Data security storage and transmission framework for AI computing power platforms.

Area of Science:

• Computer Science
• Cybersecurity
• Artificial Intelligence

Background:

• Rapid expansion of AI applications necessitates robust data security solutions.
• Existing data storage and transmission methods face challenges in AI computing environments.
• Ensuring data confidentiality, integrity, and availability is critical for AI systems.

Purpose of the Study:

• To propose a novel secure data storage and transmission system for AI computing platforms.
• To enhance data security through a hybrid encryption mechanism and decentralized storage.
• To develop an intrusion detection model for classifying cyberattack types.

Main Methods:

• Developed Secure AI-DST, integrating Amended Merkle Tree (AMerT) hashing and Secret Elliptic Curve Cryptography (SEllC).
• Utilized blockchain with InterPlanetary File System (IPFS) for secure, decentralized storage.
• Deployed an attention bidirectional gated recurrent unit-assisted residual network (Att-BGR) for intrusion detection.

Main Results:

• Reduced unauthorized access attempts by 92.7%.
• Maintained data integrity with 99.98% accuracy under simulated cyberattacks.
• Achieved a 97.6% packet validation success rate in edge-to-cloud transmissions with only 4.3% computational overhead.

Conclusions:

• The Secure AI-DST system effectively ensures end-to-end data security and resilience against cyber threats in AI computing.
• The proposed framework is scalable and suitable for real-time AI workloads.
• This research provides a foundation for next-generation secure AI computing infrastructures.