Page 347 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

intelligent decision-making, ultimately supporting the evolution of smarter and more efficient
5G infrastructure.

In addition, by significantly lowering energy consumption, the approach contributes to reducing
the carbon footprint of telecom operations. The use of a dynamic max-min planning framework 4.3 - 5G
empowers network operators to implement greener practices and actively support efforts to
mitigate climate change.

2�3 Future Work

1. Real-Time Implementation and Testing: Transition from framework and idea to simulation
and to live testbed deployment, integrating with real/emulated 5G environments.
2. Edge-AI Integration: Explore lightweight AI modules for deployment at edge nodes to
reduce latency and increase local autonomy.
3. Multi-Objective Optimization: Extend the current model to optimize for multiple
conflicting objectives like latency, throughput, energy, and reliability.
4. Security Integration: Embed security features into the SDN controller for real-time
detection of intrusions and malicious traffic in the 5G environment.
5. Scalability to 6G: Investigate how this framework can evolve for 6G technologies, including
integration with intelligent surfaces, terahertz links, and satellite backhaul.

3 Use Case Requirements

• UC1: End-to-end Traffic Monitoring by continuously monitoring end-to-end traffic
demands in different time slots to capture dynamic traffic patterns.
• UC2: Traffic Prediction: We shall work on employing traffic prediction algorithms to
forecast future traffic demands based on historical data.
• UC3: Link State Adjustment: Dynamically adjust the active/sleeping state of network links
based on predicted traffic demands and network connectivity requirements.
• UC4: Max-Min Optimization: Implement a max-min optimization algorithm to determine
the optimal network topology and resource allocation for maximizing throughput and
minimizing energy consumption.
• UC5: H-CRAN Integration: Integrate the solution within an H-CRAN architecture, utilizing
the control plane to generate max-min planning and distribute energy-efficient planning
schemes to the data plane.
• UC6: Distributed SDN Control: Utilize a distributed SDN controller to manage the
network, enabling scalability and resilience.
• UC7: Time Slot-Based Planning: Implement max-min planning for each time slot based
on traffic demand, physical topology, maximum network flow, and minimum energy
consumption.

4 Sequence Diagram

The sequence diagram illustrates the interaction between network entities, including the
user, RRH, BBU, SDN controller, and monitoring agent, highlighting how real-time traffic and
energy metrics are used to optimise routing and activate or sleep links for energy-efficient
data transmission.

311