ASSOCIATION CONTROL IN VEHICULAR WIRELESS NETWORKS

Open Access
Author:
Barreto, Alejandro César
Graduate Program:
Computer Science and Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
June 16, 2010
Committee Members:
  • Guohong Cao, Thesis Advisor
Keywords:
  • cross-layer interaction
  • vehicular ad-hoc networks
  • wireless networking
Abstract:
Association control algorithms play a significant factor in the performance of mobile clients traversing a dense wireless mesh network. In this paper, we introduce TAR, a novel association control algorithm which uses trip prediction alongside previous knowledge of the wireless network topology in order to select the APs which the client will associate with. We note that different applications have different wireless requirements. Interactive applications will demand a minimum amount of bandwidth and the least number of handoffs which introduce unwanted hiccups in live service (e.g. VoIP or live video). Non-interactive applications have no time demands but desire the maximum throughput possible. With cross-layer communication, the application can notify TAR about its bandwidth requirements and whether it wishes to perform handoff reduction or maximum throughput. Furthermore, we propose a distributed fairness mechanism around a weighted max-min scheme to allow interactive users to gain priority during the association cycle, perform a sort of load balancing, and act as a quality-of-service admission control. This can either be done with assistance from the AP or through peer-to-peer coordination with a mobile peer-to-peer database of client-AP associations. We then show that these algorithms perform better than some commonly known and deployed association control algorithms, which tradeoff simplicity for poor performance in mobile scenarios.