A Graph-based Approach to Compute Multiple Paths in Mobile Ad Hoc Networks

Similar documents
AODV-PA: AODV with Path Accumulation

Gateway Discovery Approaches Implementation and Performance Analysis in the Integrated Mobile Ad Hoc Network (MANET)-Internet Scenario

Performance of Ad-Hoc Network Routing Protocols in Different Network Sizes

QoS Routing By Ad-Hoc on Demand Vector Routing Protocol for MANET

Performance Comparison of Ad Hoc Routing Protocols over IEEE DCF and TDMA MAC Layer Protocols

An Efficient Routing Approach and Improvement Of AODV Protocol In Mobile Ad-Hoc Networks

A Reliable Route Selection Algorithm Using Global Positioning Systems in Mobile Ad-hoc Networks

A New Energy-Aware Routing Protocol for. Improving Path Stability in Ad-hoc Networks

Performance Comparison of AODV, DSR, DSDV and OLSR MANET Routing Protocols

A Performance Comparison of MDSDV with AODV and DSDV Routing Protocols

Analysis QoS Parameters for Mobile Ad-Hoc Network Routing Protocols: Under Group Mobility Model

A COMPARISON OF IMPROVED AODV ROUTING PROTOCOL BASED ON IEEE AND IEEE

1 Multipath Node-Disjoint Routing with Backup List Based on the AODV Protocol

A Highly Effective and Efficient Route Discovery & Maintenance in DSR

ROUTE STABILITY MODEL FOR DSR IN WIRELESS ADHOC NETWORKS

3. Evaluation of Selected Tree and Mesh based Routing Protocols

IMPACT OF MOBILITY SPEED ON PROACTIVE AND REACTIVE ROUTING PROTOCOLS IN MOBILE ADHOC NETWORKS

Performance Evaluation and Comparison of On-Demand Routing Protocols for Ad Hoc Networks: DSR, AODV, AOMDV, TORA

Analysis of Network Traffic in Ad-Hoc Networks based on DSDV Protocol

Anil Saini Ph.D. Research Scholar Department of Comp. Sci. & Applns, India. Keywords AODV, CBR, DSDV, DSR, MANETs, PDF, Pause Time, Speed, Throughput.

Impact of Hello Interval on Performance of AODV Protocol

Efficient On-Demand Routing for Mobile Ad-Hoc Wireless Access Networks

Simulation & Performance Analysis of Mobile Ad-Hoc Network Routing Protocol

Dynamic Source Routing in ad hoc wireless networks

AODV Multipath Extension using Source Route Lists with Optimized Route Establishment

Comparative Study for MCDS and DSR Which Are Used For Packet Forwarding In Ad Hoc Network

ANewRoutingProtocolinAdHocNetworks with Unidirectional Links

Performance Evaluation of Routing Protocols for MAC Layer Models

Dynamic Route Switching Protocol in Mobile Ad Hoc Networks

Routing Protocols in MANETs

DYNAMIC ROUTES THROUGH VIRTUAL PATHS ROUTING FOR AD HOC NETWORKS

PERFORMANCE EVALUATION OF DSR USING A NOVEL APPROACH

Performance of Route Caching Strategies in Dynamic Source Routing

TCP Performance over Multipath Routing in Mobile Ad Hoc Networks

Performance Evaluation of Various Routing Protocols in MANET

A Performance Comparison of Multi-Hop Wireless Ad Hoc Network Routing Protocols

Content. 1. Introduction. 2. The Ad-hoc On-Demand Distance Vector Algorithm. 3. Simulation and Results. 4. Future Work. 5.

Performance Analysis and Enhancement of Routing Protocol in Manet

Behaviour of Routing Protocols of Mobile Adhoc Netwok with Increasing Number of Groups using Group Mobility Model

Performance Analysis of Three Routing Protocols for Varying MANET Size

Secure Enhanced Authenticated Routing Protocol for Mobile Ad Hoc Networks

Overhead Analysis of Query Localization Optimization and Routing

Speed Performance of Intelligent Ant Sense Routing Protocol for Mobile Ad-Hoc Personal Area Network

Relative Performance Analysis of Reactive (on-demand-driven) Routing Protocols

Performance Evaluation of MANET through NS2 Simulation

A Performance Comparison of Multi-Hop Wireless Ad Hoc Network Routing Protocols. Broch et al Presented by Brian Card

A SURVEY OF ROUTING PROTOCOLS IN MOBILE AD HOC NETWORKS

INVESTIGATING THE SCALABILITY OF THE FISH-EYE STATE ROUTING PROTOCOL FOR AD HOC NETWORKS

PERFORMANCE ANALYSIS OF AODV ROUTING PROTOCOL IN MANETS

Efficient On-Demand Routing for Mobile Ad-Hoc Wireless Access Networks

Research Paper GNANAMANOHARAN ET AL., INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING TECHNOLOGY E-ISSN

Maharishi Markandeshwar University

Mobility and Density Aware AODV Protocol Extension for Mobile Adhoc Networks-MADA-AODV

Control Traffic Analysis of On-Demand Routing Protocol. in Mobile Ad-hoc Networks

A Study of Bellman-Ford, DSR and WRP Routing Protocols with Respect to Performance Parameters for Different Number of Nodes

Throughput Analysis of Many to One Multihop Wireless Mesh Ad hoc Network

Routing in Ad Hoc Wireless Networks PROF. MICHAEL TSAI / DR. KATE LIN 2014/05/14

Exploring the Behavior of Mobile Ad Hoc Network Routing Protocols with Reference to Speed and Terrain Range

ENERGY EFFICIENT MULTIPATH ROUTING FOR MOBILE AD HOC NETWORKS

CA-AODV: Congestion Adaptive AODV Routing Protocol for Streaming Video in Mobile Ad Hoc Networks

An Extended AODV Protocol for Multipath Routing in MANETs

Enhancing the Performance of Mobile Ad Hoc Networks with the Aid of Internet Gateways 1

A COMPARISON OF REACTIVE ROUTING PROTOCOLS DSR, AODV AND TORA IN MANET

Measure of Impact of Node Misbehavior in Ad Hoc Routing: A Comparative Approach

Impact of Node Velocity and Density on Probabilistic Flooding and its Effectiveness in MANET

Power aware Multi-path Routing Protocol for MANETS

Performance Comparison of AODV and AOMDV Routing Protocols in Mobile Ad Hoc Networks

Performance Evolution of Proactive and Reactive Routing Protocols in Mobile Ad Hoc Networks

Performance Comparison of Two On-demand Routing Protocols for Ad-hoc Networks based on Random Way Point Mobility Model

Performance analysis of aodv, dsdv and aomdv using wimax in NS-2

MSDM: Maximally Spatial Disjoint Multipath Routing Protocol for MANET

Volume 2 No. 1 ISSN Journal of Emerging Trends in Computing and Information Sciences CIS Journal. All rights reserved.

Evaluation of Routing Protocols for Mobile Ad hoc Networks

An Adaptive Routing Strategy Based on Dynamic Cache in Mobile Ad Hoc Networks*

Dynamic AODV Backup Routing in Dense Mobile Ad-Hoc Networks *

2013, IJARCSSE All Rights Reserved Page 85

Performance Analysis of MANET Routing Protocols OLSR and AODV

Analysis of TCP and UDP Traffic in MANETs. Thomas D. Dyer Rajendra V. Boppana CS Department UT San Antonio

Zone-based Proactive Source Routing Protocol for Ad-hoc Networks

Probabilistic Mechanism to Avoid Broadcast Storm Problem in MANETS

Performance Comparison of MANETs Routing Protocols for Dense and Sparse Topology

Mobile Ad Hoc Network Routing Protocols: A Comparative Study

Multicasting in Ad-Hoc Networks: Comparing MAODV and ODMRP

PERFORMANCE ANALYSIS OF RANDOMIZED REVERSE AD HOC ON DEMAND DISTANCE VECTOR ROUTING PROTOCOL IN MANET

Performance evaluation of reactive and proactive routing protocol in IEEE ad hoc network

Study of Route Reconstruction Mechanism in DSDV Based Routing Protocols

Performance Analysis of Routing Protocols for Mobile Ad-hoc Networks

Figure 1: Ad-Hoc routing protocols.

Performance evaluation of reactive routing protocols for IEEE

[Kamboj* et al., 5(9): September, 2016] ISSN: IC Value: 3.00 Impact Factor: 4.116

AWERProcedia Information Technology & Computer Science

Estimate the Routing Protocols for Internet of Things

A Novel Interference Aware Optimized Link State Routing Protocol for Power Heterogeneous MANETs

Performance Evaluation and Comparison of AODV and AOMDV

Analysis of Routing Protocols in MANETs

A Comparative Analysis of Energy Preservation Performance Metric for ERAODV, RAODV, AODV and DSDV Routing Protocols in MANET

Computation of Multiple Node Disjoint Paths

The CMU Monarch Project s Wireless and Mobility Extensions to ns

Dynamic Load-Aware Routing in Ad hoc Networks

A Performance Comparison of Multicast Routing Protocols In Ad hoc Networks

Transcription:

A Graph-based Approach to Compute Multiple Paths in Mobile Ad Hoc Networks Gunyoung Koh, Duyoung Oh 1 and Heekyoung Woo 2 1 School of Electrical Engineering and Computer Science Seoul National University, Seoul, Republic of Korea {kgy, ody}@popeye.snu.ac.kr 2 Division of Information Technology Engineering, Soonchunhyang University, Asan, Republic of Korea {woohk}@sch.ac.kr Abstract. Multipath on-demand routing protocols for mobile ad hoc networks try to reduce control overhead and end-to-end delay by computing multiple paths with a single route discovery process. We propose Graph-based Multipath Routing (GMR), a novel multipath routing protocol that generate the network topology graph to compute all link disjoint paths in the network. The destination node computes link disjoint paths using the local graph search algorithm. We present our simulation results compared with DSR and Multipath DSR 1 Introduction High mobility, limited battery and valuable bandwidth resource of ad hoc networks make the classical routing protocols impractical to be used directly for ad hoc networks. Therefore a number of routing algorithms for ad hoc networks have been proposed. They may be categorized into two distinct groups - proactive protocols including Destination Sequenced Distance Vector (DSDV) [1] and on-demand protocols including Dynamic Source Routing (DSR) [2], Temporally Ordered Routing Algorithm (TORA) [3] and Ad hoc On-demand Distance Vector routing (AODV) [4]. Although some simulation studies [5] have shown that on-demand protocols incur lower routing overheads than proactive protocols, they still have some problems. On-demand routing protocols discover routes via a flooding technique. That is, a message from a source node is delivered to all other nodes. It takes a substantial amount of network bandwidth which is a premium resource in wireless networks. Although there are many research efforts [6] to find efficient flooding methods, reducing the number of flooding initiations is still an important research issue. Multipath routing schemes try to reduce these problems by finding multiple paths through a single route discovery process. The source node chooses a This work was supported by Samsung Electronics and the University IT Research Supporting Program under the Ministry of Information and Communication of Korea

path among multiple paths and starts transferring data packets. If a link on the path broke down, instead of initiating an additional route discovery process, the source node just chooses another path and continues to transfer. New route discovery process (i.e. flooding) is initiated only after all paths have failed. So the number of flooding initiations can be reduced subsequently. Several multipath routing schemes, including Multipath-DSR [7], Split Multipath Routing (SMR) [8], and Ad hoc On-demand Multipath Distance Vector (AOMDV) [9], are recently proposed. It is important that many multipath routing algorithms seek for link disjoint paths among all possible paths. Link disjoint paths guarantee the independence of failures of each path, thus increasing the path availability. But all these link disjoint paths are not required. M. K. Marina and S. R. Das argued that additional routes beyond a few provide only marginal benefit [9]. They selected 3 for the threshold. But we found that previous multipath routing schemes are unable to find existing link disjoint paths in many cases even below 3. We will show some examples and simulation results. In this work, therefore, our goal is to develop a multipath routing scheme that computes sufficient number of link disjoint paths. The reason for reducing the number of flooding initiations is to reduce routing control overhead, so a scheme using more control packets to compute multiple paths is not desirable. So our goal also includes holding down the number of control packets compared with single path routing protocols. To achieve this goal, our scheme, Graph-based Multipath Routing (GMR) uses graph information which represents the abstract network topology. This graph information is constructed by intermediate nodes in a distributed manner and enables the destination to compute link disjoint paths using local graph search algorithms. The rest of this paper is organized as follows. In section 2, previous works related to multipath routing is observed. Proposed GMR protocol is presented in section 3, while simulation results are carried out in section 4. We conclude in section 5. 2 Related Works In this section, we briefly describe the key features of previous multipath routing protocols and also show some examples in which the routing protocols cannot find existing link disjoint paths. Each multipath on-demand protocol is developed on the base of its original single path on-demand protocol, such as DSR and AODV. Multipath-DSR (M-DSR) [7] is a simple multipath extension of the popular DSR. Instead of replying only to the first received RREQ as DSR, the destination node sends an additional RREP for a RREQ which carries a link disjoint route compared with the routes already replied. However, M-DSR can t compute link disjoint paths in many cases because the intermediate nodes drop every duplicate RREQ that may comprise another link disjoint path. Figure 1 illustrates an example that M-DSR doesn t compute link disjoint paths.

Fig. 1. An example that M-DSR doesn t compute link disjoint paths. Node 3 discards one of the RREQs from node 1 and node 2 because it is a duplicate copy. Split Multipath Routing (SMR) [8] is another multipath variant of DSR. SMR introduces a different RREQ propagation mechanism in order to compute link disjoint paths that M-DSR cannot find. However, while SMR reduces the number of flooding initiations, SMR doesn t reduce routing overhead considerably. It is because each flooding in SMR requires much more control packets than that of DSR. AOMDV [9] is an AODV-based multipath routing protocol. The RREQ propagation mechanism of AOMDV is same as that of AODV except each RREQ carries an additional field called firsthop to indicate the first hop passed by the RREQ. The firsthop field is used to decide the link disjoint-ness of each RREQ. However, AOMDV also fails to catch many chances to find existing link disjoint paths. It is because the second RREQ is not re-broadcasted. 3 The Proposed Scheme We now describe our Graph-based Multipath Routing protocol based on DSR. First, we explain the basic operation of GMR. We then extend it by considering an optimization method and practical issues. The ideal goal of GMR is to compute all link disjoint paths. But too long paths are not practical because connection on such paths suffers longer end-toend delay. Therefore we focus on only for shortest or sufficiently short paths. When we mention all paths, it means all shortest (or sufficiently short) paths rather than all possible arbitrary hop paths. GMR uses source routing mechanism just like DSR. So a RREQ message contains path information from source node to the intermediate node. But instead of single node list of DSR, each RREQ includes graph information which represents abstract network topology. We call this graph as Reverse Path Graph (RPG). The route discovery process of GMR performed with similar manner to that of DSR. The main difference is that each intermediate node which receives a new RREQ message should wait for some predetermined time to gather more RREQs, if any. If the waiting intermediate node receives more than one RREQ, it merges graph information of those duplicate RREQs with its previous graph

information. After time-out, it re-broadcast one RREQ message which contains all information it gather until that moment. Waiting process enables gathering sufficient route information while it prevents generation more control packets than DSR. It is easy to show that if the time-out interval is sufficient, the re-broadcasted RPG of each node covers all RPGs of the previous hops, that is, all shortest link disjoint paths. If the propagations of RREQ occur simultaneously, the time-out interval needs not to be long. But to prevent collision, wireless MAC layer has back-off process. This makes jitters between RREQ messages. We have chosen 50 milliseconds for the interval, which is larger than MAC layer back-off interval by a factor of magnitude. Waiting some time from the first RREQ, the destination node gathers multiple RREQs. And after time-out, it computes link disjoint paths from the RPG using the local graph search algorithm (e.g. [10]) and it replies to the source through multiple RREPs through multiple paths. Finally, the source uses source routing for data packet delivery. Figure 2 shows the graph information generated during a route discovery process. Fig. 2. The basic operation of GMR. Node 4 receives upper edges from node 1 and lower edges from node 2 while it waits. And node 4 relays fully merged graph to node 6. The limitation on the size of RPG is another control parameter of GMR. If the size goes over one packet size, the number of control packet increases. So one packet size is the loosest bound of RPG size. If an intermediate node detects the merged RPG size exceeds the limitation, it performs following optimization. The optimization occurs by two steps. At first step, from the original RPG, it selects edges which form link disjoint routes form the source node and itself. New RPG is composed of those edges. This process reduces the size of RPG remarkably, but cannot assure that the size falls below the limitation. When the size of new RPG is still larger than the limitation, just one shortest route is computed to

replace RPG. It becomes identical to node list of DSR. This is the second step of optimization. Actually the size of RPG highly depends on implementation of graph information. In the simulation, we used raw edge list implementation, which is one of the worst implementation at the view of graph size. Two IP address compose one edge in RPG. So each edge takes 8 bytes, which look very high overhead. Other proper implementations may reduce the graph size significantly, but we want to show that even with poor implementation, the optimization doesn t affect much over the number of multipaths GMR finds. In the simulations, we limit the size of RPG with 64 edges. 4 Performance Evaluation To compare the performance of GMR with previous work in ad hoc routing, we simulate GMR with DSR and M-DSR. SMR is excluded because it uses more control packets to compute more paths. Also, AOMDV which bases on AODV is not compared with GMR which bases on DSR. We implement the simulator within the ns-2 [11], using the wireless extensions developed at Carnegie Mellon [12]. In our simulation model, 100 mobile hosts are placed randomly within a 1000 meter 1000 meter area and random waypoint model [5] is used as the mobility model. Traffic sources are CBR and the source-destination pairs are spread randomly over the network. There are 10 data sessions and the packet sending rate is 4 packets per second. Each simulation data shown are an average value of ten runs with different randomly generated mobility scenarios. 4.1 Simulation Results Figure 3 shows the number of paths computed by each protocol by a single route discovery process. Through simulations, we limited each protocol to compute at most four paths because more than four paths provide little benefit and make themselves stale routes which may disturb data packet delivery. As expected, DSR computes one path and GMR outperforms M-DSR. It is because GMR provides the destination with more topology information than M-DSR. Figure 3 also proves that M-DSR cannot find existing link disjoint paths even when the number of the paths is fewer than four. This is the basic motivation of GMR and also the basic reason how GMR can outperform M-DSR through the following simulation results. Figure 4 shows the routing overheads of each protocol. Normalized routing load is the ratio of the number of all control packets propagated throughout the whole network and the number of data packets successfully delivered to the destination nodes. Among three protocols, GMR shows the best performance and M-DSR follows. Both surpass DSR. Figure 4 shows the benefit of multipath routing protocols. Even though those three protocols generate almost same number of control packet in one route discovery process, the number of route

Fig. 3. The Number of Computed Paths Fig. 4. Normalized Routing Load

discoveries differs drastically. This benefit glows larger as the mobility increases. So, GMR which computes the most paths shows the least routing overheads. Figure 5 shows how many data packets each protocol delivers successfully. Packet delivery ratio is obtained by dividing the number of data packets correctly received by the destinations by the number of data packets originated by the sources. Two multipath protocols show large difference from DSR in this performance metric. It is because the destination node sends multiple RREPs in multipath protocols while DSR sends only one RREP. In moderately loaded networks, the loss of RREP packets means the failure of the route discovery process. In on-demand protocols, the source uses an exponential back-off algorithm to limit the rate at which it initiates new route discoveries. Since the back-off time is large compared to send buffer capacity, application packets within the back-off time may be dropped due to send buffer overflow. Therefore, multipath protocols which use multiple RREPs shows higher delivery ratio than single path protocols. Between multipath protocols, GMR s delivery ratio is a little higher than that of M-DSR. We think this is because GMR has less routing overheads than M-DSR so network load is lower than that of M-DSR. Fig. 5. Packet Delivery Ratio Figure 6 shows the end-to-end delay of application data packets. As mentioned before, GMR suffers the largest delay in a single route discovery because it uses waiting process. But the end-to-end delay is average delay suffered by all transmitted data packets. So the end-to-end delay of GMR is the least among three protocols. It is because GMR has the least routing overheads so route discovery failures don t happen frequently. If a route discovery fails, data packets in the send buffer are delayed until new discovery finishes. As stated before, the route discovery back-off time is much larger than the route discovery time.

Therefore, GMR which suffers fewer route discovery back-offs shows the shortest end-to-end delay among three protocols. Fig. 6. End-to-end Delay 5 Conclusion We presented the Graph-based Multipath Routing (GMR) protocol for ad hoc networks. Multipath routing schemes try to reduce the number of flooding initiations by searching for multiple paths in a single route discovery. In GMR protocols, each node generates and accumulates the graph data which represents the abstract network topology with distributed manner and relays the graph data toward the destination node. The destination node can use local graph search algorithm to compute link disjoint paths. We have studied the performance of GMR relative to M-DSR and DSR under a wide range of mobility scenarios. We observed that GMR shows the least control packet overhead, the highest packet delivery ratio and the shortest end-to-end delay among three protocols. References 1. C. E. Perkins and P. Bhagwat.: Highly Dynamic Destination-Sequenced Distance- Vector Routing (DSDV) for Mobile Computers. In Proceedings of the ACM SIG- COMM (1994) 234-244 2. D. Johnson and D. Maltz.: Dynamic Source Routing in Ad Hoc Wireless Networks. In T. Imielinski and H. Korth, editors, Mobile computing, chapter 5. Kluwer Academic (1996)

3. V. D. Park and M. S. Corson.: A Highly Adaptive Distributed Routing Algorithm for Mobile Wireless Networks. In Proceedings of the INFOCOM (1997) 1405-1413 4. C. E. Perkins and E. M. Royer.: Ad Hoc On-Demand Distance-Vector Routing. In Proceedings of the IEEE Workshop on Mobile Computing Systems and Applications (1999) 90-100 5. J. Broch, D. Maltz, D. Johnson, Y. C. Hu, and J. Jetcheva.: A Performance Comparison of Multi-Hop Wireless Ad Hoc Network Routing Protocols. In Proceedings of the IEEE/ACM MOBICOM (1998) 85-97 6. B. Williams and T. Camp.: Comparison of Broadcasting Techniques for Mobile Ad Hoc Networks. In Proceedings of the ACM MOBIHOC (2002) 253-264 7. A. Nasipuri, R. Castaneda, and S. R. Das.: Performance of Multipath Routing for On-demand Protocols in Mobile Ad Hoc Networks. ACM/Kluwer Mobile Networks and Applications (MONET) (2001) 6(4):339-349 8. S. J. Lee and M. Gerla.: Split Multipath Routing with Maximally Disjoint Paths in Ad Hoc Networks. In Proceedings of the IEEE ICC (2001) 3201-3205 9. M. K. Marina and S. R. Das.: On-demand Multipath Distance Vector Routing in Ad Hoc Networks. In Proceedings of the International Conference for Network Protocols (ICNP) (2001) 14-23 10. Charu C. Aggarwal and James B. Orlin.: On Multi-route Maximum Flows in Networks. Networks N39 (2002) 43-52 11. K. Fall and K. Varadhan.: ns notes and documentation, December 2000. Available from http://www.isi.edu/nsnam/ns/. 12. The CMU Monarch Group: Wireless and Mobility Extensions to ns-2. Available from http://www.monarch.cs.cmu.edu/cmu-ns.html.