A Systematic Approach for Performance Efficiency of Distributed Networks Using Machine Learning Techniques and Network Simulator 2
A. Parveen Akhther, N. Veera Subrahmanyam, Adabala Murali Veera Sri Sai, Prasanth V. S · 2025
In the rapidly evolving domain of distributed networks, including cloud computing and edge systems, achieving optimal performance efficiency is critical. Existing methods often fail to effectively handle dynamic policies, resource allocation, and scalability. This project presents a robust framework leveraging machine learning (ML) techniques integrated with NS2 simulation to address these challenges. Data collection from network nodes is the first step in the process. Preprocessing is then done to manage missing values, eliminate noise, and normalize the data for model training. Supervised learning methods like regression and classification optimize traffic prediction and routing protocols, while unsupervised techniques, including clustering, identify resource patterns for better allocation. Reinforcement learning dynamically adjusts policies to improve fault tolerance and resource management. NS2 simulation validates the model's effectiveness across various network scenarios. A remarkable accuracy of 98.77% and a minimal error rate of 1.22% are attained by the claimed framework.