Mobility Impact on MANET Routing Protocols In Grid Networks

Mobile Ad-Hoc Networks (MANETS) is a collection of wireless mobile nodes that are able to dynamically form a temporary network without any aid from fixed infrastructure or centralized administration due to no wired backbone.Ad Hoc networks are formed spontaneously and the nodes are highly mobile.Thispaper presents performance evaluations, comparisons, andanalysis for three routing protocols (AODV, DSR, and OLSR) to bring out their relative meritsundervarying network size and mobilitywith various speed and pause times. The simulation is carried out using OMNET++ simulator based on the quantitative basic parameters like throughput, Packet transmission Ratio (PTR),packet transmission time delay and protocol overhead .The nodes are distributed randomly in a grid network topology and mobile nodes moving using Random Waypoint mobility models. The results demonstrate that, undervarious node speeds and pause timesfor different network size, AODVoutperforms DSR and OLSR protocols with respect to networkthroughput (by 0.9% and 4.4%, respectively). For protocoloverhead,DSR is about 13.4% and 65.5% below AODV and OLSR protocols respectively. In contrast, thepacket transmission time delay when using OLSR is shorter than whenusing both AODV (by 81.7%) and DSR (by 76.7%).


III. ROUTING PROTOCOLS IN MANETS
Researcher Community has broadly classifiedthe routing protocols into two categories. These are Proactive (Table  Driven) routingprotocols, because the routing table of node is periodically exchanged and updated.An example of proactive routing protocol is the Optimized Link State Routing Protocol (OLSR) [6,4].A reactive protocol on the other handfinds the route only to needed nodes only ondemand.In reactive or on demand protocols the up-todate routing table is not retained if there is no communication. It will discover the route only when it is having the data tosend. Examples of N o v 0 1 , 2 0 1 3 reactive routing protocols are the Dynamic Source Routing (DSR) and Ad Hoc On-demand Distance Vector routing (AODV) [6].
In order toget the advantages of both table driven and on-demand routing protocols, one can combine theapproaches to form a hybrid protocol. Various protocols have been developed under thesecategories over the years. Zone Routing Protocol (ZRP) offers a hybrid protocol [9].
The next sub-section describes the basic features of these protocols. For a more detailed description, the reader is referred to the respective RFCs [15][16][17].

IV. DYNAMIC SOURCE ROUTING (DSR)
The Dynamic Source Routing (DSR) protocol is a simple and efficient routing protocol designed specifically for usein multi-hop wirelessAdHoc networks of mobile nodes [6]. Also it is a reactive, on-demandrouting protocol based on a method known as source routing, a routing technique in which the sender of the packet determines the complete sequence of the nodes through which to forward the packet. The sender explicitly lists this route in the packet"s header, identifying each forwarding "hop" by the address of the next node to which to transmit the packet on its way to the destination host [1].
Nodes in DSR "learn" and cache multiple routes to each destination (either as a response to a request, forwarding, or overhearing) to be used in case of route loss. In addition, this also helps in reducing routing overheads. The on-demand feature of DSRreduces the bandwidth use, especially in cases where the mobility is low.
The DSR protocol consists of two mechanisms: Route Discovery and Route Maintenance, which work together to enable nodes to discover routes to destinations, and to maintain the routes to prevent any loss [6,10].

Route Discovery Mechanism in DSR
When a source node S wants to send message to the destination node D, as shown in Figure 1, it checks its routing table. If no route is found to the destination, then node "S" activates the route discovery procedure to find the route to the destination, node "D" by broadcasting the RREQ packet to its neighbors (A, E, F). The intermediate nodes (A, E, F) on receive the RREQ packet rebroadcast the packet to its neighbors until the request reaches the requested node "D". Any node participating in route discovery can learn routes from passing packets and gather this routing information into its route cache.When the destination node D receives two or more RREQ packets from the same source through different routes, it checks its cached routes tofind the two best routes based on the numberof hopes. The route which has least number of hops it becomes primary<S, F, G>.Otherwise it uses the reverse of the route that has been used by the request message to send the reply [6,18].

Route Maintenance Mechanism in DSR
Route Maintenance is used to detect if the network topology has changed such that the link used by this packet is broken, and has to inform the source about the route breakage by sending a route error message (RERR). When the source node is informed about the route breakage, it removes the broken route from its cache and starts repairing the route to the destination by looking for another route in its cached routes. If no route is found tothe destination, then the source starts a new route discovery mechanism [1,6].

V. AD HOC ON-DEMAND DISTANCE VECTOR(AODV)
AODV is "on-demand" ora destination based reactive protocol, whichprovides on-demand route discoveryand maintain a route to another node until the two needs to communicate. This protocol inherits the feature of route discovery from DSR. However, AODV resolves the problem of large headers found in DSR. AODV maintains routing tables on the nodes instead of including a header in the data packet. The routing decisions are made using distance vectors, i.e. distances measured in hops to all available routers [6,19].
Route finding in AODV is based on a route discovery cycle involving a broadcast network search and a unicast reply containing discovered paths. Each node maintains a sequence number which saves a time stamp and for loop freedom, N o v 0 1 , 2 0 1 3 and a routingtable which contains routes to destinations. Sequence numbers are used to determine the freshness of routes (thehigher the number, the fresher the route which allows the older one to be discarded). Since this is an ondemand distance vector scheme (routersmaintain distances of only those destinations that they need to contact or relay information to). Each active routeis associated with a lifetime stored in the table; if a route is not utilized within the lifetime period, the route is marked as invalid and later on, removed. Otherwise, each time the route is used, the lifetime period is updated so that the route is not prematurely deleted [8,9].
The two main mechanisms used by the AODV protocol to establish and maintain the connection are [6]: 1. Route Discovery mechanism.
Four types of control packets are used to discover the path in the network if required. As shown in Figure 2 [Chakeres and Royer] [20].

Route Discovery mechanism in AODV
Thesource node initiates the route discovery process in the same way as in DSR. An intermediate node may replywith a route reply (RREP) only if it knows a more recentpath than the one known by the sender node to thedestination. A new routerequest generated by the sender node is tagged with ahigher sequence number and an intermediate node thatknows the route to the destination with a smaller sequencenumber cannot send the RREP message. Forward links aresetup when a RREP travels back along the path taken byRREQ. So the routing table entries are used to forward thedata packet and the route is not included in the packetheader [19].

Route Maintenance mechanism in AODV
In AODV, the role of route maintenance is to provide feedback to the sender in case a link breakage occurs, to allow the route to be modified or re-discovered as follow: If a source node moves, then it must rediscover a new route. If an intermediate node moves, its upstream neighbor notices the move and propagates a link failure route error (RERR) message to each of its active upstream neighbors. These nodes in turn propagate RERR, and so on, until the source node is reached. The source node may re-initiate route discovery if a route is still desired.AODVprotocol uses hello messages , which are periodic local broadcasts made by a node to advertise its presence in its neighborhood.. A node learns that a link is broken when it does not receive a HELLO for a predetermined time [6,20,21].

VI. OPTIMIZED LINK STATE ROUTING PROTOCOL (OLSR)
It is a table-drivenproactive routing protocol. It wasdeveloped for mobile Ad Hoc networks. So the routes are alwaysimmediately available when needed.OLSR is an optimization of link-state routing;with the message flooding in OLSR is optimized topreserve bandwidth. The optimization is based on atechnique called Multipoint Relaying. Under this protocol, allthe nodes contain pre-computed route information aboutall the other nodes in network. This information isexchanged by protocol messages periodically.Each node selects a set of its neighbor nodes as MultiPoint Relays (MPRs). Only those nodes selected as MPRsare responsible for forwarding the Control Traffic. MPRsare selected such that 2-hop neighbors can be reachedthrough at least one MPR node. MPRs are used to form the route from starting node todestination node in MANET. The purpose of selecting MPRs is to reduceflooding overhead and provide optimal flooding distance as shown in Figure 3  OLSR uses two kinds of control messages: Hello andTopology Control (TC). Hello messages are used forfinding information about the link status and the host"sneighbors. With the Hello message, the Multipoint Relay(MPR) Selector set is constructed which describes which neighbor has chosen this host to act as MPR. The Hello messages are sent only one hop away but the TC messages are broadcast throughout the entire network. The TC messages are broadcast periodically and only the MPR hosts can forward the TC messages [6]. N o v 0 1 , 2 0 1 3

VII.RANDOM WAYPOINT MOBILITY MODEL
The Random Waypoint (RWP) mobility model is the simplest and most widely available model for MANETS studies [1]. The algorithm for RWP is as follows:

1.
A node chooses a random destination anywhere in the network area.

2.
The node starts moving towards the destination witha velocity randomly chosen from a speed vector[Vmin,Vmax].

3.
After reaching the destination, the node stops at the destination for a duration specified by "pause time" parameter, which is the same for all nodes.

4.
All nodes repeat this procedure until the simulationends.

VIII. PROTOCOLS PERFORMANCEMETRICS
The MANET working group within the International Engineering Task Force(IETF) has defined some unique characteristics of Ad Hoc networks for standardization in Internet Drafts or Request for Comments(RFC 2501)(request for comment) [24], which describe very nature of Ad Hoc networks .The following four quantitative [5,19] performance metrics are used for this study: 1. Network Throughput:It gives the fraction of the channel capacity used for useful transmission and is defined as the total number of packets received by the destination. It is in fact a measure of the effectiveness of a routing protocol.

Average packettransmission time delay:
This includes all possible delays caused by buffering during route discovery latency, queuing at the interface queue, retransmission delays at the MAC, and propagation and transfer times.

Routing overhead:
It is the total number of routing packets transmitted during the simulation.. All the packets sent or forwarded at network layer are considered as routing overhead.

Packettransmissionratio (PTR):
The ratio of the data packets delivered to the destinations to those generated by the traffic sources.
All these metrics are most widely used for representing performance of routing protocols because higher data delivery, lower control overhead and lower delay are always desirable. The essential parameters, which we have varied in our simulations, are mobility speed and pause time and network size [5].

IX. SIMULATION SCENARIO AND PARAMETERS SETUP
OMNeT++ version 4.1Simulator was used to model MANET scenarios as shown in Figure 4.The model was designed with an extensive set of parameters and was used to evaluate, analyze, and compare the performance of three routing protocols (AODV, DSR, and OLSR).
In this paper the studied scenarios have been done on network size of 50 and 100 nodes; node are distributed randomly in a grid network topology and mobile nodes moving with velocities ranges from 1 m/s to 30 m/s using Random Waypoint mobility models with two values of pause time (0-3) and (200-300) seconds. N o v 0 1 , 2 0 1 3

Figure 4: MANET Scenario
Thesimulation of all scenarios has been done for 500 seconds in a simulation area of 500mX500m with CBR (UDP) traffic pattern. The source-destination pairs are spread randomly over the network. The size of data packets is 1024 bytes withpacket rate of 10 packets per second. Simulations are carried out by network load with ten source nodes send data to ten destination nodes at the execution time. To overcome the effect of randomness in the output we have taken the averages of the results to get their realistic values.
The simulation parameters are shown in Table 1.

X. EXPERIMENTAL RESULTS AND PERFORMANCE COMPARISION DISCUSSION
Two tests were carried out to evaluate andanalyze the impact of node speed variationson theperformance of AODV,DSR and OLSR at using 50 and 100 network nodes density for grid network structure. In order to show the effect of pause time variation on the performance of routing protocols,for each network size, two experiments were runfor two values of pause time (0-3) and (200-300) seconds.(0 -3) s means continuously moving (high mobility)and (200-300) means stable or low mobility. When the nodes are continuously moving (0 s pause time) the number of link changes is very high and decreases with increase in pause time and converge to 0 (200s pause time). At this stage the network becomes stable. According to the performance metrics that have been explained in section VIII, the experimental results and performance evaluation of the three protocols are as follow:

Parameter name value
Size of playground 500 m * 500 m  Figure 5 shows that when network size is 50 nodes, DSR is able to achieve slightly more throughput than AODV in less stressful situations (low speeds and long pause time). This could indicate that local repair can be used more efficiently to recover from loss at low speed due to the use of route caching in DSR.While when node speed increases in stressful situations (i.e., larger number of nodes and higher mobility), the AODV protocol slightly outperforms the DSR protocol with respect to the network throughput as shown in Figure 9 for network size of 100 nodes. This is because the AODV protocol works better in stressful situation with a heavy network load and high mobility, while DSR work better in network with low density network. However, OLSR gives comparatively lower throughput as the large number of routing bits is required for keeping update network topology and fresh routes; increase in overhead reduces the throughput which will in turn have effect on the delivery ratio of packets due to the fact that when topology changes more quickly protocol control messages will increase more which consumes more bandwidth that affects network throughput and packet delivery ratio as shown in Figure 6. In reactive protocols, if there is no route to a destination, packets to that destination will be stored in a buffer while a route discovery procedure is conducted which takes some time. Moreover, OLSR continuously maintains routes to all destinations in the network. When link break occurs, it can quickly find a new route to the destination since the routing table has routes for all available hosts in the network.Figure7 also shows thatthe packet transmission time delay of DSR is the lowest as compared to AODV. This is likely because the route acquisition procedure in DSR allows more routes to be detected and cached than in AODV, which obtains a single route per RREQ. With a higher mobility and high traffic load, links are more frequently broken, for DSR, packets wait less time during route acquisition than with AODV. In OLSR, routes are known well in advance; hence the average packet transmission delay is less than that of AODV and DSR. Protocol overhead represents total number of bytes generated by a routing protocol for routing operationswithin a network. Increases in protocol overhead will negatively affect the network performance by consumingbandwidth. Figure 8 shows that DSR creates the least overhead as compared to both AODV and OLSRat high mobility (short pause time) ,this is due to the way routes are detected in DSR; also the route acquisition procedure in DSR allows more routes to be detected and cached with the same RREQ than in AODV (which obtains a single route per RREQ), and this reduces the protocol overhead in DSR.OLSR shows a higher difference as compared to the other two protocols.Because OLSR is atable-driven protocol, and the proactivity nature of OLSR makes it exchange topology information with other nodes of the network regularly and periodically, this increases the overhead. This figure demonstrates also that with high mobility, DSR overhead is increasing more, this is because of source routing mechanism of DSR.  Figure 9 demonstrates that AODV outperforms the other two protocols (DSR and OLSR) for all pause time values considered.Also this figure showsthat the performance of DSR is significantly poorer as node density increases. DSR exhibits severe degradation in performance with higher node density. This is an effect of the higher network load caused by the source routes carried in all data packets, therefore, at higher mobility and when using a heavy network load, AODV appears to be a better protocol choice than DSR for MANET applications. However, the OLSR performance is less than the performance of both DSR and AODV in dense and unstable network (short pause time). This is because a large amount of overheads generated to keep network updated, this overhead consumes the network bandwidth, consequently degrades the network performance.This figure demonstrates also that the throughput's decrease percentage is slight at about 1.8% when using the AODV routing protocol, and less than the throughput"s percentage decrease when using other protocols (DSR, by about 1.9% and OLSR, by about 2.4%). Packet Transmission Ratio (PTR) describes the loss rate that will be seen by the transport protocols, which in turn have affect the maximum throughput that the network can support. For each protocol, the network PTR starts to decrease with an increase in node speed. The PTR shown in Figure 10 reveals that AODV shows better performance than DSR in both cases (short and long pause times) because of the effect of the higher network load caused by the source routes for DSR carried in all data packets. Therefore, at higher mobility (short pause time), packet loss due to an increase in speed of the mobile nodes decreases the packet transmission ratio (PTR) which leads to make AODV and OLSR more robust than DSR Figure 10: packet transmission ratio (PTR) vs. node speed (m/s) for 100 nodes Figure 11, shows a strong rise in packet transmission time delay of DSR as the number of nodes increases from 50 to 100. This may be due to large overhead packets in DSR as compared with AODV. Higher node density increases the number of neighboring nodes and that causes more route reply messages to the source node with large packet header used in DSR, and also stale routes are held for a long time thus increasing time delay. AODV shows the lower delay than DSR for high node speed. In AODV better technique is used, as the destination replies only to the first arriving RREQ. This automatically favors the least congested route instead of the shortest route. Whilst, DSR replies to all RREQs, which makes it difficult to determine the least congestion route. OLSR reacts well to the link breakage because OLSR continuously maintains routes to all destinations in the network and when link break happens, it can find a new route, there is no data buffering for invalid routes to the destination faster than DSR. With the higher node density, overall transmission time delay for all three protocols increases as the number of hopes increases. In these cases the packet needs to cover long distance to reach the destination. AODV performs better in stressful situation (high load or high mobility).  Figure 12 shows the protocol overhead versus node speed for different pause times. It demonstrates that the DSR overhead is lower than OLSR and AODV overheads for both pause time values in case of low speed. When there is an increase in the speed of the node, the OLSR overhead still remains higher than DSR and AODV, likely due to the proactive nature of OLSR. On the other hand, DSR overhead is higher than the AODV overhead, because DSR is based on source routing algorithm and every data packet must hold the entire route from the source to the destination in its header. In this stage poor performance with using DSR may be explained by the aggressive route caching technique built in this protocol, but for a higher node density, the benefit of caching routes seems to be lost.

XI. CONCLUSION
In this paper, the simulation results show some important differences among the performance of MANETrouting protocols in terms of network throughput, packet delivery ratio, routing overhead and packet transmission time delay. The presence of high mobility implies frequent link failures and each routing protocol reacts differently during link failure.
Under varying speeds and with a varying network size, AODV outperforms DSR with respect to network throughput and packet transmission ratio in stressful situations (heavy traffic load and/or high mobility, network size is large). DSR protocol, however, consistently generates less routing overhead than AODV.It has to be noted that for the light network situation, DSR shows better performance. The OLSR is outperformed by both other protocols for all mobility and load test situations.
DSR outperforms AODV with respect to packet transmission time delay for all speed values due to the caching mechanism used by DSR. On the other hand, OLSR provide a better quality of service than both DSR and AODV, this is dueto OLSR proactive naturewhich makes routes to every destination are always available and up-todate resulting minimumtime delay. While For on-demand protocols, DSR and AODV, the source node has to wait for the route to bediscovered before transmission. This delay in route discovery might be unsuitable for real-time applications. However, this delay advantage of OLSR is on the expense of high protocol overhead which consumes some of the available bandwidth and consequently reduces network throughput.