Efficient and Accurate Link-Quality Monitor

 

Project description

We have developed a highly efficient and accurate link-quality measurement framework, called EAR (Efficient and Accurate link-quality monitoR), for multi-hop wireless mesh networks, that has several salient features. First, it exploits three complementary measurement schemes: passive, cooperative, and active monitoring. EAR maximizes the measurement accuracy by (i) dynamically and adaptively adopting one of these schemes and (ii) opportunistically exploiting the unicast application traffic present in the network, while minimizing the measurement overhead. Second, EAR effectively identifies the existence of wireless link asymmetry by measuring the quality of each link in both directions of the link, thus improving the utilization of network capacity by up to 114%. Finally, its reliance on both the network layer and the IEEE 802.11-based device driver solutions makes EAR easily deployable in existing multi-hop wireless mesh networks without system recompilation or MAC firmware modification. EAR has been evaluated extensively via both ns-2-based simulation and experimentation on our Linux-based implementation. Both simulation and experimentation results have shown EAR to provide highly accurate link-quality measurements with minimum overhead.

 

Approach

EAR is a low-overhead and high-accuracy measurement framework that is aware of asymmetric wireless links and also easily deployable in 802.11-based WMNs. EAR has the following distinct characteristics.

  • Hybrid approach: EAR adaptively selects one of three measurement schemes (passive, cooperative, and active) to opportunistically exploit existing application traffic as probe packets. If there is no application traffic over a link, EAR uses active probing on the link at a reasonable cost. Otherwise, EAR switches itself to passive or cooperative monitoring that gratuitously uses existing traffic for collecting the link-quality information.
  • Unicast-based uni-directional measurement: EAR uses unicast (instead of broadcast) in each direction of a link for measuring its quality. Unicast, which uses the same settings as the actual data transmissions, allows different schemes to generate homogeneous measurements. Moreover, since the quality of each link’s direction is independently measured via unicast, the measurement results are uni-directional.
  • Distributed and periodic measurement: EAR independently measures the quality of link from a node to its every neighbor in a fully-distributed way. This measurement is also taken periodically to cope with the varying link-quality, and the measurement period is also adapted based on a link-quality history.
  • Cross-layer interaction: EAR is composed of “inner EAR” (iEAR) that periodically collects and derives link-quality information in the network layer and “outer EAR” (oEAR) that monitors egress/cross traffic at the device driver. These two components interact across the two layers to intelligently exploit MAC-layer information without any modification of MAC’s firmware.
eararchitecture.png

 

People

Faculty

  • Kang G. Shin

Student

  • Kyu-Han Kim

 

Publications

  • Kyu-Han Kim and Kang G. Shin. Accurate and Asymmetry-aware Measurement of Link Quality in Wireless Mesh Networks. to appear in IEEE/ACM Transaction on Networking. , 2009.
  • Kyu-Han Kim and Kang G. Shin. On Accurate Measurement of Link Quality in Multi-hop Wireless Mesh Networks. Proceedings of the 12th ACM International Conference on Mobile Computing and Networking (MobiCom’06) Los Angeles. , September, 2006. PDF PDF PPT PPT

 

Software Distribution

  • Linux-based Software Code: Click Here.