Reducing Congestion Effects in Wireless Networks by Multipath Routing

  • Authors:

  • Lucian Popa;Costin Raiciu;Ion Stoica;David Rosenblum

  • Affiliations:

  • Department of Computer Science, University of California, Berkeley. popa@cs.berkeley.edu;Department of Computer Science, University College London. c.raiciu@cs.ucl.ac.uk;Department of Computer Science, University of California, Berkeley. istoica@cs.berkeley.edu;Department of Computer Science, University College London. d.rosenblum@cs.ucl.ac.uk

  • Venue:
  • ICNP '06 Proceedings of the Proceedings of the 2006 IEEE International Conference on Network Protocols
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

We propose a solution to improve fairness and increase throughput in wireless networks with location information. Our approachconsists of a multipath routing protocol, Biased Geographical Routing (BGR), and two congestion control algorithms, In-NetworkPacket Scatter (IPS) and End-to-End Packet Scatter (EPS), which leverage BGR to avoid the congested areas of the network.BGR achieves good performance while incurring a communication overhead of just 1 byte per data packet, and has a computationalcomplexity similar to greedy geographic routing. IPS alleviates transient congestion by splitting traffic immediately beforethe congested areas. In contrast EPS alleviates long term congestion by splitting the flow at the source, and performing ratecontrol. EPS selects the paths dynamically, and uses a less aggressive congestion control mechanism on non-greedy paths toimprove energy efficiency. Simulation and experimental results show that our solution achieves its objectives. Extensive ns-2simulations show that our solution improves both fairness and throughput as compared to single path greedy routing. Our solutionreduces the variance of throughput across all flows by 35%, reduction which is mainly achieved by increasing throughput oflong-range flows with around 70%. Furthermore, overall network throughput increases by approximately 10%. Experimental resultson a 50-node testbed are consistent with our simulation results, suggesting that BGR is effective in practice.