JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 9, ISSUE 2, OCTOBER 2011 25
Effect of Mobility on Energy Consumption in DSR and AODV Protocols in Mobile Ad hoc Networks Pinki Nayak, Rekha Agarwal, and Seema Verma Abstract—There are two types of wireless networks: infrastructured wireless network and infrastructureless wireless network, also called Mobile Ad-Hoc Network (MANET). The MANETs are also suitable when network setup is difficult, costly and required to be done quickly within a short period. They are characterized by nodes which are moving continuously in a confined area. Energy conservation becomes very important in MANETs as most of the nodes are powered by a battery source which has limited energy reservoir and some times it also becomes very difficult to recharge or replace the battery of the nodes. Thus, lifetime of the nodes show strong dependence on the lifetime of the batteries. In this paper, we analyze the relative energy consumption of two popular on-demand routing protocols in MANETs, AODV (Ad hoc On-demand Distance Vector) and DSR (Dynamic Source Route). Simulations are carried using NS2 Network Simulator under different mobility scenarios. It is shown that DSR performance is better under low mobility conditions. Index Terms— MANET, energy consumption, AODV, DSR.
—————————— ——————————
1 INTRODUCTION
W
ith the advancement of wireless communication technologies, small-size and high-performance computing and communication devices like commercial laptops and personal digital assistants are being used increasingly in daily life. A mobile ad-hoc network (MANET) is composed of a group of mobile, wireless nodes which cooperate in forwarding packets in a multihop fashion without any centralized administration [1]. Significant examples of MANETs include establishing survivable dynamic communications for emergency /rescue operations, disaster relief efforts, and military networks. The network topology is highly dynamic in MANETs due to the movement of the nodes (mobility) which leads to frequent path breaks. This mobility affects the energy consumption of the nodes due to the high overheads that occurs in route discovery and route maintenance phase in MANETs as compared to their wired counterparts. As each mobile node in a MANET performs the routing function for establishing communication among different nodes the “death” of even a few of the nodes due to energy exhaustion might cause disruption of service in the entire network. MANETs can also be understood as low mobility and high mobility networks. In low mobility networks, the movement of nodes is less frequent and slow like com-
munication in a conference room, between a set of robots etc, whereas for high mobility networks the speed of the nodes is very high as in urban and road cars. The energy consumption of the nodes varies as the mobility changes. A number of power aware metrics are proposed and compared [8]-[10]. This paper attempts to study the effect of mobility and pause time on energy consumption of nodes in MANETs. The analysis of energy consumption for two popular on-demand routing protocols; Ad-hoc On Demand Distance Vector (AODV) and Dynamic Source Routing (DSR) is done. The rest of the paper is organized as follows. The brief description of AODV and DSR is provided in Section 2. A detailed description of the simulation environment follows in Section 3 which contains details of the network topology, the mobility models and the traffic model used in the simulation. Section 4 gives a brief of the various simulation scenarios used. The simulation results are provided in Section 5. Section 6 gives the conclusions.
2 MANET ROUTING PROTOCOL
The existing routing protocol in MANETs can be classified as table driven (also called proactive) and ondemand (also called reactive) schemes. In table-driven scheme, each node maintains the network topology in———————————————— formation in the form of routing tables by periodically exchanging routing information. On the other hand, onPinki Nayak is with Amity School of Engineering and Technology, New Delhi, INDIA. Currently, she is also a research student at Banasthali Uni- demand schemes obtain the necessary routes when reversity, Banasthali. quired by a source node for transmitting data. For perProf. Rekha Agarwal is the Director of Amity School of Engineering and formance comparison, two most popular on-demand proTechnology, New Delhi, INDIA. Dr. Seema Verma is working as Associate Professor at Banasthali Universi- tocols have been considered namely, DSR and AODV. ty, Banasthali, INDIA.
© 2011 JCSE www.journalcse.co.uk
26
2.1 Dynamic Source Routing (DSR) [4] The key distinguishing feature of DSR is the use of source routing i.e. the sender knows the complete hop-by-hop route to the destination [6]. It allows nodes in the network to dynamically discover a route across multiple network hops to any destination. In this protocol, the mobile nodes maintain route caches for the known routes. The route cache is updated when any new route is known for a particular entry in the route cache. Routing in DSR is done using two phases: route discovery and route maintenance. When a source node wants to send a packet to a destination, it first checks its own route cache to determine whether it already knows about any route to the destination or not. If already there is an entry for that destination, the source uses that route to transmit the packet. If not, it initiates a route request broadcast. This is known as route discovery phase. This request includes the destination address, source address, and a unique identification number. Each intermediate node checks whether it knows about the destination or not. If the intermediate node does not know about the destination, it again forwards the packet and eventually this reaches the destination. A node processes the route request packet only if it has not previously processed the packet and its address is not present in the route record of the packet. A route reply is generated by the destination or by any of the intermediate nodes when it knows about how to reach the destination. 2.2 Ad Hoc On Demand Distance Vector Routing (AODV) AODV [5] is a reactive routing protocol and it minimizes the number of broadcasts by creating routes based on demand. When any source node wants to send a packet to a destination, it broadcasts a route request (RREQ) packet [7]. The neighboring nodes in turn broadcast the packet to their neighbors and the process continues until the packet reaches the destination. During the process of forwarding the route request, intermediate nodes record the address of the neighbor from which the first copy of the broadcast packet is received. This record is stored in their route tables, which helps for establishing a reverse path. If additional copies of the same RREQ are later received, these packets are discarded. The reply is then sent using the reverse path. Unlike DSR, in AODV the nodes keep information of the next hop only. To maintain the freshness of the route AODV uses sequence numbers. For route maintenance, when a source node moves, it can re-initiate a route discovery process. If any intermediate node moves within a particular route, the neighbor of the drifted node can detect the link failure and sends a link failure notification to its upstream neighbor. This process continues until the failure notification reaches the source node. Based on the received information, the source might decide to re-initiate the route discovery phase.
3 SIMULATION SETUP The simulations are carried out using the event driven simulation tool Network Simulators-2 [2] (ver. 2.34) and
the wireless extensions provided by CMU [3]. Tool Command Language (TCL) is used to create the topology structure, to configure source nodes, intermediate nodes and destination nodes and to create the statistical data trace file, etc. NS - 2 also provides an inbuilt energy model that uses the traditional method of keeping track of the residual energy of a node. Once the trace file is generated, AWK scripts were written to extract the information from the trace file. Based on the output of these AWK scripts graphs were plotted for average residual energy v/s time for the network. The simulation setup consists of a test bed of 20 nodes randomly and uniformly distributed in an area of 500 X 500 m2. The range of each node is assumed to be 250 m. The mobility model used was Random Waypoint model with different speed varying between 0-5 m/sec (low mobility), 0-10 m/sec (medium mobility) and 0-20 m/sec (high mobility) and at different pause time. Constant Bit Rate (CBR) traffic source was used with packet size of 512 bytes. Different source-destination pair (5, 10 connections) were used to establish the routes. All the simulations were run for a period of 200 sec. The initial energy of each node was set as 100 Joules with transmission and reception power of 5 W and 1W respectively. Table 1 gives the simulation parameters used. TABLE 1 SIMULATION PARAMETERS
Parameter
Value
Number of Nodes Grid Area Transmission Range Pause Time Speed Traffic Model Data Packet Size Data Packet Interval Simulation Time Initial Energy of Node Transmitted Power Received Power Idle State Power Sleep State Power Transition from sleep to active state
20 500m x 500m 250 m 0 sec, 20 sec 5 m/s, 10m/s, 20m/s CBR 512 bytes 4 packets/sec 200 sec 100 J 5W 1W 0.0005 W 0.0002 W .03W
4 SIMULATION SCENARIO Simulations have been conducted by varying the speed of the nodes and for different source – destination pair. The metric used to evaluate the performance of DSR and AODV protocol is the average residual energy of the network at different times. Total energy consumption in a network is directly proportional to various factors like size of network, arrival rate of route requests, packet arrival rate, packet size (header size and payload size),
27
packet collision and retransmissions. Total residual energy is indirectly proportional to the energy consumption. The life of the network depends on the node expiration which in turn depends upon energy consumption in a node. If the battery capacity of node reaches zero then node will die. In NS-2, the energy of the node is displayed using different colors’. 100% > energy of node > 50 % 50% > energy of node > 20% 20% > energy of node
Green Yellow Red
5 RESULTS In this section, the comparison of energy consumption is made for DSR and AODV routing protocols.
5.1 Performance of DSR and AODV at different mobility scenarios Fig. 1 and 2 give the average residual energy of the network with time for DSR and AODV protocols respectively at different speeds with pause time of 0 sec. As seen, DSR performs well in low mobility scenario. This is because DSR maintains multiple paths foe same destination in its route cache. Also, at low mobility the number of link failures is less which means infrequent route discovery and few overheads. So, average energy consumption is less.
Fig. 2. Average residual energy vs. Time for AODV at different speed
5.2 Performance Evaluation for different source destination Pairs The performance of DSR and AODV is compared for different number of sources at varying speeds. As seen from
Fig. 3. Average Residual Energy with Time for different connections at Low Mobility
Fig. 1. Average Residual Energy vs. Time for DSR at different speed.
As, the mobility of the nodes increases, DSR’s performance greatly decreases with time. The deterioration is significant at high mobility due to very frequent link breaks. Same is the case with AODV, as the number of link breaks is less at low mobility; Route Discovery is initiated only if path fails. The node in AODV has only one next hop address corresponding to the destination. It can be seen from Fig. 2 that the energy consumption in AODV is not as drastic as it was for DSR with change in mobility’s.
Fig. 3, at low speed of 5 m/sec, DSR outperforms AODV for different source-destination pairs of 5 connections and 10 connections. The average residual energy of DSR and AODV remains almost same initially till 70 sec of simulation time but after that (70 sec -180 sec) DSR shows much better performance as compared to AODV. This is because of the promiscuous listening and multiple routes that are stored in route cache in DSR [6]. Also, few link breaks at low mobility indicates route in route cache to remain fresh, thus, leading to less overheads due to route discovery.
28
At medium (Fig. 4) and high mobility (Fig. 5), not much difference between DSR and AODV is seen when the number of sources is 5. But as the source-destination pair increases to 10, DSR shows better performance over AODV.
Fig. 6. Average Residual Energy with Time for different Pause Time at Low Mobility (5 m/sec)
Fig. 4. Average Residual Energy with Time for different connections at Medium Mobility.
Fig. 7. Average Residual Energy with Time for different Pause Time at Medium Mobility (10 m/sec)
Fig. 5. Average Residual Energy with Time for different connections at High Mobility.
5.3 Performance Evaluation at different Pause Time The average residual energy is then plotted with time at different pause time and for varying speed (Fig. 6, 7, 8). Pause time of 0 sec and 20 sec has been considered. Residual Energy of the network increases with increase in pause time. Increase in pause time indicates the network to be less mobile. Fig. 8. Average Residual Energy with Time for different Pause Time at High Mobility (20 m/sec)
29
6 CONCLUSION This study compared the energy consumption of two routing protocols: DSR and AODV in terms of average residual energy for different mobility scenarios. Mobility greatly effects the energy consumption of the nodes in MANETs which is because of the high overheads incurred in Route Discovery (Route Request RREQ and route reply RREP) and Route Maintenance phase. Higher the mobility, greater the routing overheads and increased energy consumption. DSR performs better at low and medium mobility scenarios than AODV. It is possible because at low and medium mobility scenarios, there is delay in Route Discovery as the route caches of the nodes are up-to date in DSR. There is further decrease in Route Discovery overhead due to packet salvaging. In contrast, AODV initiates Route Discovery every time there is a link failure as it maintains only one next hop address for the destination. As there is increase in mobility, the performance of the two protocols reduces significantly. AODV shows almost same performance as DSR. In high mobility scenario, frequent link failure occur, which neccesitates frequent Route Discovery in DSR. In AODV, stale routes are prevented by destination sequence number. So, increased routing overheads in both protocols leads to lower energy levels of nodes. When the nodes are highly mobile, AODV may be preferred. It may also be used where the path between source and destination is long (no source routing, so less overhead).
REFERENCES [1]
S. Corson, J. Maker., “Mobile Ad hoc Networking (MANET),” Internet Draft, Oct. 1998. [2] NS-2 Manual www.isi.edu/nsnam/ns/doc/index.html [3] CMU Monarch Extensions to ns. http://www.monarch.cs.cmu.edu/. [4] J. Broch, D. Johnson, D. Maltz,“ The Dynamic Source Routing Protocol for Mobile Ad Hoc Networks,” IETF internet draft, 1999. [5] C.E.Perkins, S.R.Das, E.M.Royer, “Ad hoc On Demand Distance Vector Routing,” IETF draft, March 2001. [6] C. Perkins, E. Royer, S. Das, M. Marina, “Performance Comparison of Two On-Demand Routing Protocols for Ad Hoc Networks,” IEEE Personal Communication, vol 8, issue 1, Feb 2001. [7] Samba Sesay, Zongkai Yang, Biao Qi, and Jianhua He, “Simulation Comparison of Four Wireless Ad-hoc Routing Protocols,” Information Technology Journal 3, 2004. [8] Allard G, Minet P, Nguyen D. Q., Shresta N, “Evaluation of the Energy Consumption in MANET,” ADHOCNOW, Ottawa, Canada, Aug 2006. [9] Andrew McCabe, Mikael Fredin, Alan Cullen, Leif Axelsson, "A power consumption study of DSR and OLSR," Military Communications Conference, 2005. MILCOM 2005. IEEE [10] Marco Fotino, Antonio Gozzi, Floriano De Rango, Salvatore Marano, Juan-Carlos Cano, Carlos Calafate, Pietro Manzoni, “Evaluating Energy-aware behaviour of proactive and reactive routing protocols for mobile ad hoc networks,” 10th
International Symposium on Performance Evaluation of Computer and Telecommunication Systems (SPECTS'07) Pinki Nayak received her M.Tech (Information Technology) from Guru Gobind Singh Indraprastha University, New Delhi, India. At present, she is working as Assistant Professor in Department of ECE at Amity School of Engg. & Tech., New Delhi and also working towards the Ph.D. degree at Banasthali University, Banasthali, India. Her research interest lies in the area of Mobile Ad-hoc networks, Wireless Communucation. Dr. Rekha Agarwal is working as Professor at, Amity School of Engg & Tech., New Delhi. She graduated in Electronics and Communication Engg. from Jiwaji University, Gwalior. Later, she obtained her Masters degree from MREC, Jaipur and then Ph.D. degree from Guru Gobind Singh Indraprastha University, Delhi. She has been actively involved in teaching and research and has published many research papers. Her teaching and research interest include multiuser detection, coding techniques, Ad-hoc Networks. Dr. Seema Verma obtained her Master and Ph.D degree in Electronics from Banasthali University in 1999 & 2003. She is working as Associate Professor [Electronics] at Banasthali University. She is an active research supervisor and has presented many papers in various international conferences. She has published many research papers in various journals of repute. Her research areas are Coding theory, TURBO Codes, Wireless sensor networks, Aircraft Ad-hoc networks, Network Security & VLSI Design.