QoS-aware Routing for Mobile Ad hoc Networks Based on Multiple Metrics: Connectivity Index (CI) and Delay Kunagorn Kunavut * and Teerapat Sanguankotchakorn † Telecommunications Field of Study, School of Engineering and Technology Asian Institute of Technology, Thailand * kunagorn.kunavut@ait.ac.th, † teerapat@ait.ac.th Abstract - Mobile ad hoc networks are infrastructureless wireless networks; all nodes can quickly share information without using any fixed infrastructure like base station or access point. Thus, nodes are responsible for path selection and maintenance. Most routing protocols implemented on ad hoc networks allow each node to select the path based on hop count or delay which can lead to network congestion. Hence, performance of ad hoc networks can be improved by offering the QoS provisioning either source or hop-by-hop QoS routing. However, it is very difficult to offer QoS guarantee in ad hoc networks because of their characteristics like dynamic topologies and bandwidth- constrained. In this work, we propose an effective algorithm to compute the feasible path by using Connectivity Index (CI) and delay to provide hop-by-hop QoS routing in ad hoc networks. CI is routing metric which indicates connectivity of each node in ad hoc networks; large value of CI indicates more branch node which makes a node more robust to link failure. The performances of the proposed algorithm are investigated by simulation. The results illustrate an important improvement over the traditional ad hoc routing protocol. I. INTRODUCTION Wireless networks can be approximately classified into two main types which are infrastructure and infrastructureless wireless networks. Among these types of wireless networks, infrastructureless-based network called mobile ad hoc networks seem to be more suitable for networks which need rapid deployment and configuration, such as networks in military battlefields, search-and-rescue sites, meetings, conferences and conventions, etc. Ad hoc networks are self- creating, self-organizing and self-administering networks. Each node in ad hoc network behaves as a router and is responsible for discovery and maintenance of the paths in order to quickly share information by using specific routing protocols. Unfortunately, routing protocols which are used in wireline network cannot be simply implemented on ad hoc networks because of their characteristics such as limited bandwidth, limited power and highly dynamic topology. Thus, many ad hoc routing protocols have been proposed in decades. The routing protocols can be categorized into three main groups: Proactive routing protocols, such as Destination Sequenced Distance Vector (DSDV) [1], Topology Dissemination Based on Reverse-Path Forwarding (TBRPF) [2], Optimized Link State (OLSR) [3]; Reactive routing protocols such as Dynamic Source Routing protocol (DSR) [4], Ad Hoc On-Demand Distance-Vector routing protocol (AODV) [5]; and Hybrid routing protocols such as Zone routing protocol (ZRP) [6] and etc. Among these routing protocols, proactive routing protocols seem to be appropriate and flexible enough to support all types of applications especially multimedia and delay-sensitive applications because there is no delay in route discovery process as occurring in reactive routing protocols. Thus, OLSR is selected to implement our algorithm since it is proactive routing protocol and has been standardized by IETF MANET Working Group. OLSR is developed for ad hoc networks and operates as the table-driven routing protocol by exchanging control information to the others regularly. Each node selects a set of its neighbor nodes as "multipoint relays" (MPRs) [7]. Only nodes selected as MPRs are responsible for forwarding control information in order to reduce the number of control overhead sent to the network. However, there is no Quality-of-Service (QoS) provisioning in OLSR. Many works proposed the algorithm to provide the QoS in OLSR [13-16] in the past. However, most of the proposal considered simply only the bandwidth and delay without putting into account the important parameters such as the node or network situation. Consequently, this paper proposes the algorithm to provide QoS-aware routing on OLSR based on novel metrics: Connectivity Index (CI) and delay, called shortest-highest path algorithm in order to improve the performance over OLSR protocol. II. RELATED WORK QoS in mobile ad hoc networks has been received lots of attention since it is quite difficult to be implemented in these types of networks due to their characteristics. Many methodologies have been carried out to find the ways to provide QoS in mobile ad hoc networks. In general, QoS routing in ad hoc networks can be classified into two main paradigms: source and hop-by-hop QoS routing as described in the following topics. 978-1-4244-5607-9/10/$25.00 ©2010 ECTI 978-1-4244-5607-9/10/$25.00 ©2010 ECTI 978-1-4244-5607-9/10/$25.00 ©2010 ECTI 46