Imran Edzereiq Kamarudin, <br/><br/>
An efficient hybrid multimetric qosbased routing model for software defined networking / 
Imran Edzereiq Kamarudin 
 - xiv, 133 pages :  illustrations ;  30 cm. +  1 CD-ROM. 
<br/><br/> Faculty of Computing 
<br/><br/>Thesis (Doctor of Philosophy) -- Universiti Malaysia Pahang - 2026 
<br/><br/>Includes bibliographical references 
<br/><br/> The rise of network-intensive traffic, such as cloud computing, the Internet of Things (IoT), and latency-sensitive applications, has intensified the need for intelligent and adaptive routing mechanisms in modern network infrastructures. Traditional network architecture faces limitations in supporting dynamic traffic demands due to its tightly coupled control and data planes, which suffer from rigidity and limited programmability. In response, Software Defined Networking (SDN), with its centralized control and programmability, presents an ideal platform for deploying Quality of Service (QoS) based routing models. However, existing QoS-based routing approaches often suffer from limitations such as static metric weighting, excessive computational overhead, and inefficient traffic classification, all of which hinder scalability and responsiveness in dynamic network environments. This research proposes a novel hybrid QoS-based routing model for SDN that integrates current network metrics, adaptive weight assignment, and packet-based traffic classification with rate control. The model is developed in five progressive phases, beginning with a baseline routing framework that computes optimal paths based on available bandwidth, latency, and packet loss. It is subsequently enhanced through a Dynamic Weight Assignment (DWA) module that adjusts link cost weightings in response to Differentiated Services Code Point (DSCP) values, network utilisation, and latency. Further improvements are introduced through a low overhead traffic classification mechanism that utilises DSCP-based packet differentiation and a hierarchical rate limiter to enforce bandwidth allocation across traffic classes. The model is evaluated using Mininet simulations on the Albine network topology under diverse traffic conditions. Performance metrics, including end-to-end delay, jitter, throughput, and controller overhead, are benchmarked against existing routing schemes, such as Dijkstra Shortest Path (DSP), QoS-Aware Routing Mechanism (QARM), and Multi-Constraint Optimal Path (MCOP). Results demonstrate that the hybrid model achieves consistent performance improvement in terms of reducing delay and jitter between 20-30%. This reduction was also achieved while maintaining low control overhead between 10-15%, thus validating its efficiency, scalability, and adaptability. This study contributes to the field by introducing a comprehensive routing framework that balances computational efficiency with responsiveness, offering practical insights for deploying QoS-based solutions in SDN-based infrastructures. Future research can focus on deploying the model on real SDN hardware platforms to assess practical feasibility, benchmarking it on larger and more complex network topologies to explore scalability limits, and incorporating application-aware QoS metrics to enhance servicelevel differentiation across specialized SDN domains such as SDIoT and SD-WAN. 
<br/><br/><label>ISBN: </label>THE0010799 (Local) 
<br/><br/><label>Subjects--Corporate Names: </label><br/>Faculty of Computing--Dissertations
<br/><br/><label>Subjects--Topical Terms: </label><br/>Universities and colleges--Dissertations<br/>Thesis--Dissertations