Mobile IP movement detection optimisations in 802.11 wireless LANs

Master Thesis

2005

Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher

University of Cape Town

License
Series
Abstract
The IEEE 802.11 standard was developed to support the establishment of highly flexible wireless local area networks (wireless LANs). However, when an 802.11 mobile node moves from a wireless LAN on one IP network to a wireless LAN on a different network, an IP layer handoff occurs. During the handoff, the mobile node's IP settings must be updated in order to re-establish its IP connectivity at the new point of attachment. The Mobile IP protocol allows a mobile node to perform an IP handoff without breaking its active upper-layer sessions. Unfortunately, these handoffs introduce large latencies into a mobile node's traffic, during which packets are lost. As a result, the mobile node's upper-layer sessions and applications suffer significant disruptions due to this handoff latency. One of the main components of a Mobile IP handoff is the movement detection process, whereby a mobile node senses that it is attached to a new IP network. This procedure contributes significantly to the total Mobile IP handover latency and resulting disruption. This study investigates different mechanisms that aim to lower movement detection delays and thereby improve Mobile IP performance. These mechanisms are considered specifically within the context of 802.11 wireless LANs. In general, a mobile node detects attachment to a new network when a periodic IP level broadcast (advertisement) is received from that network. It will be shown that the elimination of this dependence on periodic advertisements, and the reliance instead on external information from the 802.11 link layer, results in both faster and more efficient movement detection. Furthermore, a hybrid system is proposed that incorporates several techniques to ensure that movement detection performs reliably within a variety of different network configurations. An evaluation framework is designed and implemented that supports the assessment of a wide range of movement detection mechanisms. This test bed allows Mobile IP handoffs to be analysed in detail, with specific focus on the movement detection process. The performance of several movement detection optimisations is compared using handoff latency and packet loss as metrics. The evaluation framework also supports real-time Voice over IP (VoIP) traffic. This is used to ascertain the effects that different movement detection techniques have on the output voice quality. These evaluations not only provide a quantitative performance analysis of these movement detection mechanisms, but also a qualitative assessment based on a VoIP application.
Description

Reference:

Collections