EASE-6G: An Energy-Aware SDN Framework with Proactive Slicing and DL-Based Overhead Mitigation for Scalable IoT...

Area of Science:

• Telecommunications Engineering
• Computer Networking
• Artificial Intelligence

Background:

• Sixth-generation (6G) networks promise high data rates (1 Tbps) and low latency (<0.1 ms) for Internet of Things (IoT).
• Increased device density in 6G networks leads to significant challenges in control plane signaling and energy consumption, impacting scalability and efficiency.
• Current network architectures struggle to manage the demands of ultra-dense IoT environments.

Purpose of the Study:

• To propose EASE-6G, an energy-aware Software-Defined Networking (SDN) framework designed for ultra-dense 6G IoT networks.
• To enhance network scalability and efficiency by shifting from reactive to proactive and predictive operations.
• To reduce control plane signaling and energy consumption while maintaining network performance.

Main Methods:

• Implemented an energy-aware SDN framework (EASE-6G) utilizing Proactive Flow Installation.
• Employed a Long Short-Term Memory (LSTM) model for traffic prediction.
• Utilized a signaling-aware Deep Q-Network (DQN) for streamlined control.
• Conducted simulations using OMNeT++/Simu5G, comparing EASE-6G against SF-RAN and DQN-ORAN.

Main Results:

• EASE-6G reduced energy consumption by 36.8%, signaling overhead by 36.7%, and latency by 35.6%.
• The LSTM model achieved a Mean Absolute Percentage Error (MAPE) of 4.2% for traffic prediction.
• The DQN agent demonstrated improved stability with 22% lower variance compared to the baseline.

Conclusions:

• The proposed predictive SDN control mechanisms in EASE-6G effectively improve energy efficiency and reduce overhead in 6G networks.
• EASE-6G offers a practical solution for scalable and sustainable IoT implementation in future 6G environments.
• Proactive and predictive network operations are crucial for managing ultra-dense IoT deployments.