Low-latency handoff inter-WLAN IP mobility with broadband network control

  • Authors:
  • Moshiur Rahman;Fotios Harmantzis

  • Affiliations:
  • AT&T Labs, 200 Laurel Avenue, Middletown, NJ 07748, USA;Stevens Institute of Technology, Hoboken, NJ 07030, USA

  • Venue:
  • Computer Communications
  • Year:
  • 2007

Quantified Score

Hi-index 0.24

Visualization

Abstract

Wide-bandwidth and low-cost Wireless LANs (WLANs) have emerged as a competitive choice, not only for wireless high-speed Internet access, but also for voice network access. High-speed Broadband Access Network-Controlled (BANC) mobility management greatly facilitates inter-WLAN IP mobility, through the integration of access technologies and higher layer mobility features. However, Mobile IP (MIP), the current IP solution for mobility, cannot be optimized in obtaining seamless handoff for all cases, unlike the BANC mobility management employing IEEE 802.11b WLAN infrastructure mode that we propose in Rahman and Harmantzis [M. Rahman, F. Harmantzis, IP mobility with high speed access and network intelligence, in: IEEE Wireless Communications and networking Conference, WCNC, Atlanta, GA, March 21, 2004]. Even with the proposed optimization techniques, the MIP is still not entirely tunnel-free. It is not efficient due to certain unavoidable data loss during handoff for time-sensitive services, such as the Voice-over-IP (VoIP). In this paper, we present a BANC-supported low-latency handoff mobility management scheme, by modifying the enhanced IAPP messaging for the WLAN infrastructure mode. We then examine main performance metrics for IP mobility management: delay, packet loss, handover execution time, throughput, and signaling overload for the fast handover MIPv6 (FMIPv6) [R. Koodli, Fast handovers for mobile IPv6, IETF Internet Draft, draft-ietf-mishop-fsat-mipv6-01.txt, February 2004] and BANC-offered mobility management schemes. Simulation results show that BANC has better performance than FMIPv6 for inter-WLAN IP mobility management. As the network server is the main component in the BANC scheme, we have evaluated the impact of the Mobile Node (MN) traffic load, the server packet forwarding rate and datagram processing schemes (central vs slot-based processing), on the CPU utilization to aid BANC service providers' capacity and traffic planning.