Mitigating Congestion in Wireless Networks

Open Access
Kumar, Raju
Graduate Program:
Computer Science and Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
May 17, 2010
Committee Members:
  • Thomas La Porta, Dissertation Advisor
  • Thomas La Porta, Committee Chair
  • Guohong Cao, Committee Member
  • Sencun Zhu, Committee Member
  • Aylin Yener, Committee Member
  • Rate Selection
  • Network Coding
  • Congestion
  • Wireless Networks
  • Multi-Rate
  • Routing
Cellular networks and Wireless LAN provide ubiquitous wireless connectivity today. These one-hop wireless networks have been extended to multi-hop wireless networks like mesh, ad hoc, and sensor networks. Since the wireless medium is broadcast in nature, multi-hop wireless networks are susceptible to congestion due to excessive traffic in a part of the network. In this dissertation, five algorithms are proposed to ameliorate the impact of congestion on different wireless networks under separate scenarios. An important aspect of congestion in wireless networks is its indiscriminate impact. Data in sensor networks may not all be equally important and congestion in such a network may lead to low priority data being delivered while high priority data is dropped en route to the destination. A differentiated routing scheme is proposed to address this issue. Network coding, heralded as a means to better utilize the medium's capacity as compared to traditional routing's approach of store-and-forward, is also susceptible to congestion. Channelization, or reserved access to the medium, has been proposed as a means to eliminate congestion by assigning parts of the medium to different links. In this dissertation, channelization schemes are proposed for intra-flow network coding of multicast flows in a wireless network. These channelization schemes effectively restore the performance of network coding. Congestion in wireless networks has also been addressed by using appropriate transmission rates in a multi-rate wireless MAC. But a multi-rate medium raises the issue of rate selection, the problem of selecting which transmission rate to use. When inter-flow network coding is used to deliver unicast traffic in such a network, it requires some packets to be multicast. A rate selection scheme is proposed to determine the appropriate transmission rate while taking network coding into account and ensuring that the throughput on the multicast link is maximized. Finally, all prominent rate selection schemes are hop-by-hop. We propose an end-to-end rate selection scheme for networks that leverage opportunistic reception as a means to increase the medium capacity to counter congestion.