Performance Comparison of AODV and DTN routing Protocols in a sparse network

Performance Comparison of AODV and DTN routing Protocols in a sparse network 1 Maria Benamar, 1Nabil Benamar, 2Jonathan Ledy, 2Benoit HILT, 3Jean Marie Bonnin, 1Driss El Ouadghiri, 1 MACS Laboratory,Faculty of sciences, University Moulay Ismaïl, Meknes Morocco MIPS Laboratory , 2University of Haute Alsace, Colmar France 3 Telecom Bretagne, Rennes, France 2 E-mail: 1mariabenamar@gmail.com, 1benamar73@gmail.com, ledy.jonathan@gmail.com, 2benoit.hilt@gmail.com, 3 jm.bonnin@telecom-bretagne.eu, 1dmelouad@gmail.com Abstract— In this paper we compare routing protocols of mobile vehicular ad-hoc networks (VANETs) and vehicular delay tolerant networks (VDTNs). We choose AODV as the reference protocol for evaluation because it is the most used in VANETs. Simulations of VDTN and AODV protocols are performed on ONE and NS-2 simulators respectively, with the identical scenario. Based on the simulation results obtained, performances of AODV are analyzed and compared in terms of delivery probability, latency and hops counts with three popular VDTN protocols. Keywords— Vehicular Delay-Tolerant Networks (VDTN), AODV, Routing, Performance Comparison, ONE, NS-2. I. INTRODUCTION MANETs consist of mobile nodes with no existing pre-established infrastructure. To establish connectivity based on multi hop routes, they connect themselves in a decentralized and self-organized manner. If mobile nodes are vehicles, then this type of network is called VANET (Vehicular Adhoc Network). One important property that distinguishes MANET from VANET is that nodes move with higher speed, on constrained ways (roads) and nodes quantity can vary from sparse to jamming. From a technical point of view, VANETs consist of vehicles On Board Units (OBU) and Road Side Units (RSU) equipped with radios devices that run dedicated Wireless Access in Vehicular Environments (WAVE) IEEE 802.11p protocol [1]. When the density of the cars is very low information must be stored for variable time before forwarding or delivering. We can therefore speak of Vehicular Delay Tolerant Network (VDTN) as depicted in figure 1. Fiat, BMW, Renault and some other companies have united to develop a car-to-car communication consortium, dedicated precisely to define common architecture for Vehicle to Vehicle (V2V) and Vehicle to Infrastructure (V2I) communication for sharing safety related information and access location based services [2]. The wealth of information that could be obtained from vehicular networks is quite INTRODUCTION enormous, ranging from location and speed of emergency alerts and request for roadside assistance. In particular, many envisioned safety related applications require that the vehicles continuously broadcast their current position and speed in so called beaconing messages. This messaging increases the awareness of vehicles about their neighbors’ whereabouts and warns drivers off dangerous situations. But the very richness of information also threatens to cause deployment to come to a grinding halt if there is adverse consumer reaction to technology. In this paper we start the discussion with the introduction of vehicular delay tolerant network. Then we describe AODV routing technique and give simulation results of AODV and VDTN routing protocols on NS2 and ONE respectively. Finally we end with conclusion and few useful references. Figure 1 : VDTN scenario II. DTN ROUTING PROTOCOLS In DTN networks the end-to-end path does not exist hence, existing routing protocol fails [3]. There are various replications and forwarding based protocol proposed to work in such environment. In forwarding only single copy of message is available and in case of replication multiple copies of message are spread across the network. Examples protocol for single copy scheme are epidemic, spray & wait , prophet, direct delivery and MaxProp. Prophet (Probabilistic Routing Protocol using History of Encounters and Transitivity) uses a probabilistic metric: delivery predictability y, that attempts to estimate, based on node encounter history, which node has the higher probability of successful delivery of a message to the final destination [4]. When two nodes are in communication range, a new message copy is transferred only if the other node has a better probability of delivering it to the destination. Direct Delivery are single copy DTN routing protocols where only one copy of each message exists in the network [5]. In Direct Delivery, the message is kept in the source and delivered only to the final destination, if the nodes meet. In First Contact, the message is forwarded to the first node encountered and deleted. The message is forwarded until it reaches the intended destination[6]. Epidemic Routing protocol [7] is flooding-based protocol, where nodes continuously replicate and transmit messages to newly discovered contacts that do not already possess a copy of the message. Consequently, epidemic routing protocol minimizes the delivery delay and maximizes the delivery ratio as messages may reach the destination on multiple paths, but spoils storage and bandwidth in comparison with other protocols[8]. Spray and Wait [9] is an n - copy routing protocol with two phases: (1) spray phase, where a message created by the source node is initially spread by the source to encountered nodes until the n copies are exhausted ; (2) wait phase, where every node containing a copy of the message performs a direct delivery to the destination. There are two variants of the protocol: normal mode, where a node gives one copy of the message to each node it discovers that does not have the message; and binary mode, where half of the n copies are given in each encounter. MaxProp protocol attempts to transfer all messages not held by the other node, when it is in communication range[10]. The protocol uses acknowledgments to clear the remaining copies of a message in the network when the destination node receives it. When nodes discover each other, MaxProp exchanges messages in a specific priority order, taking into account message hop counts and the delivery likelihood to a destination based on previous encounters. New packets are assigned higher priority, and the protocol attempts to avoid reception of duplicate packets. III. AODV ROUTING PROTOCOL AODV is an on-demand protocol, which initiate route request only when needed. When a source node needs a route to certain destination, it broadcasts a route request packet (RREQ) to its neighbors. Each receiving neighbor checks its routing table to see if it has a route to the destination. If it doesn’t have a route to this destination, it will re-broadcast the RREQ packet and let it propagate to other neighbors. If the receiving node is the destination or has the route to the destination, a route reply (RREP) packet will be sent back to the source node. Routing entries for the destination node are created in each intermediate node on the way RREP packet propagates back. A hello message is a local advertisement for the continued presence of the node. Neighbors that are using routes through the broadcasting node will continue to mark the routes as valid. If hello messages from a particular node stop coming, the neighbor can assume that the node has moved away. When that happens, the neighbor will mark the link to the node as broken and may trigger a notification to some of its neighbors telling that the link is broken. In AODV, each router maintains route table entries with the destination IP address, destination sequence number, hop count, next hop ID and lifetime. Data traffic is then routed according to the information provided by these entries [4]. IV. localization of all stationary nodes before running the simulation. PERFORMANCE COMPARISON The network scenario is based on a part of the city of Helsinki (Finland), the total number of all nodes in all simulations has been kept fix and equal to 100. Ten stationary destination nodes and ten stationary source nodes are placed at the map positions presented in this figure. Fig.2 shows the Figure 2 : The position of all nodes in the simulated network The source nodes represent sensors used in urban area to collect different types of information and measures, whereas destinations nodes are data collection gateways. The different parameters of simulations are summarized in the table 1. Note that in this table the TTL defines the living time (in seconds) of a DTN bundle in the network. This TTL is not the classical IP TTL that counts the number of hops. Table 1: summarization of the different parameter used in the scenario Simulation time 43200 seconds Buffer size 5M Movement Model Map Based Movement TTL [50;100;150;200;300;400;500] Routing protocol Interface type [MaxProp; Epidemic; Prophet; Spray And Wait;AODV] Simple Broadcast Interface Number of nodes 80 mobile nodes Velocity of mobile nodes Size of bundles 2.7, 13.9 100KB, 1MB Event used External Event Queue Map Helsinki Channel Propagation model FreeSpace Data Transmission rate 2 Mbps a) b) c) Figure 3 : bundle delivery probability (a), Latency (b) and hopcount_avg (c), as Map Based Movement for Epidemic, Spray and Wait, PRoPHET, MaxProp and AODV routing protocols Maxprop shows the best delivery probability results when the number of nodes is high, and the worst latency values for sparse and dense traffic, while Prophet Protocol shows best values in term of latency. The low latency of AODV can be explained by the fact that road is created by the first message received RREP after a RREQ. So whoever follows the path with the lowest latency, also we see that the delivery probability of AODV becomes stable from the TTL=150min which proves that the protocol is not suitable for a sparse environment. In addition, AODV that is a routing protocol in which each node maintains a routing table, one entry per destination, which records the next hop to the destination and its hop count, it’s uses a sequence number to ensure freshness of routes. AODV discovers a route through network wide broadcasting. It does not record the nodes it has passed, but only counts the number of hops, so Fig 3-e shows the performance of the studied routing protocols regarding hop count metric. AODV protocol shows the maximum value regarding to other protocols of VDTN, also this value becomes a constant value in TTL=150min. V. CONCLUSION This paper evaluated the performances and AODV and different VDTN routing protocols using a scenario with 80 nodes in a large area (Helsinki map), tow network simulators was used in this study, which is NS2 and ONE, The ONE simulator is a dedicated tool for DTN and opportunistic networks. The results observed and simulations shows that AODV routing protocol is not suitable for a opportunistic environment because of his delivery probability that is constant despite the value of the TTL increases. Next we will extend the work by considering other MANETs routing protocols like GPSR. REFERENCES [1] A. Lindgren, A. Doria, E. Davies, and S. Grasic, “Probabilistic Routing Protocol for Intermittently Connected Networks draft-irtf-dtnrg-prophet-09,” 2011. [2] T. Spyropoulos, K. Psounis, and C. S. Raghavendra, “Single-copy routing in intermittently connected mobile networks,” 2004 First Annual IEEE Communications Society Conference on Sensor and Ad Hoc Communications and Networks, 2004. IEEE SECON 2004., pp. 235–244. [3] B. Parno and A. Perrig, “Challenges in Securing Vehicular Networks,” no. 4. [4] A. Lindgren, A. Doria, E. Davies, and S. Grasic, “Probabilistic Routing Protocol for Intermittently Connected Networks“ draft-irtfdtnrg-prophet-09,” 2011. [5] T. Spyropoulos, K. Psounis, and C. S. Raghavendra, “Single-copy routing in intermittently connected mobile networks”, 2004 First Annual IEEE Communications Society Conference on Sensor and Ad Hoc Communications and Networks, 2004. IEEE SECON 2004., pp. 235–244. [6] S. Jain, K. Fall, and R. Patra, “Routing in a delay tolerant network”, Proceedings of the 2004 conference on Applications, technologies, architectures, and protocols for computer communications - SIGCOMM ’04, p. 145, 2004. [7] A. Vahdat and D. Becker, “Epidemic routing for partially connected ad hoc networks”, Technical Report,, 2000. [8] N. Benamar, K. D. Singh, M. Benamar, D. El, and J. Marie, “Routing protocols in vehicular delay-tolerant networks : a comprehensive survey”, [9] T. Spyropoulos, K. Psounis, and C. S. Raghavendra, “Spray and wait: an efficient routing scheme for intermittently connected mobile networks”, Proceedings of the 2005 …, pp. 252–259, 2005. [10] J. Burgess and B. Gallagher, “Maxprop: Routing for vehicle-based disruption-tolerant networks”, Proc. ieee …, vol. 6, no. c, pp. 1–11, 2006. [11] J.Ledy, H.Boeglen, AM.Poussard, B.HILT, R.Vauzelle, “A semi-deterministic channel model for VANETs simulations”, International Journal On Vehicular Technology, Volume 2012, Article ID 492105