This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Abstract- Routing Overhead is an important issue of any Routing Protocol in wireless networks. It incurs during broadcast of Route Request packets (RREQ) during route discovery & broadcast of HELLO packets during link connectivity. A new routing protocol called Enhanced Ad-hoc on Demand Distance Vector (E-AODV) routing protocol has been proposed in this paper which merges the Blocking Expanding Ring Search (B-ERS) & Routing packets as HELLO packets techniques to reduce routing overhead. Results shows that, performance of E-AODV routing protocol is better than existing AODV in wireless networks.
Keywords- Ad hoc network; Routing protocols; AODV; Hello Message; Stop instruction.
Wireless ad-hoc networks are self-organizing, rapidly deployable, and require no fixed infrastructure. They are comprised of wireless nodes, which must cooperate in order to dynamically establish communications using limited network management and administration . AODV routing protocol is a popular on demand routing protocol used in Wireless Mesh networks . AODV routing protocol is a reactive protocol in which the routes are created only when they are needed . Mobile Ad-Hoc Network (MANET) is a recent developed part of wireless communication. In MANET's, there is no need for established infrastructure and the task of packet forwarding from one host to another is done by participants, also called mobile nodes . Conventional AODV suffers from a lot of routing overhead. Energy consumption is a major issue in designing protocols for Ad Hoc Wireless Networks due to battery constraints. Two major factors affecting routing overhead are broadcast of Route Request & HELLO packets . In the literature, two techniques have been identified which reduce the routing overhead during the wireless communication. The first technique is called Blocking Expanding Ring Search (B-ERS) which reduces the search area during route discovery process with the help of Stop packets .The second technique make use of routing packet as hello packets for determination of link connectivity . In this work the above two techniques have been merged to enhance the existing Ad hoc On Demand routing protocol in terms of scalability and energy efficiency.
The objectives of this work are to develop routing protocols for wireless networks and to analyze their performance by realizing different environments. The analysis has been done through simulation using NS-2 (Network Simulator). Objectives of the work are:-
1. To Simulate the proposed routing protocol, Enhanced AODV for wireless networks.
2. Evaluation of Enhanced-AODV based on parameters like routing overhead, hello load and end to end delay.
3. Comparison of proposed protocol with existing protocol.
Figure 1. Structure of STOP packet
II. Enhanced Ad-hoc on demand distance vector routing protocol (e-aodv)
The performance of wireless networks is highly dependent on routing protocol. AODV is a popular routing protocol for wireless networks. AODV provides loop freedom for all routes through the use of sequence numbers. Routing overhead is an important issue of any routing protocol. Routing overhead occurs during broadcast of RREQ packets during route discovery and due to broadcast of hello packets. In this paper to develop E-AODV Routing Protocol, two techniques called Blocking Expanding Ring Search and Routing packets as hello packets have been combined and implemented to improve the performance of existing AODV routing protocol in wireless networks.
Blocking Expanding Ring uses a special stop packet which informs all the nodes in the network to stop further rebroadcast of route requests packets after the destination is found which limits unnecessary broadcast of RREQ packets. All the fields used in STOP packets as shown in fig. 1 provides information, regarding which RREQ this STOP packet has been generated for. Each RREQ is generated for a particular destination by a particular source with a unique timestamp. New RREQ either for same destination from same source will be having different timestamps. So, the information in STOP packet is enough to block correct RREQ packet. Stop List is the main data structure used for implementing Blocking Expanding Ring Search. Stop List decides whether RREQ should be forwarded further or blocked. Any node can receive multiple STOP packets for multiple RREQ packets received. Whenever a new request comes, that request will be stored in Stop List. After storing request in stop list, waiting timer will be activated for that particular request. Timer will be initiated with the time 2 x H x unit_time, where H is no of hops from the source node. Whenever a STOP packet for the particular RREQ is received, then the information in the STOP packet is matched with the entries in the stop list. If an entry is found, it signifies that, waiting timer has not yet expired & corresponding RREQ packet is not broadcasted. The found entry will be deleted from the stop list & RREQ will not be broadcasted. If information in the STOP packet does not matches with any entry from the stop list, then RREQ has been already broadcasted then the STOP packet will be broadcasted further to the next ring, up till it is not the destination node.
Other major routing overhead is due to broadcast of Hello packets after a particular interval of time. Second technique, routing packet as hello packets has been implemented to decrease Hello load in the network. This technique improvises the fact that RREQ & Route Error (RERR) are also broadcast packets. They also provide the same information as being provided by the hello packet about the presence of a link. Regular broadcast of RREQ & RERR packets reduces the frequent use of the Hello packet which in turns reduces the overall Hello load. In NS2, Hello packet is being implemented with the help of a hello timer, which is being fired after a particular amount of time which varies between Min_hello_time & Max_hello_time. That Hello timer fire the sendHello () routine & set the new interval for next timer calling.
Two different timer variables Control_Timer & Hello_Timer, is implemented to keep track of the time at which last control packet is broadcasted & last hello packet is broadcasted. Whenever a control packet (RREQ or RERR) is forwarded, Control_Timer is set to that time & when Hello packet is forwarded Hello_Timer is being set to that time. ControlT and HelloT are the variable which will store the value of Control_Timer and Hello_Timer at a particular instant of time. If the value of Control_Timer is more than the Hello_timer, that means a control packet is forwarded after a hello packet so, no need to forward hello packet at this time and a new interval will be calculated. Otherwise, Control_Timer is less than Hello_Timer, no control packet is being forwarded after Hello packet, and next hello packet will be fired at normal time. The following algorithm is implemented for calculating new time interval for rebroadcasting of hello message. If more request and error packets being forwarded, less hello packets will be used to transmit link availability information, which in turns decreases the overall Hello load.
Algorithm: Calculation of new time interval for rebroadcast of Hello packet
2. ControlT = Control_Timer; HelloT=Hello_Timer; CTime=Current_Time;
3. Interval = MinHelloInterval+((MaxHelloInterval - MinHelloInterval)*random: uniform ());
4. If ((ControlT-HelloT) > 0)
6. Timer_Algorithm (interval) //fire algorithm again after Interval.
To improve the performance of existing EAODV above two techniques have been merged and Enhanced AODV has been implemented to reduce routing overhead by using Network Simulator-2.
The aim of the simulation is to determine the effect of changing the number of nodes and mobility of nodes on the Routing Overhead, Hello Load and Average End-End Delay in Ad-hoc mobile network. A Simulation model based on NS-2 is used in our evaluation. The NS-2 simulator is a discrete event simulator widely used in the networking research community. The Monarch group in CMU developed support for simulating wireless networks. Complete with physical layer, data link layer and MAC layer models on NS-2 .
Traffic and Mobility Models
The network used for simulations consists of CBR (Continuous Bit Rate) traffic sources. The packet size is set to 512 byte and the sending rate is set to 4 packets per second. Each simulation is run for 200 simulated seconds. The mobility model uses the random waypoint model in a rectangular 1000Ã-1000 field. Each node starts its journey from a random location to a random destination with a randomly chosen speed (uniformly distributed between 0-30m/sec). Once the destination is reached, another random destination is targeted after a pause. Pause Time is kept fixed at 2.0 seconds. Each data point in the drawn graph represents an average of six runs with identical traffic models, but different randomly generated mobility scenarios. For fairness, identical mobility and traffic scenarios are used for both the protocols that compared with varying number of nodes and determine the effects on the Routing Overhead, Hello Load and Average delay.
B. Performance Metrics
In comparing the protocols, we chose the following metrics
Routing Overhead: The total number of routing packets transmitted & received by all the nodes during the simulation is known as Routing Overhead. The routing overhead metric simply shows how much of the bandwidth that is consumed by routing messages, i.e. the amount of bandwidth available to data packets.
Hello Load: The total number of hello packets transmitted and received by all the nodes during the simulation is known as Hello Load. Hello packets are used to check local link connectivity.
End-to-End Delay: It is average time a packet takes for delivery to its destination after it was transmitted. It tells how a protocol adapts or arranges for an immediate delivery of packets to its desired destination. Average delay is all possible delays like route discovery latency, queuing at the interface queue, retransmission delays at the MAC etc .
IV. RESULTS & DISCUSSIONS
By varying number of nodes and mobility of nodes (meter/seconds), performance metrics have been compared for AODV and Enhanced ADOV.
As shown in Fig. 2, the routing overhead obtained using E-AODV is almost identical to that obtained using AODV when node numbers are small. However, when there are larger numbers of nodes (i.e., more than 80), E-AODV performed better. This suggests that E-AODV is highly effective in discovering routes when network size increases.
Figure 2. Comparison of Routing Overhead for varying No. of Nodes
Figure 3. Comparison of Routing Overhead for varying Mobility (Mobility in meter per seconds)
In fig 3, it is observed that when mobility of nodes increases, Routing Overhead in E-AODV decreases considerably. The reason is that when the mobility increases, more number of links breaks which causes the generation of more STOP packets resulting lower routing overhead.
Figure 4. Comparison of Hello Load for varying No. of Nodes
The Fig.4 shows the comparison of Hello Load for varying no. of nodes, it is observed that hello load slightly improves when there is increase in number of nodes.
Figure 5. Comparison of Hello Load for varying Mobility
The Fig. 5 shows comparison of hello load for varying Mobility and it is observed that hello load is reduced in EAODV as compared to conventional AODV as mobility increases. As the speed of node increases more number of links breaks observed resulting more control packets. These control packets are being used as hello messages which reduce hello load of E-AODV.
Fig. 6 and Fig. 7 show comparison of end to end delay for varying no. of nodes and increasing mobility. In AODV, as the routes are only created when they are required thus resulting a delay which occurs during the route discovery process.
Figure 6. Comparison of End to End Delay for varying No. of Nodes
Figure 7. Comparison of End to End Delay for varying Mobility
However, in EAODV, this delay is lower than the conventional AODV because less time is consumed during route discovery and small number of hello packets avoids congestion in the network.
V. CONCLUSIONS & FUTURE WORK
This paper proposes an enhanced AODV protocol, called EAODV. It is a scheme which improves on routing overhead by utilizing Blocking Expanding Ring Search & Control packets as Hello packets techniques. Routing overhead is much improved in EAODV as compared to conventional AODV when the network size and the mobility of nodes is increased .As the speed of node increases more number of links breaks observed resulting more control packets. These control packets are being used as Hello messages which reduce Hello Load of routing protocol.
However, there are still a lot of important issues for future researches. Route Reply packet is a unicast packet which can also be used as a HELLO packet by utilizing promiscuous mode of a wireless node. In this mode, a node can listen to even unicast packets distributed in network. Future work will enhance the performance of EAODV by applying this mode. HELLO load can be further reduced by using DATA packets as HELLO packets.