Neural network-based symmetric encryption algorithm with encrypted traffic protocol identification.

Area of Science:

• Cybersecurity
• Information Theory
• Applied Cryptography

Background:

• Cryptography is vital for power grid security, protecting against cyberattacks.
• Identifying cryptographic algorithms is crucial for assessing security and understanding encryption methods (symmetric and asymmetric).
• Analyzing spatial patterns in ciphertext is key to inferring algorithms and securing power system communications.

Purpose of the Study:

• To develop and evaluate a plaintext guessing model (SCGM model) for symmetric encryption algorithms using convolutional neural networks.
• To propose a novel scheme for encrypted traffic protocol identification by integrating temporal and spatial features.
• To assess the performance of cryptographic algorithms in both symmetric and asymmetric frameworks for power grid security.

Main Methods:

• Developed a symmetric cryptographic guessing model (SCGM model) utilizing convolutional neural networks.
• Proposed a new scheme for encrypted traffic identification that combines temporal and spatial data features.
• Evaluated the effectiveness of the proposed methods on various symmetric and asymmetric cryptographic algorithms.

Main Results:

• The SCGM model demonstrated learning efficacy and practical applicability for symmetric encryption.
• The integrated temporal and spatial feature scheme proved effective for encrypted traffic protocol identification.
• Experimental results confirmed the proposed scheme's potential to enhance power grid security.

Conclusions:

• The developed SCGM model and the novel traffic identification scheme are effective for cryptographic algorithm identification.
• The integrated approach shows significant promise for improving the security and resilience of power systems against cyber threats.
• This research contributes to advancing cryptanalysis techniques for critical infrastructure protection.