AI-Driven MPLS Traffic Optimization Using Real-Time Composite Scoring in GNS3

This paper presented an AI-driven traffic optimization framework for MPLS networks, implemented and validated in a GNS3 simulation environment. The proposed system employs a real-time weighted composite scoring function to dynamically select the optimal Label Switched Path based on live measurements of latency, packet loss, and available bandwidth. A Flask-based web dashboard provides continuous visibility into network performance across both candidate paths. Experimental evaluation across three traffic scenarios demonstrated significant improvements over static MPLS path selection: up to 34% reduction in average latency, up to 50% reduction in packet loss, and 27% improvement in bandwidth utilization. The system's low deployment complexity - requiring no SDN controller, no training data, and no specialized hardware - distinguishes it from prior RL-based approaches and makes it practically deployable in production enterprise MPLS networks.

1.                E. Rosen, A. Viswanathan, and R. Callon, "Multiprotocol Label Switching Architecture," IETF RFC 3031, Jan. 2001. 2.                D. Awduche et al., "RSVP-TE: Extensions to RSVP for LSP Tunnels," IETF RFC 3209, Dec. 2001. 3.                R. S. Sutton and A. G. Barto, Reinforcement Learning: An Introduction, 2nd ed. MIT Press, 2018. 4.                GNS3 Technologies Inc., GNS3 Network Simulator Documentation. [Online]. Available: https://www.gns3.com. [Accessed: Jun. 2026]. 5.                D. O. Awduche et al., "Requirements for Traffic Engineering Over MPLS," IETF RFC 2702, Sep. 1999. 6.                E. Rosen and Y. Rekhter, "BGP/MPLS IP Virtual Private Networks (VPNs)," IETF RFC 4364, Feb. 2006. 7.                H. Mao, R. Netravali, and M. Alizadeh, "Neural Adaptive Video Streaming with Pensieve," in Proc. ACM SIGCOMM, Aug. 2017, pp. 197-210. 8.                A. Valadarsky, M. Schapira, D. Shahaf, and A. Tamar, "Learning to Route," in Proc. ACM HotNets, Nov. 2017, pp. 1-7. 9.                K. Rusek et al., "RouteNet: Leveraging Graph Neural Networks for Network Modeling and Optimization in SDN," IEEE J. Sel. Areas Commun., vol. 38, no. 10, pp. 2260-2270, Oct. 2020. 10.             L. Gong and A. Rezaeipapan, "A Fuzzy-Based Dynamic Routing Algorithm for QoS Provisioning in MPLS Networks," J. Netw. Syst. Manag., vol. 29, no. 3, 2021.