You are on page 1of 25


Seminar Report on

A Unified Analysis of Routing Protocols used in MANETs

Submitted in partial fulfilment of the requirement for the award of the degree of BACHELOR OF TECHNOLOGY in ELECTRONICS & COMMUNICATION ENGINEERING

Under the Guidance of: Astt.Prof.

Submitted By: Trivendra Joshi 060070102154

Department of Electronics & Communication Engineering Dehradun Institute of Technology, Dehradun 2012-13(ODD SEMESTER)



I hereby certify that the analysis work on research paper which is being presented in the seminar report of PEC-754 entitled A Unified analysis of routing protocols used in manets in partial fulfilment of the requirements for the award of the Degree of Bachelor of Technology in Electronics and Communication and submitted in the Department of Electronics and Communication Engineering. Dehradun Institute of Technology, Dehradun, Uttarakhand Technical University Uttarakhand, is an authentic evidence of my own analysis work under the guidance of Astt.Prof. Mr.,Department of Electronics and Communication Engineering, Dehradun Institute of Technology, Dehradun. TRIVENDRA JOSHI

This is to certify that the above statement made by the candidate is correct to the best of our knowledge.

DATE:29/11/2012 Guide name with destination Mr. Kuldeep Chaudhary (Astt. Prof) Seminar Coordinators Mr. Dhruva Chaudhary Mr. Udit Gupta Head of Department Prof. (Dr.) Sandip Vijay

1. List of figures 2. Abbriviations used 3. Introduction about topic 4. Theory 5. Analysis and Methodology used 6. Result discussion 7. Area of application and present research status 8. Conclusion 9. References

Page No.
4 4 5-6 7-12 13-14 15-21 22-25 25 26-28

1.List Of Figures
Fig. no. 1 2 3 4 5 6 7 8 9 Title A multi-hop data transmission scenario in wireless sensor network Simulation Parameter Comparision on transmission cost when p routing = 25% Comparision on transmission cost when p routing = 50% Comparission on transmission cost when p routing = 75% Comparision on end-to-end packet delay Comparision on energy consumption per packet Basic modle of a two way relay system Example of a multi hop relay system Page no. 8 17 18 18 19 20 20 22 23

2.Abbreviation used
1. AODV 2. MANET 3. MAC 4. ARQ 5. COMAC 6. CBF 7. MIMO 8. PRR 9. SNR 10. CSMA 11. SIFS 12. DIFS 13. NAV 14. MANET 15. RDSTC 16. QOS 17. EZW 18. ROI Access On Demand Vector Routing Protocol Mobile Ad-Hoc Networks Medium Access Control Automatic Repeat Request Cooperative Medium Access Control Cluster Based Forwarding Multiple Input Multiple Output Packet Reception Rate Signal to Noise Ratio Code Sense Multiple Access Short Inter Frame Signal Distributed Inter Frame Signal Network Allocation Vector Mobile Ahoc Network Randomized Distributed Space Time Code Quality of Service Embeded Zerotee Wavelet Region of interest

3. Introduction to Topic:
Mobility brings fundamental challenges to the design of routing protocols in mobile ad hoc networks (MANETs). The mobility of nodes implies that the routing protocols of MANETs have to cope with frequent topology changes while attempting to produce correct routing tables. To accomplish this, two types of routing protocols have been proposed proactive routing and reactive (or on-demand) routing. Proactive routing protocols provide fast response to topol- ogy changes by continuously monitoring topology changes and disseminating the related information as needed over the network. However, the price paid for this rapid response to topology changes is the increase in signaling overhead, and this can lead to smaller packet-delivery ratios and longer delays when topology changes increase. In the worst case, broadcast-storms" [1] can result in congesting the entire network. Reactive routing protocols operate on a need to have basis, and can, in principle, reduce the signaling overhead. However, the long setup time in route discovery and slow response to route changes can offset the benefits derived from on-demand signaling and lead to inferior performance. Given these striking differences between proactive and on- demand routing in MANETs, a basic question to ask is whether one routing protocol always performs better than the other. Extensive simulations have already given the answer: both of routing algorithms exhibit advantageous and inferior performance to the other one depending on different network configurations, particularly with respect to different node mobility, node density and traffic load. Given that proactive and reactive routing schemes for MANETs have relative advantages and disadvantages, com- paring the two is important. Significant work (e.g., [2][5]) has been conducted to evaluate and compare these protocols under network profiles of various mobility and traffic configurations. Such performance comparisons have been mostly conducted via discrete-event simulations. Simulation-based studies of routing schemes are a powerful tool to gain insight on their performance for specific choices of network parameters. How- ever, it is difficult to draw conclusions involving multidimen- sional parameter spaces, because running several simulation experiments for many combinations of network parameters is impractical. Few if any analytical studies have been pursued on this topic, and prior analytical work has been mostly restricted to the analysis or comparison of routing control on routing metric overhead [6][8], information theoretic aspects [9], [10] or multipath routing protocols [11], [12]. Furthermore, these works do not evaluate the effects of signalling overhead on unicast capacity at nodes, and none reveals the underlying connection between protocol performance and network parameters. Hence, the following two questions have remained open: Does there exist a unified analytical framework, simpli-fied but able to provide the characterization of essential behaviors of routing protocols? If there is one such framework, can it also provide network scalability?

This paper provides a unified analytical framework that parameterizes and quantitatively evaluates the performance of routing protocols from a joint characterization of the routing logic and MAC layer. The solution emerges from a combina- torial model that synthesizes the evaluation of both the routing logic and the operation of the MAC layer, with each of them being parameterized individually. The MAC layer at nodes is as well as simplified as a two-customer queuing model, where the packet loss probability and delay at nodes can be effectively computed. When an- alyzing the performance of the MAC layer, we consider the cases of a scheduled MAC (TDMA) and a contention-based MAC (802.11 DCF MAC). In the combinatorial model, the computed metrics are synthesized along with the routing logic to produce quantitative measures of the routing protocols in terms of end-to-end packet loss probability and delay. It is important to note that this work does not attempt to model or compare between specific proactive or reactive routing protocols. Rather, the intent is to capture the essential behaviour and scalability limits in the network size of both classes of protocols by quantifying their performance within a unified framework. The simulation results in this paper are not intended to provide an exact match with our analysis; they are provided only as a supporting evidence for the conclusions regarding to protocol behaviors observed in literature. With the aid of simulations, we show that the analytic results agree with the simulation findings. We model each wireless node as a two-customer queue without priority, where the two kinds of customers are unicast and broadcast packets. By analyzing the service time and waiting time in queue for any unicast packet, we derive the transmission delay time for any link; by summing up the transmission delay time for each link of any routing path, we derive the end-to-end delivery delay. The paper is organized as follows. Section II presents related work in the modeling of routing protocols. Section III presents the mobility model, traffic model and simplified mod- els of routing algorithms used in the analysis in Section IV. Section V characterizes the performance of the MAC layer and provides performance metrics in evaluating protocol per- formance. Section VI compares our analytical results against Qualnet simulations based on scenarios on various traffic loads, mobility and node density configurations. The results illustrate the correctness of our analytical framework, which captures the essential behaviors of protocols and is capable of pinpointing the effect of various parameters analytically. Section VII concludes this paper.

4. Theory:
This paper presents a mathematical framework for the evaluation of the performance of proactive and reactive routing protocols in mobile ad hoc networks (MANETs). This unified framework provides a parametric view of protocol performance, which in turn provides a deeper insight into protocol operations and reveals the compounding and interacting effects of protocol logic and network parameters. The parametric model comes from a combinatorial model, where the routing logic is synthesized along with the characterization of MAC performance. Each wireless node is seen independently as a two-customer queue without priority, where the two types of customers are unicast and broadcast packets. The model captures the essential behavior and scalability limits in network size of both classes of routing protocols, and provides valuable guidance on the performance of reactive or proactive routing protocols under various network configurations and mobility conditions. The analytical results obtained with the proposed model are in close agreement with simulation results obtained from discrete- event Qualnet simulations. Parameterizing and comparing the performance of routing protocols analytically is a very complex problem. Conse- quently, the characterization and comparison of routing pro- tocols has been limited to simulation-based approaches [2] [5], [13][16], under various configurations. The performance evaluation metrics used in these simulations or experimentalbased approaches include packet delivery ratio, delay and throughput. Network configurations vary on traffic pattern, mobility and network density. for proactive and reactive protocols to evaluate the individual routing control overhead. Zhou and Abouzeid [7] presented an analytical view of routing overhead for reactive protocols, assuming a static Manhattan-grid network and studied the scalability of reactive protocols. These studies concentrate on the impact of traffic patterns and they also provide [8] a mathematical and simulation-based framework for quantifying the overhead of reactive routing protocols. Zhou and Abouzeid extended their analytical studies to information theoretic techniques to derive analytic expressions for specific metrics, such as control message routing overhead and memory size requirement [9] where entropy is utilized to derive bounds for these metrics. They have made funda- mental contributions [10] toward a rigorous modeling, design and performance comparisons of protocols by deriving lower bounds on the routing protocol. Tsirigos et al. [11], [12] studied multipath routing protocols with the aim of increasing packet delivery ratio in inherently unreliable networks and developed an analytical framework for evaluating them. Their work took mobility and topology changes into considerations. Our work has been inspired by research work in [17] [20] which analyzes and evaluates the performance of routing protocols by constructing mathematical models of the routing operations. Yang et al. [17] set up a mathematical model that imitates the operation of proactive routing and is utilized to optimize the operation of routing protocols in order to strike a balance between protocol overhead and accuracy. However, this work is specific for proactive protocols operating under certain conditions. Lebedev [18] showed an analytical tool for both proactive and reactive protocols and proposed a model to study the operation of two classic reactive (Ad-hoc reactive Distance Vector Routing (AODV) [21]) and proactive (Optimized Link State Routing (OLSR) [22]) protocols in the presence of faulty links. However, this work does not cover other aspects of

network parameters on the performance of routing protocols. Nogales [19] models routing protocols with the consider- ation of the behavior of connections at the link layer and MAC protocol. However, that model is greatly simplified when modeling the packet routing and scheduling process and does not take into consideration traffic load and the topology of the network. Jacquet and Laouiti [20] compare proactive and reactive routing protocols performance by proposing probability-based models. However, their models focus on calculating specific performance metric values, rather than evaluating the general operations of these routing protocols. Their study is based on random graph models that are simplified and idealized, such that mac and link layer failure effects are not taken into consideration. 4.1. NETWORK MODE L 4.1A. Mobility Model Nodes are mobile and initially they are distributed equally over the network. The movement of each node is independent and unrestricted, i.e., the trajectories of nodes can lead to anywhere in the network. For node i V = {1, 2, . . . , N }, let {Ti(t), t 0} be the random process representing its trajectory and take values in Y , where Y denotes the domain across which the given node moves. To simplify the model, we make the following assumption on the trajectory processes. Assumption 1: [Stationarity] Each of the trajectory processes (Ti(t)) is stationary, i.e., the spatial node distribution reaches its steady-state distribution irrespective of the initial location. The N trajectory processes are jointly stationary, i.e., the whole network eventually reaches the same steady state from any initial node placements, within which the statistical spatial nodes distribution of the network remains the same over time. The above assumption is quite fundamental in the sense that it lays the foundation for the modeling of node move- ment. Most existing models, (e.g., random direction mobility models [23], random waypoint mobility models [24], [25] and random trip mobility model [26]) clearly satisfy our assumption. In other words, our assumption ensures that, on the long run, the network converges to its steady state and the stationary spatial nodes distribution can be used in the performance analysis of the network. 4.1.B. Traffic Model We consider a new traffic flow or simply a new session as one that is associated by the arrival of a new application-level session request at a node i with some destination j, j = i, in the network. Traffic flows are randomly generated with uniformly distributed sources and destinations. Long-lived traffic flows are assumed in order to investigate protocol performance under the steady state of nodes mobility and traf- fic distributions. Short-lived traffic flows, reflecting transient behaviors, are beyond the scope of this paper. Furthermore, wellconnected networks are assumed, i.e., if an existing path for any traffic session is broken, there is always an alternative path (with high probability) available to support continuing operations of the traffic flow. 1

4.1.C. Neighbor Sensing Neighbor sensing protocols, such as periodic broadcasts of HELLO messages, are effective approaches used in routing protocols (both proactive and reactive) to detect the availability of links between neighbor nodes. New links are detected when HELLO messages are received from nodes not included in the neighbor list. Existing links are declared as failed if none of HELLO messages from the neighbor node is received during a certain amount of time window. 4.1.D. Routing Protocols Model We provide descriptions of generic proactive and reactive routing protocols, which we believe capture the essential behavior of many designs and implementations of existing routing protocols. However, this analysis, and hence the generic protocols below, does not consider any protocol- specific techniques, such as multi-point relay, local repairs and route caching mechanisms. Note that the alternative path is not necessarily disjoint with the former broken path. 1) Proactive Routing Protocol: In proactive routing proto- cols, every node maintains a list of destinations and updates its routes to them by analyzing periodic topology broadcasts from other nodes. When a packet arrives, the node checks its routing table and forwards the packet accordingly. Every node monitors its neighboring links and every change in its neighbors results in a topology broadcast packet. That is flooded over the entire network. Other nodes update their routing tables accordingly upon receiving the update packet. In a well-connected network, the same topology broadcast packet could reach nodes multiple times and therefore enjoy a good packet reception probability. In the paper, we assume that every node reliably receives topology packets from other nodes. 2) Reactive Routing Protocol: In reactive routing protocols, nodes maintain their routing tables on a needed basis. This implies that when a new traffic session arrives, nodes have to set up the paths between sources and destinations before starting to deliver data packets. The process of path setup is called route discovery. Complementarily, another process called route maintenance is necessary to find an alternative path if a former path was broken. More specifically: a) Route Discovery: a mechanism initiated by a node i upon the arrival of a new traffic session in order to discover a new path to a node j. Node i floods the whole network with route request (RREQ) packets. Upon receiving the RREQ packet, node j sends out a route reply packet (RREP) along the reverse path to i. As a result, node i usually gets a shortest path to node j. b) Route Maintenance: a mechanism by which a node i is notified that a link along an active path has broken, such that it can no longer reach the destination node j through that route. Upon reception of a notification of route failure, node i can initiate a route discovery again to find a new route for the remaining packets destined to j. In reactive routing protocols, each node does not maintain routing tables before a routing task is triggered. They only find a route on demand by flooding the network with RREQs, i.e., before sending data packets sender broadcasts router request and initiates a route discovery

process. If a link breakage is detected during packets delivery, a new RREQ is generated. The main disadvantages of such algorithms are high latency time in finding routes and excessive flooding when traffic load is high. Assume that a successful delivery between source node S and destination node U takes K hops, that at the first step a route discovery is initiated at S and that after time 0 source node S receives a RREP and starts sending data packets. Assume that a link breakage occurs at a relay node F with probability pi and then a new RREQ is generated for F and that it takes time i to restart delivery. Let us suppose that the transmission delay for any link i is Ti . Then the end-to- end delivery delay Dp between S and U can be formulated as Dp = 0 + K (pi i + Ti ). Compared to the local recovery time in proactive routing protocols, the route discovery time in reactive routing proto- cols is much larger.

5. Analysis and Methodology used

5.A. Models Previous studies [2] [9] have pointed out that the simple radio models (e.g., the ideal binary model assumes that there are perfect links within a given communication range, beyond which there is no link) is unlikely to be valid in any realistic environment. Actually, the reliability of a link is influenced by a lot of factors, such as the transmitter-receiver distance, the emitting power and the interference noise. To model the reliability, we choose to use the one derived in [2] and [19], which is based on the log-normal path loss model. The packet reception rate (PRR) p for a transmitter-receiver distance d is given by p(d)={1-exp[-(d)BN/2R]/2}
8f (6)

length in byte. The signal to noise ratio (SNR) at the receiver can be given as where B N is the noise bandwidth, R is data rate in bits,

is the encoding ratio, and f is the frame

(d)= P t -PL(d 0 )- 10log 10 (d/d 0 )+N(0,)-P n


where Pt is the output power, PL(d 0 ) is the power decay for the reference distance d 0 , is the path loss exponent (rate at which signal decays with respect to distance), N(0, ) is a Gaussian random variable with mean 0 and variance 2 (due to multi-path effects), and P n is the noise floor. In this model, low-power wireless links are classified based on transmitter-receiver distance in distinct reception regions: connected (p1), transitional (0 < p < 1), and disconnected (p 0). Then, the communication range R c can be defined as

R c =max{d|P(d)>p thres p thres >0 p tres 0}


We model a sensor network by an undirected graph G = (V, E), where V is the set of sensor nodes and the set of wireless communication links, E, is defined as E={(u,v)V2|uvd u,v R c } H(u)={vV|vu(u,v)E} (9)

The neighborhood set H(u) of a node u is defined as (10)

5.B. Assumptions A wireless sensor network composed of a large number of resource-constrained sensor nodes and several more powerful sinks in an area is considered in this paper. Sensor nodes are randomly distributed in the area and remain stationary after deployment, and all of them have similar capabilities and equal significance. Each sensor node sends its data packets to sink nodes through multi-hop wireless links. All nodes in the network are supposed to implement a CSMAlike MAC (with ARQ) to share the physical medium. Due to link unreliability, the network is assumed to employ a more optimal metric [19] [20] instead of the conventional shortest-pathfirst policy for setting up an optimum routing path (referred to as intended path) between source and sink. Each node knows and keeps record of the PRR between itself and any neighboring node.

6. Result Discussion
6.1 PERFORMANCE EVALUATION: ANALYTICAL RESULTS Based on reasonable assumptions, we compare the energy efficiency of our EECC and the traditional methods, i.e., pure retransmission-based and FEC-based. According to previous researches [23], wireless communication is responsible for the greatest weight in the energy budget when compared to sensing and data processing. Assume M sensor nodes are randomly

deployed in an area of size S. Let E t and E r be the energy consumed by a single transmission and reception respectively. Theorem 1. Let the per hop energy consumption for FEC, pure retransmission and node cooperation are denoted by EFEC, E retrans , and EEECC respectively, we have EFEC E retrans > EEECC. Proof: The energy consumed for each transmission attempted by a sensor node is E signa l=E t +(MR2 c/S-1)Er

Ideally, we assume that a node retransmits for infinite times. Thus, in the retransmission-based mechanism, the expected energy consumption for a successful delivery from s to r is

E retrans = p s,r E single +......+np s,r (1-p s,r )n-1E single = n k=1 kp s,r (1-p s , r )k-1 E single

then (12) can be rewritten as

In the FEC mechanism [4], it is required that

Now we analyze the energy consumption of EECC in three different cases of cooperation scenario. From (13), it is more energy efficient to relay packet on link (t, u) than (r, u). Thus, to simplify the analysis, we only consider the energy consumed for data packet transmission in one hop range, i.e., from s to its next hop relay. Let the cardinality of T and Q is denoted by |T| and |Q| respectively.

where Therefore,we have

where and lQ is the node that relays the packet as the cooperator. Because p l,r > p s,r , we have

Summarizing the above analytical results in (11) ~ (19), the following inequality holds: EFEC E retrans > EEECC. Now we give the performance bound of EECC Lemma 1. In idealistic cases with perfect links (p s,r = 1), the data forwarding in one hop range succeeds at the first time. Thus, the lower bound of per hop energy consumption is E single . Lemma 2. In realistic cases with imperfect links, the packet can be received by the receiver and overheard by the cooperator candidates. The overall reception rate provided by these nodes is Hence,transmission cooperation is more likely to be realized when the node density is relatively high. In sparsely deployed scenarios, we can know that the per hop energy consumption is upper bounded by E retrans when there are few cooperation opportunities.

6.2 PERFORMANCE EVALUATION: SIMULATION RESULTS The performance of EECC was evaluated by conducting simulations via the ns-2 simulator. Here, we adopted a model which is widely used in previous researches to calculate energy

consumption for transmission. To receive l bits at the receiver with a distance of d from the transmitter, energy consumption for a single transmission and reception is given by

where E elec is the electronics energy, f s d2 and mp d4 are the amplifier energy in the free space and in the multi-path fading environment respectively, and d 0 is a distance threshold that depends on the environment. In our simulation experiments, the wireless link model is the same with that in [2], and other parameters are set as shown in Table I.

Fig: 2 The first set of simulation experiments is performed to verify our analysis above in analytical result. Thus, we manually deployed two sensor nodes and a single sink to create a simple2-hop data transmission topology as that has been introduced in analytical result. Different amount of sensor nodes are randomly deployed as the neighbours of the intended data source node to create different cooperation scenarios, i.e., (|T| > 0, |Q| = 0), (|Q| > 0, |T| = 0), and (|T| > 0, |Q| > 0). To make the results more pronounced, for each specific transmission scenario the experiment is carried out independently for multiple times by network redeployment. The performance metric selected is called transmission cost, which is defined as the ratio of the total number of data transmissions by the total number of non redundant packets sent out by the data source. This metric reflects both communication overhead and energy efficiency. To fully investigate on the performance effect by EECC, experiments are performed under different routing reliability conditions (That is, the PRR metric for becoming an intended receiver, p routing , is set as low

(25%), middle (50%), and high (75%) respectively). The simulation results for this experiment set are shown in Fig. 3 to Fig. 5.

We have several observations concerning these results. First, EECC can actually help to reduce transmission cost to some extent, as compared to the pure retransmission mechanism (i.e., (|T| = 0, |Q| = 0)). Such cost reduction is even more obvious and stable when p routing of the intended receiver is relatively low (e.g., p routing = 25%) and when the total amount of cooperator candidates is relatively high (e.g., |T| = 3 or |Q|= 3). On the other hand, node cooperation contributes little to the cost reduction when |T| = 1 and p s,t is extremely low (e.g., Experiment Index = 12, 16, 20 in Fig. 5(a)) or when |Q| = 1 and p q,r is slightly greater than p s,r (e.g., Experiment Index = 9, 10 in Fig. 3(b)). This observation helps to validate our analysis on EECC performance in analytical result. Second, relaying by the receivers cooperator (tT) is more efficient than by the senders cooperator (qQ) in general. The reason is that the receivers cooperator is able to reach the next hop of the receiver directly, while the senders cooperator just substitutes for the original sender to retransmit the packet to the receiver with a relatively better link. We could even notice that it becomes less effective to invoke the cooperation by the senders cooperator when p routing becomes relatively higher, as shown in Fig. 3(b), Fig. 4(b) and Fig. 5(b). Therefore, EECC can be an optional choice if the sender has a good transmission link to the receiver while there are no receivers cooperators.

The above two observations imply that a proper network density can provide EECC with more and better cooperation opportunities to minimize excessive transmission costs caused by unreliable links and packet losses. We argue that it is not so difficult to satisfy this condition

since many applications have already required high density deployment of sensor nodes. In the second set of experiments, EECC is evaluated in the multi-hop transmission scenarios, by measuring end to- end packet delay and energy consumption per packet. The network is densely deployed so that for each intended node, |T| 3 and |Q| 3. The routing metric is the same as that introduced in [19]. The results are averaged from multiple simulation runs. As shown in Fig. 6, EECC significantly reduces end-to-end packet delay as compared to the pure retransmission method. An examination of ns-2 simulation traces reveals that this is attributed to a considerable decrease in the number of packet retransmissions: without EECC, the intended sender will timeout and resend the lost packet. The delay caused by such time-outs is far longer than that introduced by the new NAV setting in EECC. On the other hand, EECC is able to decrease such time-outs by node cooperation on data transmission, and thus effectively reducing the end-to-end delay. Meanwhile, the decrease in retransmissions also results in the energy saving for packet transmission, as shown in Fig. 7. These results justify the application of EECC in large-scale sensor networks with energy constrained sensor nodes and unreliable wireless links.

7. Area of applications and present research status

The research in the area of cooperative communications dates back to the pioneering work in the 1970s, where the capacity of relay channels was studied for the problem of information transmission over three terminals. Then the channel capacity of relay networks over non-faded channel was examined. Because of its advantage over conventional one-way communications, the relaying concept has gained great attention in recent research for an extension to fading channels .Many important aspects of relay networks have been extensively studied. For example, the capacity of relay networks over Rayleigh fading channels has been investigated .The diversity-multiplexing tradeoff of DF and AF relays has been investigated . User cooperation which is the generalization of relay networks to multiple sources has been investigated .Relaying/cooperation has been shown to offer a performance enhancement in terms of the capacity .Although the problems of relaying and cooperations have been studied for years in many aspects such as communications, signal processing, networking, and information theory, they still attract the research community as a new paradigm for wireless and mobile networks. Recently, the relaying method has been introduced in the WiMAX standard and is expected to spread into many other commercial standards .

Multi-hop Relay Networks

The multi-hop communication in relay networks is a very promising approach to improve the transmission coverage of cellular and ad hoc networks

Fig.8 Basic modle of a two way relay system In multi-hop relay networks, a transmission between a source node and destination node is composed of multiple hops as illustrated in Fig. 8. Multihop transmission can significantly reduce the transmit power compared to direct communications. In addition, because of transmit power constraints, multi-hop transmission also leads to remarkable coverage extensions by dividing a total end-to-end transmission into a group of shorter paths. On the other hand, due to the rigorous effect of fading and shadowing, the lack of line-of-sight can severely degrade the system performance. Using multi-hop transmission can overcome this dead-end spots problem and thus another advantage of multi-hop relaying has been pointed out, especially for rural areas with low traffic density and sparse population. In these areas, there is no economic benefit in building the whole cellular networks but multi-hop relay networks. Hence, the multi-hop technique is a promising candidate to meet the desirable goal of contemporary wireless systems with respect to providing seamless communications.

Fig.9 Example of a multi hop relay system Mobile Multimedia over Cooperative Communications Future mobile communication systems are expected to satisfy the increasing demand for services with stringent constraints on quality of service (QoS) such as mobile multimedia, mobile video applications, and mobile streaming on-demand. Due to the source encoding that is usually imposed to multimedia applications, e.g., speech, audio, image, and video, some parts of the produced multimedia format have a higher priority for the signal reconstruction than others. Specifically, errors to headers and markers of such formats that may be induced during transmission over the radio channel can severely degrade the multimedia quality regardless of the reception of the remaining parts. A coded image sequence based on the wavelet transform, so-called embedded zerotree wavelet (EZW), is divided into two priorities. The higher priority contains the most significant information of image pixels consisting of the wavelet coefficients of dominant pass and subordinate pass extracted from several early scanning steps. The lower priority conveys less important information. Using cooperative communications, two preassigned relays cooperate to form a distributed-Alamouti code which help the source to transmit the low priority sequence of the image over error-prone wireless channels. Among the whole relay nodes in the networks, two relay nodes are selected which provide the maximum signal-tonoise ratio to convey the higher priority. With this strategy, unequal error protection is implicitly achieved for image transmission without controlling power or varying modulation and coding at the transmitter at the expense of increasing networks overhead (i.e. feedback information). The

selection of relays needs limited feedback, hence, making the proposed scheme feasible for practical applications. Apart from prioritized data structures caused by source encoding techniques, the actual content of a multimedia application may result in some parts being more important to the end-user than others. In particular, as far as context or content of images is concerned, the region of interest (ROI) prioritizes those areas that attract more attention over the background (BG) content. Viewers will commonly pay attention to ROI rather than BG. Accordingly, it appears to be more annoying to observe distortions in the ROI of an image although the BG has high quality. Hence, it is beneficial from a perceptual quality point of view to rather improve the quality of an ROI than spending too much resources on the BG. Clearly, this is particularly helpful for wireless imaging and mobile video services where bandwidth is scarce and expensive. For image transmission, for example, a more efficient use of available radio resources such as bandwidth and power may be unequally allocated between ROI and BG in the way that the former receives a higher portion. In other words, resources could be controlled to increase the quality of the portion of an image that would be of most interest to the viewer. For instance, unequal error protection for ROI coded images over fading channels has been proposed .A novel ROI-based image transmission scheme for wireless fading channels to improve the quality of ROI using cooperative diversity is proposed. Specifically, image communications with three terminals, i.e., source, relay, and destination, is considered. In the first-hop transmission, the source broadcasts the ROI frame to both the relay and the destination. In the second-hop transmission, the source sends the BG frame to the destination while the relay assists to enhance the quality of the ROI at the destination by forwarding the ROI received from the first-hop transmission. With this transmission scheme, the relay does not consume the power used to transmit BG as in the conventional communications. Therefore, to satisfy the total power constraint, the relay can allocate more power for the ROI-based image transmission by using the remaining power reserved for the BG. At the destination, the ROI frames from the first-hop and second-hop transmission are combined using the MRC technique to enhance the quality of ROI. It has been shown that the proposed scheme significantly improves the ROI and the quality of the image compared to conventional approaches. Using relays to transmit videos has recently also gained an increased interest in the research community . Specifically, a video multicast strategy for a two-hop cooperative transmission

scheme has been proposed in .A H.264/SVC encoded video with two layers, namely, base and enhancement layers is considered so that receptions of more layers lead to better video quality. The randomized distributed space-time codes (RDSTC) are used to provide users different video quality based on their channel conditions. The performance of the proposed relaying scheme has been shown to outperform a conventional multicast systems.

8. Conclusion:
We presented an analytical framework to evaluate the behavior of generic reactive and proactive protocols. In the model, the operation of the routing protocol is synthesized with the analysis of the MAC protocol to produce a parametric characterization of protocol performance. The effectiveness and correctness of the model are corroborated with extensive simulations. The model enables in-depth understanding of routing protocol performance, and points out the need to design routing protocols that are capable of confining signaling overhead to those portions of the network where the routing information is needed, in order to operate efficiently under different types of mobility and traffic patterns. A similar conclusion can be derived by looking at the capacity of ad hoc-networks under different types of information dissemination [39]. In our future work, we plan to incorporate realistic physical layer effect into our modeling framework to make it more practical and comprehensive.

9. Refrences:
[1] S.-Y. Ni, Y.-C. Tseng, Y.-S. Chen, and J.-P. Sheu, The broadcast storm problem in a mobile ad hoc network," in ACM Mobicom 1999, NY, 1999, pp. 151-162. [2] S. Das, R. Castaneda, J. Yan, and R. Sengupta, Comparative perfor- mance evaluation of routing protocols for mobile ad hoc networks," in 7th International Conf. Computer Commun. Netw. (IC3N), Lafayette, Lousiana, USA, Oct. 1998, pp. 153-161. [3] F. Bertocchi, P. Bergamo, G. Mazzini, and M. Zorzi, Performance comparison of routing protocols for ad hoc networks," in IEEE GLOBE- COM, vol. 2, San Fransisco, CA, USA, Dec. 2003, pp. 1033-1037. [4] S. R. Das, C. E. Perkins, E. M. Royer, and M. K. Marina, Performance comparison of two on-demand routing protocols for ad hoc networks," IEEE Personal Commun. Mag. special issue Ad hoc Netw., vol. 8, no. 1, pp. 16-28, Feb. 2001. [5] J. Broch, D. A. Maltz, D. B. Johnson, Y. C. Hu, and J. Jetcheva, A performance comparison of multi-hop wireless ad hoc network routing protocols," in Proc. 4th Annual ACM/IEEE International Conf. Mobile Comput. Netw., New York, NY, USA, Oct. 1998, pp. 85-97. [6] L. Viennot, P. Jacquet, and T. H. Clausen, Analyzing control traffic overhead versus mobility and data traffic activity in mobile ad-hoc network protocols," ACM/Springer Wireless Netw., vol. 4, no. 10, pp. 447-455, July 2004. [7] N. Zhou, H. Wu, and A. A. Abouzeid, Reactive routing overhead in networks with unreliable nodes," in Mobicom03, San Diego, Sept. 2003. [8] N. Zhou, H. Wu, and A. Abouzeid, The impact of traffic patterns on the overhead of reactive routing protocols," IEEE J. Sel. Areas Commun., Special Issue Mobile Ad-Hoc Netwo., vol. 23, no. 3, pp. 547-560, Mar. 2005. [9] N. Zhou and A. Abouzeid, Routing in ad hoc networks: A theoretical framework with practical implications," in Proc. 24th Annual Joint Conf. IEEE Computer Commun. Societies (INFOCOM), vol. 2, Miami, FL, USA, Mar. 13-17 2005, pp. 1240-1251.

[10] , Information-theoretic lower bounds on the routing overhead in mobile ad-hoc networks," in Proc. IEEE International Symp. Inf. Theory (ISIT), vol. 2, Yokohama, Japan, June 2003, pp. 455. [11] A. Tsirigos and Z. J. Haas, Analysis of multipath routing, part 1: The effect on the packet delivery ratio," IEEE Trans. Wireless Commun., vol. 3, no. 2, pp. 500-511, Mar. 2004. [12] , Analysis of multipath routing, part 2: Mitigation of the effects of frequently changing network topologies," IEEE Trans. Wireless Commun., vol. 3, no. 1, pp. 138-146, Jan. 2004. [13] P. Johansson, T. Larsson, N. Hedman, and B. Mielczarek, Scenario- based performance analysis of routing protocols for mobile ad-hoc networks," in Proc. 5th ACM/IEEE MOBICOM, Seattle, WA, USA, 1999, pp. 195-206. [14] S. J. Lee, M. Gerla, and C. K. Toh, A simulation study of table-driven and on-demand routing protocols for mobile ad hoc networks," Netw., IEEE, vol. 13, no. 4, pp. 48-54, 1999. [15] J.-W. Chen, Y.-Z. Lee, D. Maniezzo, and M. Gerla, Performance comparison of aodv and oflsr in wireless mesh networks," in IFIP MEDHOCNET, Lipari,Italy, June 2006. [16] A. Al-Maashri and M. Ould-Khaoua, Performance analysis of manet routing protocols in the presence of self-similar traffic," 31st IEEE Conf. Local Computer Netw., vol. 0, pp. 801-807, Jan. 2006. [17] P. long Yang, C. Tian, and Y. Yu, Analysis on optimizing model for proactive ad hoc routing protocol," in Military Commun. Conf. (MILCOM), vol. 5, Atlantic, NJ, USA, Oct. 2005, pp. 2960-2966. [18] D. Lebedev, A framework for the comparison of aodv and olsr protocols," Technol. Advanced Heterogeneous Netw., pp. 98-111, 2005. [19] I. Bacarreza Nogales, Model and performance analysis of mobile ad- hoc wireless networks," in 17th International Conf. Radioelektronika, Prague, Czech Republic, Apr. 2007, pp. 1-6. [20] P. Jacquet and A. Laouiti, Analysis of mobile ad-hoc network routing protocols in random graph models," 2000. [Online]. Available: [21] C. Perkins and E. Royer, Ad-hoc on-demand distance vector routing," in Proc. 2nd IEEE Workshop Mobile Computing Syst. Appl., vol. 2, pp. 90-100, 1999. [22] P. Jacquet, P. Muhlethaler, T. Clausen, A. Laouiti, A. Qayyum, and L. Viennot, Optimized link state routing protocol for ad hoc networks," IEEE INMIC., pp. 62-68, 2001. [23] C. Bettstetter, Mobility modeling in wireless network: categorization, smooth movement and border effects," ACM Mobile Computing Com- mun. Review, vol. 5, pp. 55-67, Jan. 2001. [24] W. Navid and T. Camp, Stationary distributions for the random waypoint model," IEEE Trans. Mobile Computing, vol. 3, no. 1, pp. 99-108, 2004. [25] J. Yoon, M. Liu, and B. Noble, Sound mobility models," in ACM Mobicom 2003, San Diego, 2003. [26] J.-Y. L. Boudec and M. Vojnovic, Perfect simulation and stationarity of a class of mobility models," in INFOCOM 2005, Miami, 2005. [27] G. Rodolakis, Analytical models and performance evaluation in massive mobile ad hoc networks - chapter 2.2," Ph.D. dissertation, Ecole Polytechnique, 2006. [28] L. Kleinrock, Queuing Systems, Volume 1: Theory. John Wiley & Sons, Inc, 1975. [29] X. Wu, H. R. Sadjadpour, and J. J. Garcia-Luna-Aceves, Modeling of topology evolutions and implication on proactive routing overhead in manets," Elsevier Computer Commun. J., Special Issue Algorithmic Theoretical Aspects Wireless Ad Hoc Sensor Netw., 2007. [30] D. P. Heyman and M. J. Sobel, Stochastic Models in Operations Research. McGraw-Hill, NY, 1982, vol. I. [31] T. Clausen and P. Jacquet, Optimized Link State Routing Protocol (OLSR). IETF RFC 3626, 2003.

[32] C. E. Perkins and E. M. Royer, Ad hoc on-demand distance vector routing," in Proc. 2nd IEEE Workshop Mobile Computing Syst. Appl., New Orleans, Feb. 1999, pp. 90-100. [33] X. Wu, H. R. Sadjadpour, J. J. Garcia-Luna-Aceves, and H. Xu, A hybrid view of mobility in manets: Analytical models and simulation study," Elsevier Computer Commun. J., Best Paper Series Issue Perf. Evaluation Commun. Netw., vol. 31, pp. 3810-3821, 2008. [34] X. Wu, H. R. Sadjadpour, and J. J. Garcia-Luna-Aceves, Link lifetime as a function of node mobility in manets with restricted mobility: Modeling and applications," in WiOpt 2007, Apr. 2007. [35] I. Rubin, Access-control disciplines for multi-access communication channels: Reservation and TDMA schemes," IEEE Trans. Inf. Theory, vol. 25, no. 5, pp. 516-536, 1979. [36] R. Nelson and L. Kleinrock, Spatial TDMA: A collision-free multihop channel access protocol," IEEE Trans. Commun. [legacy, pre-1988], vol. 33, no. 9, pp. 934-944, 1985. [37] P. Gupta and P. R. Kumar, The capacity of wireless networks," IEEE Trans. Inf. Theory, vol. 46, no. 2, pp. 388-404, Mar. 2000. [38] T. A. F. Kostas, A. H. Haddad, and C. C. Lee, Performance analysis of a dynamic bandwidth allocation algorithm for a mixed traffic tdma network," IEEE Trans. Veh. Technol., vol. 56, no. 4, pp. 2306-2320, 2007. [39] Z. Wang, H. R. Sadjadpour, and J. J. Garcia-Luna-Aceves, A unifying perspective on the capacity of wireless ad hoc networks," in IEEE Infocom 2008, Apr. 2008.