Traffic Aware Routing in P4 Based Software Defined Networks

نوع: Type: Thesis

مقطع: Segment: masters

عنوان: Title: Traffic Aware Routing in P4 Based Software Defined Networks

ارائه دهنده: Provider: Ahmad Hamid

اساتید راهنما: Supervisors: Reza Mohammadi (Ph. D)

اساتید مشاور: Advisory Professors:

اساتید ممتحن یا داور: Examining professors or referees: Shakoor Vakilian (Ph. D) - Mehdi Abbasi (Ph. D)

زمان و تاریخ ارائه: Time and date of presentation: 2025

مکان ارائه: Place of presentation: سالن خان محمدی

چکیده: Abstract: The main challenge in Software-Defined Networking lies in how to efficiently manage network traffic under dynamic and constantly changing conditions. Traditional routing methods, which rely on static paths or slow updates, often fail to adapt to rapid variations in traffic patterns. This limitation reduces overall network performance, degrades Quality of Service, and prevents optimal utilization of network resources. Therefore, there is a growing need for an approach capable of reacting to changes in real time and dynamically optimizing data paths based on the current state of the network. To address this challenge, this research introduces an integrated framework for traffic-aware routing in SDN. In the proposed architecture, a programmable data plane and centralized decision-making in the control plane operate in coordination to dynamically select and reconfigure routing paths according to real-time network conditions. By establishing a feedback loop between the data and control planes, the framework enables continuous monitoring of the network and allows the controller to make routing decisions based on updated link and flow metrics. The design principles emphasize separation of responsibilities between control and data layers, minimal processing overhead in switches, compliance with network policies, and scalability across diverse topologies. In the experimental implementation, the data plane utilizes the P4 programming language to monitor traffic behavior at the switch level, collecting lightweight statistics such as throughput, delay, and packet loss. The ONOS controller analyzes this telemetry data and applies weighted algorithms, such as Dijkstra’s algorithm, to compute optimal paths. Experimental results demonstrate that the proposed method balances network load more effectively, mitigates congestion, and significantly improves user-perceived Quality of Experience, while maintaining low processing complexity in the data plane. Overall, this study presents a practical and transferable framework for programmable networks that establishes a balance between control-plane agility and data-plane efficiency. The proposed approach provides a solid foundation for future research in multi-criteria routing, QoS assurance, and the integration of machine learning techniques to achieve more adaptive and intelligent network control.