Securing O-RAN with Zero Trust Architecture and Large Language Models
摘要
The Open Radio Access Network (O-RAN) architecture is critical for the development of 6G networks, offering flexibility and interoperability through disaggregated components. However, this openness exposes O-RAN to new security vulnerabilities, including unauthorized access, data breaches, and malicious xApp deployments. To address these challenges, we propose DistillORAN, a novel Zero-Trust architecture designed specifically for O-RAN. DistillORAN features two core components: (1) a blockchain-based decentralized trust management system for secure verification, authentication, and dynamic access control of xApps, and (2) a lightweight intrusion detection module powered by DistilBERT, a transformer-based model optimized for resource-constrained environments. DistilBERT’s ability to analyze network activities and detect anomalies in real-time allows it to identify complex security threats and multi-step attack scenarios within the O-RAN ecosystem. Its lightweight nature makes it ideal for O-RAN’s distributed infrastructure, where computational resources may be limited. By combining blockchain technology for trust management with DistilBERT’s powerful pattern recognition for intrusion detection, DistillORAN enforces a Zero-Trust security model, ensuring continuous monitoring and verification of all network components. This comprehensive solution enhances the security and resilience of O-RAN networks, aligning with the dynamic needs of next-generation mobile infrastructures.