): 1706–1715.

*computing and networking*, ACM, New York, NY, USA, pp. 114–128.

in multi-channel mobile ad hoc networks, *Ad Hoc Netw.* 7(1): 63–78.

*Proceedings. First International Conference on*, pp. 344–354.

*Networking, IEEE/ACM Transactions on* 17(1): 267–280.

Society, Washington, DC, USA, pp. .370 – 375.

**7. References**

*Netw. ISDN Syst.* 47: 445–487.

New York, NY, USA, pp. 216–230.

York, NY, USA, pp. 39–46.

*networks* 9(2): 75–88.

Lund University,.

NJ, USA, pp. 75–85.

1: 273–282.

	- [33] Sun, W., Cong, R., Xia, F., Chen, X. & Qin, Z. [2010]. R-ca: A routing-based dynamic channel assignment algorithm in wireless mesh networks, *Ubiquitous, Autonomic and Trusted Computing, Symposia and Workshops on* 0: 228–232.
	- [34] Tang, J., Xue, G. & Zhang, W. [2005]. Interference-aware topology control and qos routing in multi-channel wireless mesh networks, *MobiHoc '05: Proceedings of the 6th ACM international symposium on Mobile ad hoc networking and computing*, ACM, New York, NY, USA, pp. 68–77.
	- [35] Tarn, W.-H. & Tseng, Y.-C. [2007]. Joint multi-channel link layer and multi-path routing design for wireless mesh networks, *INFOCOM 2007. 26th IEEE International Conference on Computer Communications. IEEE*, pp. 2081–2089.
	- [36] Wakikawa, Perkins, C., Nilsson, A. & Tuominen, A. J. [2001]. Global Connectivity for IPv6 Mobile Ad Hoc Networks, IETF Internet Draft, Work in process.
	- [37] Wan, P.-J. [2009]. Multiflows in multihop wireless networks, *Proceedings of the tenth ACM international symposium on Mobile ad hoc networking and computing*, MobiHoc '09, ACM, New York, NY, USA, pp. 85–94.
		- URL: *http://doi.acm.org/10.1145/1530748.1530761*
	- [38] Yue, X., Wong, C.-F. & Chan, S.-H. G. [2011]. Cacao: Distributed client-assisted channel assignment optimization for uncoordinated wlans, *Parallel and Distributed Systems, IEEE Transactions on* 22(9): 1433–1440.
	- [39] Zhuang, L., Liu, Y. & Liu, K. [2009]. Hybrid gateway discovery mechanism in mobile ad hoc for internet connectivity, *Proceedings of the 2009 International Conference on Wireless Communications and Mobile Computing: Connecting the World Wirelessly*, IWCMC '09, ACM, New York, NY, USA, pp. 143–147.

22 Will-be-set-by-IN-TECH

248 Wireless Mesh Networks – Effi cient Link Scheduling, Channel Assignment and Network Planning Strategies

[33] Sun, W., Cong, R., Xia, F., Chen, X. & Qin, Z. [2010]. R-ca: A routing-based dynamic channel assignment algorithm in wireless mesh networks, *Ubiquitous, Autonomic and*

[34] Tang, J., Xue, G. & Zhang, W. [2005]. Interference-aware topology control and qos routing in multi-channel wireless mesh networks, *MobiHoc '05: Proceedings of the 6th ACM international symposium on Mobile ad hoc networking and computing*, ACM, New York, NY,

[35] Tarn, W.-H. & Tseng, Y.-C. [2007]. Joint multi-channel link layer and multi-path routing design for wireless mesh networks, *INFOCOM 2007. 26th IEEE International Conference on*

[36] Wakikawa, Perkins, C., Nilsson, A. & Tuominen, A. J. [2001]. Global Connectivity for

[37] Wan, P.-J. [2009]. Multiflows in multihop wireless networks, *Proceedings of the tenth ACM international symposium on Mobile ad hoc networking and computing*, MobiHoc '09, ACM,

[38] Yue, X., Wong, C.-F. & Chan, S.-H. G. [2011]. Cacao: Distributed client-assisted channel assignment optimization for uncoordinated wlans, *Parallel and Distributed Systems, IEEE*

[39] Zhuang, L., Liu, Y. & Liu, K. [2009]. Hybrid gateway discovery mechanism in mobile ad hoc for internet connectivity, *Proceedings of the 2009 International Conference on Wireless Communications and Mobile Computing: Connecting the World Wirelessly*, IWCMC '09,

IPv6 Mobile Ad Hoc Networks, IETF Internet Draft, Work in process.

*Trusted Computing, Symposia and Workshops on* 0: 228–232.

*Computer Communications. IEEE*, pp. 2081–2089.

URL: *http://doi.acm.org/10.1145/1530748.1530761*

New York, NY, USA, pp. 85–94.

*Transactions on* 22(9): 1433–1440.

ACM, New York, NY, USA, pp. 143–147.

USA, pp. 68–77.
