Computer Science

Permanent URI for this collection

https://hdl.handle.net/10530/366

Browse

Browsing Computer Science by Subject "Auto-Configuration"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    IP address auto-configuration for wireless ad-hoc networks
    (University of Zululand, 2019) Mutanga, Murimo Bethel; Adigun, M.O.
    In an ad hoc network, nodes collaborate to allow communication without the presence of network infrastructure. Lack of manual management in ad hoc networks means that automatic configuration is highly desirable. The need for automatic configuration capabilities will become significantly more intense when one considers the networked home of the future, with IP-enabled appliances, such as microwave ovens, thermostats, alarm clocks, speakers, and various kinds of sensors. High levels self-organisation provides an out-of-the-box functionality such that very little technical expertise is required to set up a network. However, efficiently providing unique IP addresses in ad hoc networks is still an open research question. This study is a successful attempt to investigate automatic IP addressing in wireless ad hoc networks as both Multicriteria decision making (MCDM) problem and a challenge of building a system that converges towards the global desired goal. Consequently, the solution proposed in this thesis is inspired by observing swam systems’ ability to converge towards a global goal from local interactions. The investigation reported in this thesis first answered the question of how different network conditions affect the address auto-configuration process. Experiments to investigate the effect of mobility, network traffic and DAD timeout period on address auto-configuration were conducted. The results of these experiments informed the design of the new protocol, D-DAD, proposed in this work. The proposed IP address auto-configuration mechanism was simulated in Ns2 and results were compared with existing Wise-DAD and StrongDAD protocols. We performed five experiments to investigate the effect of network size, node density, node arrival rate, mobility, and network traffic on communication overhead, address uniqueness and latency. The results showed that D-DAD outperformed StrongDAD in all the metrics used for comparison. However, in some instances, D-DAD recorded more communication overhead in comparison to Wise-DAD but had better latency and fewer address conflicts.