Randomized dynamic route maintenance for adaptive routing in multihop mobile ad hoc networks

  • Authors:

  • Wook Choi;Sajal K. Das;Jiannong Cao;Ajoy K. Datta

  • Affiliations:

  • Center for Research in Wireless Mobility and Networking (CReWMaN), Department of Computer Science and Engineering, University of Texas at Arlington, Arlington, TX 76019-0015, USA;Center for Research in Wireless Mobility and Networking (CReWMaN), Department of Computer Science and Engineering, University of Texas at Arlington, Arlington, TX 76019-0015, USA;Department of Computing, Internet and Mobile Computing (IMC) Lab., Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong;Department of Computer Science, University of Nevada at Las Vegas, Las Vegas, NV 89154-4019, USA

  • Venue:
  • Journal of Parallel and Distributed Computing
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Several approaches have been proposed for designing multihop routing protocols in mobile ad hoc networks (MANET). Many of them adopt a method, called flooding, to discover a routing path. Due to the time-varying nature of the route in MANET, the discovered route needs to be dynamically maintained for optimality in terms of traffic load, hop-distance, and resource usage. It is easy to see that flooding incurs significant overhead and hence is inappropriate for the dynamic route maintenance. In this paper we propose a randomized, dynamic route maintenance scheme for adaptive routing in MANET. The scheme makes use of a nomadic control packet (NCP) which travels through the network based on a random walk, and collects its stopovers as a traversal record. The NCP uses the traversal record to probabilistically provide the nodes with clue for routing path updates. From the clue, the nodes can find the routing path update information that is up-to-date and optimal (less-loaded and shorter), thereby adapting to the dynamic network topology and traffic load conditions. We present an analytical model for measuring the effectiveness of NCP in terms of its frequency of visits and probability of finding the clue from the NCP traversal record. The proposed randomized scheme serves as a routing protocol supporting layer and can be easily applied with minimum modifications to the existing on-demand routing protocols such as AODV and DSR. In our experimental study, we modified the AODV protocol to maintain routing paths using NCPs' traversal record. Simulation results show that NCPs help the routing protocol to notably reduce average end-to-end packet delay with increased route optimality and better control on traffic congestion.