A New Efficient Multi-path Routing Protocol for Mobile Ad hoc Networks

Phu Hung Le
2017 IJARCCE  
A Mobile Ad hoc NETwork (MANET) is a network that does not require any presence of infrastructure. Mobile nodes can move at any time and cannot predict beforehand. Each node acts as a router to communicate with other nodes in the network. The movement of nodes is a big challenge for routing. Another challenge is interference. Interference in the network is one of the most important problems to study. Some interference-aware multi-path protocols were proposed to minimize the impact of
more » ... e. However, these protocols are not highly efficient or have a very high computational complexity (NP-hard). In this paper, we propose 1) a definition of interference, 2) a formula of interference and 3) develop a novel Linkdisjoint Interference-Aware Multipath protocol (LIA-MPOLSR) for mobile ad hoc networks. LIA-MPOLSR finds paths only in polynomial time. From our simulation results, we show that LIA-MPOLSR significantly improves performance in terms of packet delivery fraction, end-to-end delay, routing overhead and normalized routing load compared to the prominent protocol Ad hoc On-demand Multipath Distance Vector (AOMDV). I. INTRODUCTION Wireless networks can be divided into two types: infrastructured wireless networks and infrastructureless wireless networks. Infrastructured wireless networks have a wire backbone including stationary base stations (access points). Each access point (AP) has a coverage range. Nodes within this coverage range can directly exchange signals with the AP. In general, mobile nodes do not communicate directly with each other. In contrary to infrastructured wireless networks, infrastructureless wireless networks (ad hoc networks) are the networks without any pre-existing communication infrastructure. Wireless mobile nodes can freely and dynamically self-organize into arbitrary and temporary network topologies. Because of their unique characteristics, ad hoc networks have various applications such as disaster recovery, emergency services, defense, healthcare, education, corporate conventions/meetings, indoor and personal networks, as well as sensor networks. However, ad hoc networks are also faced with many challenges, for example, limited bandwidth, low battery, high loss rate, frequent link breakage, etc. Routing is one of the most important problems in a network. Routing in ad hoc networks includes three catalogs of protocols: Proactive (table-driven), reactive (on-demand), and hybrid. Proactive (table-driven) routing protocols reduce delay by maintaining and updating the routing table and the network topology information at each node. However, this approach has high traffic overhead. On-demand routing protocols only calculate paths when the source node needs to transmit packets. Thus, on-demand protocols reduce routing overhead but the delay is high. Hybrid protocols are designed to take advantages of both proactive and on-demand routing protocols. In ad hoc networks, topology is very dynamic and route is frequently broken. There exist some first protocols for ad hoc networks such as on-demand protocols Ad hoc On-Demand Distance Vector (AODV) [2], Dynamic Source Routing (DSR) [3] and proactive protocols Destination Sequenced Distance-Vector (DSDV) [4], The Optimized Link State Routing (OLSR) [1]. Because single path protocol is not stable, high loss rate, and high traffic overhead, many multipath protocols have been proposed to increase the network performance. Multipath is the technique that uses multiple routes to send packets. Therefore, they can reduce packet loss rate, delay and increase stability. Multi-path can be divide into three categories: -Node-disjoint multi-path: the paths have only common source and destination. -Link-disjoint multi-path: the paths can share a few common nodes but not links. -Hybrid multi-path: the paths may be some common links and nodes. In node-disjoint multi-path when nodes in a path move, they do not affect the other paths. This is different from linkdisjoint and hybrid multi-path in the sense that in link-disjoint and hybrid multipath if common nodes in a path move, they impact on the other paths. In a dense network, node-disjoint multi-path is easier to be formed. Therefore, nodedisjoint multi-path has many advantages compared to link-disjoint and hybrid multipath. However, in a quite sparse network, node-disjoint multi-paths are difficult to be established. In contrary, Link-disjoint multi-path can be formed in a dense and quite sparse network. In a quite sparse network, the Another interference protocol is the Expected Transmission Count (ETX). The ETX finds paths that deliver a packet to its destination with the lowest expected number of transmissions (including retransmissions) required. The metric uses per-link measurements of packet loss ratios in both directions of each wireless link to predict the number of retransmissions. The ETX metric can obtain high throughput in static networks. However, the authors of [21] demonstrated that these metrics do not perform well in mobile ad hoc networks because they do not rapidly adapt to the change of topology. Some protocols such as [14] , and [15] are based on the Expected Transmission Count (ETX) metric, thus they are not efficient for mobile ad hoc networks. Our protocol is implemented in NS2. We use an area of 550m x 550m to randomly place 40 nodes that have a carrier sensing range of 400 meters, and a transmission range of 160 meters. The nodes move with their speed from 1m/s to 30 m/s. The pause time of a node is 5s. We use the distributed coordination function (DCF) of IEEE 802.11 as the MAC layer. Traffic source is CBR and channel capacity is 11 Mbps. Constant Bit Rate (CBR) is kept at 100 Kbps. The Two-Ray Ground and the Random Waypoint models are also used in our simulations. The simulation time is 120 seconds. Each data column represents an average of two runs with different random mobility scenarios. The RTS/CTS mechanism is turned on and turned off alternatively.
doi:10.17148/ijarcce.2017.6454 fatcat:pxp6f4gesvhv5g2uyshv5nmu5y