Enhancing Datacenter Network Efficiency: AI-Based Path Selection in Software Defined Networks
by researcher Ahmed nadium and supervisor Dr.Ahmed Saad
Traffic steering is a crucial aspect of datacenter networks, where Software-Defined Networking (SDN) and Service Function Chaining (SFC) play significant roles SDN centralizes network management, enabling streamlined control and programmability for efficient traffic handling. This approach facilitates quick, uniform optimization across the entire network infrastructure. SFC provides a flexible framework for directing network traffic through a chain of service functions, enhancing network security and performance. Together, SDN and SFC empower datacenter operators to dynamically steer traffic, improving resource allocation and ensuring reliable service delivery. Traffic steering in datacenters is associated with several critical challenges. These include scalability, load balancing, fault tolerance, and efficient resource utilization. Addressing these problems is crucial to ensure optimal performance, reliability, and cost-effectiveness in datacenter networks. The primary objective of this thesis is to develop a comprehensive suite of intelligent function chains, enhanced with robust firewall and security measures, and optimized File Transfer Protocol (FTP) performance. These function chains are designed to adaptively optimize network performance in response to dynamic traffic demands. Compared with traditional software-defined networks, the simulation results of the proposed system showed an improvement in throughput of up to 76%, along with a reduction in the level of link congestion. The results also exhibit an improvement of up to 54% compared with state-of-the-art load balancing. Additionally, the proposed system showcased a noteworthy enhancement of up to 20% in FTP performance compared to the existing state-of-the-art approaches.