Wildcard Fields-Based Partitioning for Fast and Scalable Packet Classification in Vehicle-to-Everything

Area of Science:

• Computer Science
• Network Engineering
• Cybersecurity

Background:

• Vehicle-to-Everything (V2X) communication demands high-speed, secure networks.
• Exponential growth in connected devices causes network congestion and security vulnerabilities.
• Packet classification is fundamental to network performance and security but struggles with large rule sets.

Purpose of the Study:

• To develop a packet classification algorithm that balances speed and scalability.
• To address the limitations of existing algorithms that prioritize speed over scalability with rule-set size.
• To provide a solution essential for emerging services like V2X and the Internet-of-Things (IoT).

Main Methods:

• Proposed a new packet classification approach applicable to decision tree-based algorithms (e.g., HyperCuts, EffiCuts).
• Incorporated explicit consideration of rule duplication for memory efficiency.
• Focused on reducing decision tree size while minimizing performance impact.

Main Results:

• Achieved a balance between classification speed and scalability.
• Demonstrated memory-effective algorithm design through explicit rule duplication handling.
• Substantially reduced decision tree size with minimal loss in classification performance.

Conclusions:

• The proposed approach enables simultaneous high-speed packet classification and scalability.
• This method is crucial for supporting the demands of modern services like V2X and IoT.
• Enhanced network efficiency and security are attainable even with massive rule sets.