University of Glasgow
Development on quality of service and network resilience based on structured geometric topologies
Abstract
dc:description.abstractThe topology of a computer network is the basis for building a high-capacity, low-congestion network infrastructure. At present, the research on computer network topologies rests mainly on typically unstructured networks such as mesh networks. Network infrastructures do not widely employ structured geometric network topologies. Compared to a structured geometric topology, they possess numerous advantageous characteristics, such as connectedness, which are currently not being utilised in the construction of the network structure. Simultaneously, by suggesting additional network topologies, we can select an appropriate network configuration that enhances the availability and robustness of networking infrastructures. Our study evaluates the potential use of three structured geometric network infrastructure designs, namely cross polytope (CP), hypercube (HC) and triangular pyramid (TP), based on their network performance. In this paper, we use simulation modelling to analyse structured network geometries and evaluate their network performance by Riverbed Modeler. We compare the simulation results of an unstructured network design with three structured network topologies for both time-independent and time-dependent applications. The simulation results illustrate that the CP and TP topologies have better results than the unstructured network in response time and network delay under a high-load configuration. Furthermore, we evaluate the performance of structured geometric network infrastructure designs under three QoS disciplines, namely weighted fair queuing (WFQ), priority queuing (PQ) and modified weighted round robin (MWRR). In this paper, we simulate structured geometric and unstructured network structures using Riverbed Modeler 18.9 to evaluate network performance. The simulation results illustrate that the structured geometric network topologies have better results than the unstructured network in response time and network delay under a high-load configuration, and the QoS algorithm provides advantages for voice services by prioritising the transmission of vital services.
Degree
thesis:*- Level dc:type.qualificationlevel
- PhD
- Grantor dc:publisher.institution
- University of Glasgow
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Rao, Haotian
Subjects
dc:subject × 3Rights
- Language dc:language
- en
Identifiers
dc:identifier.*- DOI dc:identifier.doi
- https://doi.org/10.5525/gla.thesis.86112